═══ 1. About this forum ═══ TALKLINK Welcome to the world of TalkLink, a unique, new conferencing service offered by IBM. Using TalkLink, you can "talk" to customers, IBMers and personnel of our many affiliates around the world - anywhere, anytime. TalkLink is designed for people like yourself, who want to get the most out of their investment in computing products and services. TalkLink "links" you directly into the development and sup- port communities of customers, IBM and affiliates. Our goal is to help you be as effective as you can with access to many conferences, vast databases and rich libraries. HIGHLIGHTS TalkLink provides the following capabilities for customers to review information electronically, submit questions re- quests and receive answers information. - Conferences and forums -- question and answer bulletin boards for you to ask questions and help others with a wide variety of products and services. Items posted on this bulletin board are shared between BBS users, IBM personnel and affiliates worldwide. - Customer-to-customer messaging -- customers can send private messages to each other. - Support -- customers can submit or view problem re- ports. These entries will be responded to by the conference/forum owner, normally within a certain time period. Users can also place orders for product related materials. - Software library -- downloads application programs and uploads contributions. - News and announcements -- contains recent information regarding products and services. - Modes of operation: - Interactive -- Users are connected online to TalkLink while using its facilities. - Batch -- Users can download the forums to their personal computer and read the entries offline. Users can also respond offline and the entries will be saved and sent to the host the next time the user is connected in batch mode. Initially, this facil- ity is available for OS/2 (R) users only. - Access: 1 - Advantis -- local call from most cities within the US. - Internet -- telnet avaialable now, full ftp soon. SEARCH SERVICE - Search IBM databases via interactive mode -- This search facility includes a large question and answer database of items, IBM System Center Flashes, and IBM's prob- lem management libraries. This database contains infor- mation about problems and fixes. Not all general discussion forums have related information in these da- tabases. TALKLINK FEES CHARGES - $15.00 Registration Fee (one-time charge for each user registration) - $18.00 Monthly Availability Charge (includes unlimited hourly service) - This can be charged to a credit card or invoiced (for an additional administrative fee of $7.00 for each monthly invoice). HOW TO ORDER TALKLINK Call 1-800-547-1283 to sign up for TalkLink. Outside of the United States please contact your local IBM sales office for availability and access. ********************************************************************** ═══ 2. Introduction ═══ ===== OS2SPM20 CFORUM created at 22:03:42 on 92/11/16 GMT (by HOPKINS at AUSVM8) IBM UNCLASSIFIED ------------ The information in this forum is considered by IBM to be external and may be shared by everyone who is given access. This forum is shared with the internal IBM employee conferences. Forum: OS2SPM20 CFORUM Forum Topic: OS/2 System Performance Monitor/2 This CFORUM is provided as a conferencing tool to promote the discussion of issues, questions, and problems associated with the use of OS/2 System Performance Monitor/2 Product Number 96F8379 IBM System Performance Monitor/2 2.0 (3.5) SPM/2 96F8380 IBM System Performance Monitor/2 2.0 (5.25) SPM/2 96F8383 IBM System Performance Monitor/2 2.0 Add License 96F8384 IBM System Performance Monitor/2 2.0 Add License Upgrade 96F8385 IBM System Performance Monitor/2 2.0 Distributed 96F8386 IBM System Performance Monitor/2 2.0 Upgrade Distributed Contributions must not contain confidential or proprietary information. Each contributor agrees, in placing material in this forum, that IBM and other users are free to prepare derivative works based upon it, and to use, reproduce, perform, display and distribute inside and outside their company such material and such derivative works, without obligation or charge. IBM is not responsible or liable for the use, copying, or distribution of materials placed on this forum, and has therefore no obligation to control such acts. All subscribers are required to abide at all times by the conference rules described in the Bulletin Board Rules under "News and Announcements". All IBM personnel are required to abide at all times by the IBMPC RULES and the IBM Business Conduct Guidelines. IBM personnel will read the CUSTOMER FORUM RULES in IBMPC RULES before contributing. Contributions violating these rules will be deleted by the CFORUM Owner or their delegate, and in some cases by an IBM conference Administrator. The OS2BBS CFORUM is where anyone may raise concerns regarding rules violations found in this and other CFORUMS. The IBM Owner of this forum is: Name: SPM/2 Product Support and Development Mailing Address: IBM Corporation 11400 Burnet Road Internal Zip 9171 Austin, Texas 78758 IBM Node/Userid: AUSVM1/ATIG1 The IBM Administrator of this forum is: Name: Paul Cheatham Mailing Address: Boca Electronic Support Team ( BEST ) 1000 N.W. 51st Street Internal Zip 1426-012/W2030 Boca Raton, FL 33429 IBM Node/Userid: BCRVM1/CHEATHAM TalkLink Userid: LNK9/LNK9030 --------------------------------------- Here is the list of Netbios requirements by SPM/2 component: -------------------------------------- Component NBNames NBCs Sessions Remarks SPMCtrl 1 2 2 Only during query, no retry. SPMMon 1 2 2 Only during query, will retry. NBQRspnd 1 2 1 Only needed when any other (Query function) query user is running SPMILog 1 2 2 For EACH SPMILOG, persists SPMNBL 1 2 2 Persists Standalone collection on a machine on a network (for example) requires: -------------------------------------- Standalone 4 8 7 The Maximum requirement: ------------------------ Max 29 58 57 **************************************** * Attached below are the appends that we made to the * PRIVATE SPM20 CONFERENCE DISK. THEY ARE PROVIDED FOR YOUR REVIEW.. * See OS2SPM20 931CFORU on OS2BBS for appends to 92/10/21 (4945 lines) See OS2SPM20 932CFORU on OS2BBS for appends to 93/02/01 (4436 lines) See OS2SPM20 933CFORU on OS2BBS for appends to 93/04/27 (4417 lines) ..... 4417 lines pruned at 19:47:07 on 93/04/27 GMT (by OS2 at YKTVMV) ..... ********************************************************************** ═══ 3. SMP20 under March Beta. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Do I need any patches to SPM/2 2.0 to run under the GA+SP or for the 6.498 March Beta? (I ordered it In January, our purchasing lost it once, Crowntech lost the order too, IBM had this in back order, BUT I HAVE IT NOW!!) Geza Szivos TEAM-OS2 NOVACOR Chemicals (403)-290-6583 Fax (403)-290-5900 NHO-32 801-7th Ave Calgary, Alberta, CANADA T2P 2N6 ********************************************************************** As soon an I started the THESEUS2 program in my Compaq 486 Deskpro I the program beeped 4-5 times and i got a TRAP D. I'm running 6.498G with the HPFS and REXX fixes, CM/2, LS 3.0 requester TCPIP 1.2.1 I don NOT have any .ini fixes installed. Geza Szivos TEAM-OS2 NOVACOR Chemicals (403)-290-6583 Fax (403)-290-5900 NHO-32 801-7th Ave Calgary, Alberta, CANADA T2P 2N6 ********************************************************************** [Reference to append at 19:07 on 93/04/27 by XXNOVA24 at TORVMCOP ] Geza, You don't need any patches to SPM to run on OS/2 2.0 GA + SP. You need one patch to run on the OS/2 2.1 6.498 beta. There is a THESEUS2 PACKAGE on OS2TOOLS that contains fixes to make it run right on 6.498. Without these fixes Theseus2 will trap. This version of Theseus2 is also available on OS2BBS and on Compuserve. Pam Lilly ********************************************************************** Thanks. So it is safe to install it to our LS 3.0 Adv and Entry PRODUCTION servers. (Running GA+SP). I'm planning to montor them from my 2.1 beta. Geza Szivos TEAM-OS2 NOVACOR Chemicals (403)-290-6583 Fax (403)-290-5900 NHO-32 801-7th Ave Calgary, Alberta, CANADA T2P 2N6 ********************************************************************** Yes, it is safe to install SPM on your OS/2 2.0 + SP production systems. Just remember to get the Theseus2 fix for your OS/2 2.1 system. Pam Lilly ********************************************************************** Pam, Does the Theseus2 fix enable it to remaining working on 2.0 GA+SP? Frank ********************************************************************** I try to make the THESEUS2 fixes so that the previous versions of the OS (at least the ones that have been shipped out) still work. The 2.0i version of THESEUS2 works on 2.0GA, SP, 2.1 beta 1, & 2.1 beta 2. The 2.0n version works on those + 2.1 development (post beta 2). (Remember that the old THESEUS works on 1.0, 1.1, 1.2, & 1.3. Those systems were FAR more incompatible than the 2.x systems.) James Summers (OS/2 Performance) ********************************************************************** ═══ 4. Paging rate and CPU utilization ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: appended at 15:19:12 on 93/04/26 GMT (by SUMMERS at AUSVM1) James, Thanks for the tips and hints. Looking forward to your findings on relationship between pgin/pgout and CPU utilization. Jackson ********************************************************************** ═══ 5. Compaq Triflex ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We are trying to use SPM/2 on a Compaq SystemPro/XL with the Service Pack applied. The machine stops completely when SPM/2 2.0 is started with reference to a missing device driver OEMHELPЫ. Any suggestions? If this requires an APAR fix - please tell me how to get them (SPM/2 ones are not on OS2CSD) and I can't find the previous OS2SPM20 fora... ********************************************************************** Is it possible that you have installed the SP16MB patch on top of the Service Pak before 2/5/93? This patch had to do with OS/2 not recognizing more than 16M of memory on OEM machines. The first version of that patch caused SPM to break just like you are describing. After 2/5/93, the patch was replaced with another version that SPM will work with. The affected file is OS2LDR in the root directory of the boot drive. SP16MB fix is available on OS2BBS under that name. Pam Lilly ********************************************************************** ═══ 6. Theseus fixes. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is the SPM Theseus fixes needed to the 6.498 level only or to ALL the following (2.1 GA) releases too? can I treat it as CSD ? Geza Szivos TEAM-OS2 NOVACOR Chemicals (403)-290-6583 Fax (403)-290-5900 NHO-32 801-7th Ave Calgary, Alberta, CANADA T2P 2N6 ********************************************************************** No - the Theseus2 patch given out is good for the 6.498 beta and earlier, but it doesn't work on later OS/2 2.1 builds. After OS/2 2.1 ships, we will be able to provide a CSD. Until then, please consider this version a temporary fix, not a formal CSD. Pam Lilly ********************************************************************** ═══ 7. Running SPM/2 from LAN Server? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Can I run the distributed feature of SPM from a server on which I have installed SPM/2 and made it available via an alias? What else besides the would HAVE to be local for the remote monitoring to work? I want as little as possible on the local machines I'm monitoring for maintenance reasons... Thanks, G. Mobley ********************************************************************** You can run the SPM distributed feature off of a network drive, but you have to be careful about violating the license agreement by allowing more users to use SPM than you have licenses. Following is an excerpt from an old append that has been pruned out of the forum that describes how to set up SPM in this way. BTW, you are right that the only file you need on the monitored systems is THESEUS2.SYS. **************************************** How about this for a solution using the existing product? I checked with our Business Practices folks, and they say it doesn't violate the license. Let's suppose that you have 20 developers, but you only want to allow 5 max to use SPM/2 concurrently. - Buy a full copy of the product, plus the 4 additional licenses to you need to support your 5 concurrent users. - Install this code on the server in a subdirectory of it's own, give it an alias and use the LAN Server functions to limit the number of folks that can access this alias concurrently to 5. - Give each of your developers the THESEUS2.SYS device driver, tell them to put it in some subdirectory on their systems, and have them put the following DEVICE statement in each of their config.sys files: DEVICE=X:/path/THESEUS2.SYS where X:path is the path to the device driver on their system. (This is the only device driver ... you're not violating anything here since you can't do anything with this device driver unless you use other parts of the product.) - Have all your developers add the following other things to their config.sys files. (You can figure these out by looking at the config.sys file on a system where you've installed SPM/2 2.0 locally): - The appropriate additions to LIBPATH, PATH, DPATH, HELP, BOOKSHELF statements in order to have access to the appropriate location on the server that has the SPM/2 2.0 code. - The appropriate SET SPMSNAP statement that again points to the location of the code on the server. (This doesn't execute any code, but just sets an environment variable.) - When someone wants to use SPM/2, they need to do the following: - NET USE to the appropriate location on the server (NOTE: they MUST use the alias for which you limited the access ... It's possible that you'd have other aliases that encompasses a larger subset of the drive (or the whole drive), that had no limits on them, and they could get to the code that way. You're on your honor.... but then we ALL are when it comes to copying software, right??) - START SPMNBL /LISTEN - START SPM - As long as all 5 copies haven't been used up yet, they should get access. If 5 copies are already in use, then they have to wait. **************************************** Pam Lilly ********************************************************************** Pam, Thanks for the info...Greg Mobley ********************************************************************** ═══ 8. Tabular report in unit form ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * If a tabular report is generated in unit form (instead of %), what are the "units" for each column? For example, CPU and IRQ services, what does the value represent? Jackson ********************************************************************** Jackson, On a tabular report by unit, the values are reported in the units they are measured in rather than as a percentage. For example, CPU Busy Time is the amount of actual time in seconds that theCPU was busy. IRQ Service Time is the amount of time that the CPU spent in interrupt handling. On a tabular report by percentage, CPU busy % would be the percentage of the sum interval that the CPU was busy, and IRQ Service % would be the percentage of the sum interval that the CPU spent in interrupt handling. So units show the true values of the measurements, while percentage relates the values to the interval of time that was measured. Pam Lilly ********************************************************************** [Reference to append at 11:04 on 93/05/04 by HUIJACK at HKGVM8 ] Ref: appended at 16:57:06 on 93/05/04 GMT (by SPM2 at ASUVM1) Pam, From a report I have generated, the CPU Busy% is 54.84 and the CPU BUSY TIME is 440.9345s. The SUM INTERVAL is 00:11:30. If you do a little bit of arithmetics, you will find some descrepancies. Am I missing something? Jackson ********************************************************************** Jackson, No I don't think you're missing anything. Something sure looks wrong with those numbers. Is it possible that you do not have Service Pak installed on the machine that gave those numbers? If you do have Service Pak (at least), would you please open a problem report by linking to IBMPC and running OS2DFECT? The support folks will help you out with this. Pam Lilly ********************************************************************** ═══ 9. Programming and Reporting ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am working with a vendor who has modified the "Ring 3 Applications" sample code in order to test user metrics. He registered his metric, included the header in the program, recompiled, and performed the monitoring. His report format used the dump, but the report showed zero. The program updates a counter using the SPMUpdateCtr API and the counter is being increased. This has been verified with the debugger. The log file increased in size after the monitoring session. Do you have any ideas why the report is showing zero? From what I can tell, the report format is correct. Any ideas? Let me know if you need any other information. Thanks. David Eyerman ********************************************************************** My guess is that the definition of his metrics in his .SPM file does not match what is being registered on the SPMRegister call. That is the first thing to check. If that doesn't help, try changing the SPM error logging level to information level. In his CONFIG.SYS find the line that says SET SPMSNAP=E D:\SPM2V2\SPMSNAP.ERR (where D:\SPM2V2\SPMSNAP.ERR is his error log path and name). Change the 'E' (Error) following the '=' to an 'I' (Information). Reboot and start SPM and his program. There may be some useful error messages in the error log if SPM had problems collecting his data. Use SPMSNAPP to format the error log so you can read it. Pam Lilly (SPM/2 Development) ********************************************************************** Pam - The vendor tried all the suggestion and still no luck. We cannot determine if his counter and timer data are being transferred from memory into the log file and then into the report. We know the timer and counter are being incremented. Some where along the way, the data is disappearing. The report shows both the timer and counter as zero. Do you have any other ideas? Do you have a working sample program that you could send to me? I do not remember if one was included on the distribution diskettes and I do not have them with me. Is there a utility that will format the logfile into readable text? Do you want to see a program and metric listing? Thanks for all the help. David Eyerman ********************************************************************** [Reference to append at 22:27 on 93/05/05 by SPM2 at AUSVM1 ] Pam - The vendor tried all the suggestion and still no luck. We cannot determine if his counter and timer data are being transferred from memory into the log file and then into the report. We know the timer and counter are being incremented. Some where along the way, the data is disappearing. The report shows both the timer and counter as zero. Do you have any other ideas? Do you have a working sample program that you could send to me? I do not remember if one was included on the distribution diskettes and I do not have them with me. Is there a utility that will format the logfile into readable text? Do you want to see a program and metric listing? Thanks for all the help. David Eyerman ********************************************************************** ═══ 10. Additional Programming Questions ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Some additional programming questions: 1) What is the purpose of the SPM semaphore? 2) How is the SPMaddQword API used? 3) Can a metric file be modified and reinstalled? I am assuming yes. Thanks David Eyerman ********************************************************************** David, 1) The purpose of the SPM semaphore is so that SPM doesn't collect data while it is being updated. If the semaphore is set, the DCF retries 3 times and if it is still set it gives up. Also, if the semaphore changes state while DCF was collecting, it retries. 2) The SPMaddQword is for adding quad words. None of the programming languages know how to add quad words, so this does it for you. 3) Yes, a metric file can be modified and reinstalled. SPMAPPIN is smart enough to replace the old definition with the new one. Pam Lilly & James Summers ********************************************************************** Oops. I asked the wrong question yesterday. I was thinking something else. Instead of asking how the SPMaddQword API is used, I need to know when do you use it. Can you give me an example? Thanks. David Eyerman ********************************************************************** The SPMAddQword and SPMSubQword functions are provided so that you can do your own timing functions. We have been telling people how to use the SPMStartTimer and SPMStopTimer functions to do their own timing. You may want to do your own manipulations of either timers or bulk_counters, both of which are Qwords. Because C generates some truly arcane code when you try to manipulate the Qwords, and some other languages don't have a clue how to do it, we provide these functions to make your life easier. James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** ═══ 11. SPM2V2 Distributed Feature and PerfCap2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * SPM2V2 Distributed Feature and PerfCap2 (LMU/2) We plan to use PerfCap2 (as supplied with LAN Management Utilities/2) to monitor performance and capacity on our distributed OS/2 2.0/LAN Server 3.0 devices. It is our preference to run the PerfCap2 on each server and forward the collected information to the LMU/2 database (as opposed to running one instance of PerfCap2 on the LMU Managing System for each managed device.) To accomplish this, we planned on loading LMU/2 and the SPM2V2 Distributed Feature on each managed server. However, we find that we cannot start PerfCap2 with this configuration... we must load the entire SPM2V2 product in order to get PerfCap2 to function. My questions pertain to the licensing of SPM2V2 and the components included in the Distributed Feature. Since we are not planning to use the SPM2 Control Panel, nor will we using any of the PM functions of the full SPM2V2 product, I would like to see the PerfCap2 function with the Distributed Feature subset of SPM2V2. What pieces are we missing from the Distributed Feature to allow compatibility with PerfCap2? Shouldn't these components be added to the Distributed Feature installation and license? Larry Larry Pisko, TheTravelers ********************************************************************** I spoke to Larry Pisko offline, but I wanted everyone to know what the answer is here. PerfCap2 runs on top of SPM/2 2.0, and it requires the SPM API which is part of the full product but not part of the distributed feature. This is an oversight by SPM/2. The SPM API should have been included in the distributed feature. We will fix this on the CSD which we are working on. I don't have a firm date yet for when we can ship a CSD, but it probably won't be before July 1993. Pam Lilly (SPM/2 Development) ********************************************************************** Is the csd available to fix this problem yet? I have a customer who also requires SPM/2 Distftr to talk to managing LMU/2 using PERFCAP/2. If so, where is it as I've had a look on OS2CSD?? can you send it to me? MPURKIS@HVTVM4. Martin Purkis ********************************************************************** ═══ 12. SPM Fixes. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Do you have new fixes to the 6.514 or I should use the one from CIS what was designed to 6.498 level? Geza Szivos TEAM-OS2 NOVACOR Chemicals (403)-290-6583 Fax (403)-290-5900 NHO-32 801-7th Ave Calgary, Alberta, CANADA T2P 2N6 ********************************************************************** Geza, There is a new version of Theseus2 - version 2.0o, which will run on OS/2 2.1 6.504 and later. The version on CIS now is 2.0i, which will not run on 6.504 and later. I have sent the update to OS2BBS, CIS, and OS2TOOLS and it is just a matter of the time it takes for those files to get refreshed. Please keep checking for the new version. Pam Lilly (SPM/2 Development) ********************************************************************** Thanks. I'll check it on CIS. Geza Szivos TEAM-OS2 NOVACOR Chemicals (403)-290-6583 Fax (403)-290-5900 NHO-32 801-7th Ave Calgary, Alberta, CANADA T2P 2N6 ********************************************************************** ═══ 13. Metric Definition Groups ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is it possible to register multiple metric definition groups in a single file or must they be defined in separate files? If a single file is employed, what are some of the implications? Will different handles be used or needed? When should the groups be registered and deregistered? Any other information on this would be appreciated. Thanks in advance. David Eyerman ********************************************************************** David, Yes, you can define multiple metric definition groups in a single .SPM file. Each group must have a different ordinal - the number you put in the groupord field. Start another group by beginning with the groupnme field and repeating the definitions for the new group. You don't have to repeat the spmversn, os2versn, sftwnme, and prgmnme fields. You should register your metric definition group at the beginning of your program and deregister before your program ends. You should define an exitlist and deregister in the exitlist so that if your program terminates abnormally, your user metrics will still get deregistered. If you don't do this, SPM doesn't get notified that your program is no longer running, and it keeps trying to collect your hooks which may have been stored in memory that no longer exists. This can cause a trap. In the upcoming CSD, we will fix this problem by defining our own exitlist for your process and deregistering for you if you have not already deregistered. But this fix will require you to relink with a new SPMUHx86.LIB, so it is better to do it right yourself now. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 14. Internal Processing Error - Help ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I'm running OS/2 2.0 with the service pak and SPM/2 2.0. While recording and graphing I received: System Detected an internal processing error at location ##0160:fff64c32 - 000d:9c32 60004,9084 04860064 Internal Revision 6.466, 92/10/15 The system is stopped. Record the location..blah blaf woof woff This only happens with SPM/2 running. More information available upon request. Any thoughts????? Dave Ferrey - EDS - Bethlehem Steel ********************************************************************** Dave, Unfortunately, I need to know the "blah blah woof woof" stuff to figure out where the trap occurred. Did you write it down or is it gone forever? Also, it would help me to know the following things: - Is this a standalone machine, or does it have LAN Server or Requester installed and if so, what version? - Were you monitoring your local machine or remote machines? - What resources were you collecting? - Were you recording, graphing, or both? - What exactly were you doing at the time the trap occurred? Pam Lilly (SPM/2 Development) ********************************************************************** Pam, Somehow I missed your response. If by blah blah woof woof you mean error messages, my append listed all! The SPM error log showed nothing unusual. - Is this a standalone machine, or does it have LAN Server or Requester Stand alone, LS&LR installed but not running. -Were you monitoring your local machine or remote machines. local - What resources were you collecting. defaults plus process and threads - Were your recording, graphing or both? both - I noticed no problems (or may not have waiting long enough) recording and graphing seperately. - What exactly were your doing at the time the trap occured. (embarrassingly) training someone. At the exact time, I was talking, I was performing no other function on the computer. Dave Ferrey - EDS - Bethlehem Steel ********************************************************************** Dave, It looks like the trap is in the OS/2 kernel. I can't tell any more than that. Is this something that you can recreate or did it only happen once? Would you please open a problem report on this so our official support folks can work with you to figure this out? Sorry I can't be of immediate assistance, but it doesn't sound like any problems I know about. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 15. Check to see if SPM/2 is running ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am looking for a way to have my application check to see if SPM/2 is running. Is there a named semaphore or queue I can try to open in order to do this? Steve Woolet ********************************************************************** ═══ 16. Where can SPM/2 be found? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is SPM/2 for OS2 2.0 available on any conferencing disks? I found the version of SPM/2 for OS2 1.2 and 1.3 but not for 2.0. Any help is appreciated. Sandra Lucero-Jacquez ********************************************************************** Sandra, The easiest way to get SPM/2 2.0 is to order it through PCWARES. The shrink-wrapped product will be shipped to you. If you need it before it arrives at your doorstep, you can request an electronic copy with the following VM command: REQUEST SPM2A FROM LSYSINFO AT AUSVM1 You will get a package with a form to fill out. The form asks you for the order number from PCWARES. Send the form back per the instructions, and you will receive an electronic copy to hold you over until the shrink-wrapped one arrives. You will find that more and more products are not available on OS2PROD these days. This is the current trend. Please use the LSYSDIST FORUM for any discussion of this distribution method. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 17. Monitoring an Advanced LAN Server 3.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We have been remotely monitoring an Advanced LAN Server 3.0 system (OS/2 2.0 GA + SP base and ES 1.0) for the past month. For the past month, we have also had stability problems (ie. an average of 1 trap per week). I am not sure if this is a coincidence or not. The remote dial-ins from BOCA have not been conclusive as to the cause of the traps. I'd like to eliminate SPM as a possible cause of the instability. Are there any known problems with remotely monitoring an Advanced LAN Server 3.0 system? John Czernuszka ********************************************************************** John, I don't know of any traps when SPM monitors a LS 3.0 Server. We did extensive testing in that environment, but I can't tell from your description if you're doing anything different from what we did. I guess I can't really eliminate SPM from the picture. I can only say that there are no known problems in this area. Have you tried not running SPM for a while to see if you still get the trap? Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 18. SPM and printer resource administration ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Are there some restrictions with Lan Server administration when there is SPM monitoring on on the server? We have got errors when having SPM control running from a client station to a monitored server machine when trying to add printer aliases. The error code added to the lan server logfile is something like OS005. When the monitoring from the client is stopped we can add new printer aliases normally. Any idea ? R. Silvonen, Helsinki ********************************************************************** I don't know why SPM would affect defining printer resources on a LAN Server. I haven't heard anything like this before. Would you mind opening a problem report? Our support people will recreate the problem here so we can figure out what is happening. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 19. Report Summarization by Application ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I cannot seem to get SPM report description set up for Summarization by Application. My application resides on the LAN in path X:\CSP. I would have thought I could choose APPLICATION then NEW, then DRIVE of X:, and that my directories would show up. They do not. The directory shows up as X:\. Why does not panel "SPM/2 DEFINE NEW APPLICATION" provide me with the associated directories when I choose the drive ? I can switch to another window, cd to X:, enter "dir *." and see the CSP directory. I am running on os2.1 6.498. Thanks, Howard Hoyt ********************************************************************** I have tried to narrow this down- it appears I get a correct directory listing it is on my local server. The directory information I need is on a remote server that I am NET USEing into as X: Howard Hoyt ********************************************************************** Howard, I tried to reproduce this, and I couldn't. Let me go through the steps to make sure I did what you are doing. I created a new RDF file, summarized by application. On the application panel, I selected NEW to define a new app. On that panel, I clicked on the dropdown arrow for drive and selected a network drive (F:). Then I clicked on the dropdown arrow for directory (which saidF:\). I then selected a directory from that dropdown list. All EXEs were then displayed in the Available box. Are you doing something different? Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 20. MEMORY UTILIZATION always > 99 % ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * In all measurements with SPM/2 that I have made so far, the MEMORY UTILIZATION has been shown as more than 99 %, even on machines where almost noting is done. Is this reasonable? Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** Erich, Actually, memory utilization of 99.XX % is normal. Memory utilization is the amount of used memory in the system. In OS/2, almost all memory is used. A better indication of the state of your machine with regard to memory is the working set. Working set is the amount of memory that is actually being touched. The closer the working set is to the used memory, the more memory bound your machine is. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 21. Getting comm error trying to use DBM with SPM installed ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I get a comm error trying to use remote data services to connect to an OS/2 DBM server when I have SPM in my config.sys. When I comment out all the config entries for SPM the problem goes away. Anyone know if there is a special order in which SPM and DBM need to be loaded? Anyone else have this problem? Thanks, Dan ********************************************************************** Dan, I haven't heard this problem before, but that doesn't mean it doesn't exist. Would you mind opening a problem report so that our support group can reproduce it and help us figure out what is happening? Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 22. MEMSIZE causes 90 % CPU utilization ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * At a customer's site I measured a test server with SPM/2 2.0 and was surprised to see 100 % CPU utilization for a period of 5 minutes, though there cannot have been much activity. A process summary report revealed CompuServe's MEMSIZE program to consume 90 % CPU. While I did some more tests, I watched MEMSIZE's window; it showed almost no CPU utilization. Obviously it does not measure itself. It looks like MEMSIZE manages to grab all CPU power that is left by the concurrent processes. This appears strange to me. Can anybody comment on this? Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** Is this program a DOS program? If so, and you have background execution on for the DOS session, the program is probably looping waiting for keyboard input. R.C. Morrill ********************************************************************** ═══ 23. cannot get application data using SPMLOGF command ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When I start monitoring using the SPMLOGF command, I do not get any application data. Obviously there is no parameter that would do the equivalent of marking the checkbox "Collect application/process/thread level data". Has this been forgotten or does the parameter exist and is just not described? Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** Erich, There is a way to tell SPMLOGF to collect app/process/thd data, but we didn't make it very easy to figure out. You need to collect ordinal #4 - Threads. That is the same thing as checking the checkbox on the panels. Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 16:01 on 93/05/17 by SCHOBBE at DUESVM2 ] Ref.: appended at 00:15:17 on 93/05/18 GMT (by SPM2 at AUSVM1) Pam, thanks a lot for the really fast response. It works. From all the FORA I have seen so far (and I have seen quite a few ones), this one appears to give the fastest response. Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** ═══ 24. I/Os from LAN Requester on Server not shown on Server ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * For testing I measured a LAN Requester and a LAN Server simultaneously for 5 minutes. MonitorING station was the Reqester with Peer Services, monitorED station the Server. During that time I ran a procedure in the Requester that copied a file from the Server to the Requester's VDISK a few hundred times. The report shows these I/Os in the files section only for the Requester's network drive K: but not for the Server's real drive + path F:ОB43. Other I/Os on the same disk F: of the Server are shown. I consider that an error and reported it to Defect Support. Or is my expectation wrong? Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** Erich, Strangely enough, this is not an error. Let me explain why. I'm assuming that your LAN Server is an Advanced server with HPFS386. In OS/2, there are performance hooks in the HPFS file system that allow SPM to collect file information. When you install LAN Server with HPFS386, the original OS/2 HPFS file system is replaced with HPFS386. There are no performance hooks for file information in HPFS386. So when you copy files from a server to a requester, here is what happens. The copy request causes dosreads & doswrites to be issued to the requester's file system. There are performance hooks in the requester's file system, so the reads and writes are seen by SPM. The reads and writes are sent over the network to the server, but they go to the HPFS386 file system, where there are no performance hooks. So SPM can't see that file activity. So this is not a bug, the information is just not there for us to collect. The workaround is to do just what you are doing - monitor the server's file activity on a redirected drive. Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 08:00 on 93/05/19 by SCHOBBER at STUTVM3 ] Ref.: appended at 14:59:59 on 93/06/01 GMT (by SPM2 at AUSVM1) Thanks again. Yes, we do have HPFS386 on that Server. But your answer triggers immediately 2 new questions: 1. Will there be performance hooks for file information in HPFS386? If yes, when? If no, why not? This appears important to me: consider the case that one out of 50 LAN requesters does many file I/Os to a network drive, so the other users complain about bad response times. But nobody knows which one it is. A good way to find out the sinner would be SPM. Being able to see which application accesses which files may be important. You don't want to measure 50 requesters simultaneously for that purpose, but 1 server alone. 2. In that very same measurement, local I/Os on the Server to the same physical drive, where the network drive is located, are shown by application in the Server report. How is that possible if there are no performance hooks at all for file information? It looks like your explanation needs to be qualified as pertaining only to network file I/Os. Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** [Reference to append at 08:00 on 93/05/19 by SCHOBBER at STUTVM3 ] Ref.: appended at 10:09:52 on 93/06/17 GMT (by SCHOBBE at DUESVM2) Is it possible that my ref. APPEND got forgotten? I am still waiting eagerly for the answers. Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** Erich, We're sorry ... we've been max'ed out with other development of the next release and some other stuff that ended up being "more than high priority." I'll see if I can help find answers this week. Laura Adams ********************************************************************** [Reference to append at 10:52 on 93/06/17 by SCHOBBE at DUESVM2 ] 1. There is currently no way to determine Server resource usage by Requester by monitoring the Server. I agree, that would be very useful. There are several reasons why it's not currently possible: -- With hundreds of Requesters, thousands of files, and millions of requests, it would require substantial Server overhead to keep track of all resource usage by file or by requester. Instead, we chose to aggregate the information. -- The IBM OS/2 LAN Server Advanced product development team has spent years improving LS performance to the point that it is the best in the industry, and they aren't about to put SPM instrumentation in the main path of that mean-lean code. -- Because of the way LS Advanced is designed, it's difficult to track resource usage by requester. Disk usage by requester is not possible. Memory & CPU by requester is not possible because the one LS process handles all Requester requests from a single CPU-eating, memory- eating process. 2. Local I/Os (at the Server) go through the local OS/2 file system, and so are tracked by the OS/2 file metrics. For LS-Advanced, remotely-initiated I/Os go through the remote's file system, then to the lean-mean Ring 0 Server code at the server, circumventing the file system at the Server for performance sake. (LS-Entry works differently; It is a Ring 3 application, so remotely- initiated requests are handled by LS-Entry then go down through the OS/2 file system, just like local I/Os at the Entry Server. By-file stats are therefore available for ALL I/O.) I know these answers are not very satisfactory, but I hope you understand the operation and design better based on my explanation. ********************************************************************** ═══ 25. disk read/write requests <> sum of file read/write requests ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When I take a report of type APPLICATION ( SUMMARY ) and add up the read and write requests of the single files, I expect the number of read and write requests shown for each disk to be equal that sums. But this is not the case with the 2 reports that I just inspected. Am I misinterpreting something? For instance, in the report for SPMT4.LOG (for my reference) I see 226 + 10.7K + 25 + 18 read requests and 69 + 6 write requests. The corresponding DISK 01 shows 1157 read requests and 488 write requests. Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** Erich, I would expect the numbers in the FILE section to be less than the numbers in the DISK section. The reason is that SPM doesn't show file activity for SWAPPER.DAT. The FILE section is counting the number of DosReads & DosWrites for a given file. When pages are read or written to/from SWAPPER.DAT, OS/2 bypasses the DosRead & DosWrite calls and uses a faster method. SWAPPER.DAT is the most referenced file in the system and there are no performance hooks measuring it's FILE activity because the hooks would get executed so often that it would adversely affect the performance of your system. But the DISK section is counting number of reads or writes to the physical disk. It has nothing to do with the file system or DosReads or DosWrites. So any reads or writes to SWAPPER.DAT would show up in this section of the SPM/2 report. Could this explain the differences in the numbers you are seeing? Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 16:13 on 93/05/19 by SCHOBBER at STUTVM3 ] Ref.: appended at 13:49:52 on 93/05/26 GMT (by SPM2 at AUSVM1) Pam, if I set aside those 10.7K read requests, which were reported for fetching the program XCOPY and therefore may play a different role than data requests (?), you are right: both read and write requests reported for the disk were more than the ones in the file section. Is there an explanation for the read requests for program code? Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** Erich, How's this for an explanation of the 10.7K reads for XCOPY. The read & write activity in the file group is counting number of DosRead and DosWrite requests. The read & write activity in the disk group is counting number of physical reads or writes to the disk. In config.sys, you can define a disk cache to improve your system performance. Following is an excerpt from the OS/2 command reference about diskcaches. "The disk cache allows a portion of the system storage to be used as an additional hard disk buffer. DISKCACHE speeds up application programs that read hard disks by keeping hard disk data frequently accessed in a cache buffer. When an application program requests hard disk data that is already in the cache buffer, the disk cache sends the data directly to the application program. This method of accessing data is much faster than if the data had to be read from the disk each time." So when data is loaded into the diskcache, DosReads or DosWrites of that data never access the disk. SPM would show DosReads or DosWrites to that file, but there would not be any corresponding physical disk activity. In your case, XCOPY probably got loaded into the diskcache, so that is why there was no corresponding disk activity for the file activity. Check out the section of the OS/2 command reference for more information on diskcaches. From the OS/2 command line, just type "help diskcache". Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 16:13 on 93/05/19 by SCHOBBER at STUTVM3 ] Ref.: appended at 20:56:39 on 93/05/26 GMT (by SPM2 at AUSVM1) Pam, thanks a lot for your explanations. They fit quite well into the a new test measurement I made. Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** ═══ 26. Happy ending to a strange story ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * One of our customers was having a problem with THESEUS2 trapping. He picked up the update from OS2TOOLS but still had the problem. I could not determine what was wrong, but his persistence paid off. Following is his last note to me (his identity will be kept secret to protect his pride): 'Out of desperation I opened settings for THESEUS2 and almost collapsed: When installing version 2.0O I carefully renamed "old" program to THESEUS2EXE.ORG (I am using HPFS) and copied "new" THESEUS2.EXE into SPM/2 directory. In settings, program section, program name was THESEUS2EXE.ORG. I guess when I renamed old program, WPS changed name of the program in settings too, so I was always running old version!' Moral of this story: Be VERY careful with renaming objects that "belong" to the desk top manager. He tries very diligently to keep things straight, but that is not what we always want. James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** ═══ 27. SPM0025 RC=53 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * . I have SPM/2 V2 distributed feature running on my server and the full package running on one of my requesters (also a server, but I bring it up in requester and/or server mode with the same result). I can monitor my own machine fine, but the server stays in a pending status when I start reporting and/or graphing. Both machines are at 2.0 with 6055 applied. I have done the following: - increased srvpipes to 20 (max) on server - increased sesstimeout to 90 on server and req - increased maxconnections on server - increased maxusers on server - increased numreqbuf on server - granted RWC to GUESTS for \PIPE I can see the server when I query for active stations, but it just sits in a pending status. I have logged on as admin on the local domain and as guest on another domain, no change. SPMNBL /LISTEN is running on both machines. What am I doing wrong? Thanks for your help... Additional info: SPMSNAPP on the requestor is ok but the server shows SPM0025 reason code 53. The 53 means: "Network path not found. The server you specified does not exist or has not been started." Huh????? What network path is it looking for? Thanks! Ken Motta - SE (MCI Team) Colorado Springs (719) 535-1374 (719) 525-0232 pager KMOTTA @ WASVMIC1 ********************************************************************** The server (monitored machine) can't open the pipe that it needs to communiate with the monitoring machine. That's why you're getting the SPM0025 message. The question is why are you getting this error. You said that your monitoring machine is both a server and a peer server? That's not a configuration that we've tried, so it may have something to do with that. When you bring it up as a peer server, do you have user level security or share level security? There are different setup steps depending on the security level. Check out my append in this forum: OS2SPM20 CFORUM appended at 15:16:07 on 93/04/16 GMT (by SPM2 at AUSVM1) Subject: SPM/2 and Peer Services It describes the steps you need to take to setup your peer server. I can't tell you what you are doing wrong, since I don't know how you have set it up. If this doesn't help you, I suggest you open a problem report and let our support people work with you to figure this out. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 28. Theseus/2 Update Traps TRAP D ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I downloaded an update to SPM/2 2.0 from FIXES on IBMLINK which contails new Theseus2.exe, spmws.exe, theseusr.dll. I tried to run theseus2 on IBM PS/2 Model 85 with 48 Meg Memory and it Traps with TRAP D. It runs fine with GA code I am interested in knowing what are the problems in GA code that are fixed on the update. I am using OS/2 2.0 + Service Pack. Is this fix really works with the service pack level or this is the problem with 48 Meg Memory? Thanks Shishir Jain (309)-557-3226 Country Companies Ins. Bloomington, IL ********************************************************************** We will be sending an updated version to OS2TOOLS and the other repositories later today. The updated version works on OS/2 2.0 GA, 2.0 Service Pak, and 2.1 GA (but not the 2.1 betas). James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** I could not find Theseus/2 newest release on "IBM Selective Fixes". The one present there is dated 05-10-93. Is that the latest one? Or, shall I still wait for another release of theseus/2? Does that has fix for the TRAP-D I experienced? Please let me know what problems of the GA Theseus/2 these releases fix. Thanks a lot! Shishir Jain (309)-557-3226 Country Companies Ins. Bloomington, IL ********************************************************************** Shishir, Sorry about the delay. James gave me the Theseus2 update a couple of days ago, but I just now sent it to the person who will update the package on OS2BBS. Hopefully it will be there later today. Something to keep in mind - this version of Theseus2 will no longer run on the OS/2 2.1 betas. It requires OS/2 2.0 + SP or OS/2 2.1 GA. Pam Lilly (SPM/2 Development) ********************************************************************** Pam, Thanks a lot for the Theseus/2 fix. I have a small question regarding this fix. I was getting Memory errors before installing fix. It seems that OS/2 used to fail to DosMemAlloc or DosAllocSeg calls in all the applications when I had 48 Meg of memory and >200 Meg on disk on which SWAPPER.DAT resides. But suddenly and mysteriously this problem vanished with this fix and I was not able to recreate this problem. Do you have any idea if it has anything to do with spmws.exe or any of the theseus/2 files (which are part of the fix). I am curious to find out the source of this problem. Any information would be helpful. Thanks Shishir Jain (309)-557-3226 Country Companies Ins. Bloomington, IL ********************************************************************** The THESEUS2 changes would not have affected errors from the operating system. James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** ═══ 29. SPM/2 documentation - Urgent ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None I urgently need SPM/2 2.0 documentation. Can some kind soul send me a soft copy of the SPM/22.0 manual please?!! Thanks very much in advance! Cheers, Naru Narayanan Perth, Australia 1993/05/25 10:07 EST ********************************************************************** I just sent you the file. You'll need to unterse it with the '(unpack' option. Laura Adams SPM/2 development ********************************************************************** ═══ 30. Apars ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Help - I've got the IP06434, IP06435 and the IP06906 Apars, but two of them contains a new OS/2 Kernel and no information. What's going on here ? Can anyone explain what EXACTLY I need to get an up-to-date bug aand fly free SPM/2 v 2 up and running ? Aloha from Ole Kahler (OLEK at DKIBMVM2) ********************************************************************** Ole, APAR PJ06434 is required to get accurate disk data. The OS/2 file that is affected is OS2DASD.DMD, and the correct version gets installed when you install SPM. You should not have to apply this fix after installing SPM. PJ06435 fixes a trap that may occur while collecting thread or file data. The OS/2 file that is affected is OS2KRNL. Both PJ06434 & PJ06435 are described more fully in the readme file that comes with SPM/2 2.0. PJ06906 also affects OS2KRNL. It is necessary if you are running a virtual device driver that contains SPM performance hooks. Any program that uses SPM performance hooks must register themselves with me, and I can tell you that no one has registered any virtual device drivers to date. So you are safe for now without the fix. It would be up to the owner of the virtual device driver to tell you that you must install PJ06906, so you will know when you need it. So all you need to do is install the OS2KRNL which fixes PJ06435 and you are in business. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 31. any way to set up a report description file by command? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is there any command to set up an .RDF file instead of using the PM interface, in particular to group programs into applications? In addition, is there an easy way to tell SPM that the contents of a particular subdirectory is to make up an application? My customer has a large network of LANs and wants to do this remotely in his central site, because there are no specialists in all branch locations. Erich Schobbert (SCHOBBE AT DUESVM2) ********************************************************************** ═══ 32. Explanation for PERFCAP2 errors. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am using LMU/2 2.0 (CSD 106) PERFCAP2 to gather performance data. There are few errors getting logged in my error log. LMU5630W Node S8256102 - Collection Errors 0x0050 LMU5630W Node S8256102 - Collection Errors 0x0051 LMU5630W Node S8256102 - Collection Errors 0x0011 I am looking for an explanation to these error codes which are returned by SPM. These errors should have been defined in SPM.H but I was not able to find 0x0050 and 0x0051. Thanks Shishir Jain (309)-557-3226 Country Companies Ins. Bloomington, IL ********************************************************************** ═══ 33. Error Message ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Could you tell me what this error message means and what the recommended fix actions are? @ 1993-05-28-10.48.12.38 P:58 T:1 L:16 SPM1041: SPMRPT2: SPMAPI Query Error ->0157:ba40: 09 00 0a 00 ................ JOE ABENDSCHAN *Contractor* (8/347/3209) NRJOEA at BLDVMA ********************************************************************** ═══ 34. Training Offerings ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Anticipating training dates for SPM/2 - Any idea? Estimates available for product license? Thanks G. A. Heiniman @amri055 ********************************************************************** I'm not sure I totally understand your questions ... - If you are looking for training on SPM/2 2.0, there are currently two different courses offered: 1. USING SPM/2 IN OS/2 LAN SERVER ENVIRONMENT (Course G3779) 2. OS/2 LAN SERVICES: PERFORMANCE WORKSHOP (Course # G4760) You'll have to call Skill Dynamics at 1-800-IBM-TEACh (426-8322) for specific course schedules. - What are you asking for in your question: "Estimates available for product license?" Laura Adams SPM/2 Development ********************************************************************** ═══ 35. Processor utilisation ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am a new user to SPM2V2. I ran a monitor and reported by CPU usage by process. The utilisation rate are as follows: Total(top line) 22.07 CM - extended 5.74 LAN Server 0.00 SPM/2 2.0 6.90 unassigned 7.51 The sum of the individual groups is 20.15. Can someone please the dsicrepancy? John Tong ********************************************************************** ═══ 36. Another Pending Forever User ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am attempting to set up remote monitoring. The monitoring station is runing LS 3.0 requester with Peer services. The monitored station is running LS 3.0 adv. srvr. Both machines have OS/2 2.0 with the service pack. The monitored station has only the distributed feature of SPM/2. I have tried to faithfully follow all of the instructions regarding logging the workstations on to the domain, and I have done the NET SHARE IPC$ (I'm using share level security). I feel certain that I have plenty of NetBIOS resource available. Yet I am still unable to establish a remote monitoring session. The monitored station does show up on the QUERY list, but it stays pending/off-line when I start a graphing or recording session. At some point in my testing, i left my desk for about an hour and came back to find an smp0025 error message with rc=231. This makes little sense to me, as I was under the impression that the IPC$ resource could be shared by an unlimited number of users? Can you give me any suggestions on how to determine what's going on? --spk Scott Kantner/Meridian ********************************************************************** Make sure that the userid of the Monitoring machine is defined in the domain to which the Monitored server belongs. I had trouble doing remote monitoring once upon a time, and it seemed to be related to not being in the same domain as the Monitored station. Once I added my userid to the domain of the Monitored station, it fixed the problem. Hope that helps. Jeff Cates ********************************************************************** ═══ 37. Error in syslog ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * SPM2 is putting an error into the syslog with error code 00000139-0c90000 apart from this there is no other indication that anything is wrong. Does anyone know what could be causing this and if ther is a fix? Jonathan ********************************************************************** ═══ 38. APARS for SPM/2 (Theseus) for OS/2 2.1 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Are there any fixes available for SPM/2 and Theseus to make them work with OS/2 2.1? Thanks, David Young - IBM Western Area Systems Center, San Francisco ********************************************************************** You need an update to Theseus2 to make it work with OS/2 2.1. You can get this update from any of the following places: - OS2TOOLS - the THESEUS2 package - OS2BBS - the THESEUS2 package - Compuserve - OS2DF2 forum, Lib 9, file THES2.ZIP Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 15:27 on 93/06/15 by DEYOUNG at LSAVMIC1 ] I have retrieved the Thes2.zip file from compuserve and have installed it on my system. but, The process names on the Process Hierarchy display appear as either blanks or as smily faces, and selecting SYSTEM ->GENERAL SYSTEM ->MODULES from the action bar hangs SPM/2. are there any fixes for these problems? Duane Green - Allnet Communication Services ********************************************************************** Check the version of THESEUS2 (either the product panel during initialization or the Product Information under Help). The version on Compuserve is supposed to be 2.0r. If your version shows 2.0, open the "settings" in the icon and check the path name. If it is wrong, fix it. (If you deleted or renamed THESEUS2.EXE before applying the fixes, the icon path tracks the old version of the file and you have to manually change it in the settings.) If the version shows 2.0r, please re-append and I will try again. (2.0r works on the 2.1 machines in our lab, so it should be OK for you, but then, one never knows. ;-) ) James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** ═══ 39. Monitor .EXE files of monitored machines. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello| I have installed SPM/2 2.0 on a LS 3.0 machine. Actually I monitor both requesters (3.0) and servers (3.0). On each monitored machine I increased NetBios values in IBMLAN.INI as follow : 30,70,16. I am able to collect .EXE files datas of the monitoring machine, and my question is : How could I monitor .EXE files of monitored station Thanks for answer Jean-Pierre Descamps: ETSS / Tool connect. ACL/ABL/SGL Essonnes lines ********************************************************************** ═══ 40. Lost of .LOG files ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When the monitoring is running, .LOG files selected are actives. When the monitoring "crashes" and re-starts, theses .LOG files become unreachable. SPM/2 answers " xxxx.LOG is not a LOG file or is badly corrupted ". Is there any procedure to restore theses files, or are they lost ? Thanks for answer Jean-Pierre Descamps: ETSS / Tool connect. ACL/ABL/SGL Essonnes lines ********************************************************************** I did not see an answer to the above append. Is there a procedure to recover a .LOG file that has been "corrupted" because of a trap or other fatal error? Tom Friends - Account Systems Engineer - Services Co. Boston (T/L) 362-1245 (617) 895-1245 ********************************************************************** In the forthcoming CSD for SPM 2.0, there is a new command called SPMLOGFX to fix up corrupted log files. It fixes log files that are corrupted due to shutdown or power-off during logging. It may work for other types of corruption, too. If you MUST have SPMLOGFX now, send me a note. ********************************************************************** How forthcoming (when) is the CSD? Thanks! Dave ********************************************************************** ═══ 41. 100% CPU shown running DOS program and 0% resident memory ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * My customer is using SPM/2 V2.0 in a LAN Server 2.0 Entry environment. We notice that after we start a DOS 3270 emulation program in a virtual DOS machine, the CPU graph show 100% utilisation even if we are doing nothing on the machine. However, if the DOS emulation is stopped, CPU graph seems to be normal. In addition, the resident memory curve is always 0%. It seems to be unreasonable because at least OS/2 kernal and LAN Server have parts and must be stayed resident in memory. Can anyone explain why it is so? Millie Mak Systems Engineer, IBM Hong Kong. ********************************************************************** Millie, It is normal for a DOS program to take 100% of the CPU. That's just the way DOS programs work. If you want to see resident memory and working set on the RAM graph, you must collect the Working Set resource. You select this when you create your log file. From the create monitor session panel, select the Preferences pulldown, then select the Resources menuitem. You will see a panel with 2 listboxes - one for available resources and one for selected resources. The default resources we collect are CPU, Disk 1, and Memory. Page down in the Available box until you see Working Set, then double click on it to add it to the Selected box. Now start your monitoring session and you should see lines for Resident Memory and Working Set. We didn't include Working Set as a default resource because it is some- what intrusive to collect. Pam Lilly (SPM/2 Development) ********************************************************************** >It is normal for a DOS program to take 100% of the CPU. That's just the >way DOS programs work. Errr.... That's the way a thundering lot of DOS programs work -- but surely not _all_? John W. Kennedy - Hoechst Celanese, Somerville, NJ (The OS/2 Hobbit\Team OS/2) ********************************************************************** ═══ 42. SPM/2 for Education use ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: none I ordered the SPM/2 package from LSYSINFO and in the "ORDER" file it says if you want a copy for demo or education use look in the package file. I looked there and found nothing. Are you trying to tell me something? Where's the information on getting a copy for use in a Skill Dynamics class I'm preparing for OS/2? Thanks, Marty ********************************************************************** Marty, Could you please re-direct your concern to the LSYSDIST forum on IBMPC? We here in development don't have anything to do with this internal distribution of the product -- it's being handled by a totally different group, and they have the LSYSDIST forum to handle any concerns you have. Thanks! Laura Adams ********************************************************************** ═══ 43. Query Terminates SPM/2 Control Panel ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi there, I have been running SPM/2 V2.0 with OS/2 2.0 + SP for a while, and things were going very smoothly. I'm now running OS/2 2.1 and I've added the THESEUS2 fixes but I have been encountering some problems. One of them is using the Query function to determine which workstations have their SPMNBL up. The query seems to run for a while and then the entire SPM/2 control panel disappears with no error messages. I checked the SPMSNAP.ERR and there aren't any messages there either. The query runs successfully if I try it enough times. I only have four workstations on the LAN so I don't think it is a timeout problem. Any ideas? Lauren Wu Systems Integration G5/344/3600/MKM (416) 946-6535 ********************************************************************** I have just completed a similar upgrade to OS/2 2.1 and I am experiencing the exact same problems! I have contacted our IBM person who will look into this for us Dave Moulds Royal Bank of Canada (Toronto) ********************************************************************** ═══ 44. SPM1026 Error when SMP Control Panel Starts ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: none Every time I try to start SPM/2 under OS/2 2.1 GA I get a message box stating "SPM1026: SPMCNTL: SPMGetOrdNames failed". What does this mean? Since SPM/2 behaves very strangely after this message appears I presume that this is not a good error to get. Thanks, Norm ======================================== NOTICE: The entire universe, including this append, may one day collapse back into an infinitely small space. Should another universe later emerge, the validity of this information can not be guaranteed. ======================================== ********************************************************************** There is a file that comes with SPM that is called SPMAPP.INI. From the error you give me, I believe that file has somehow gotten corrupted. I could tell you how to replace that one file, but I don't know that something else hasn't also been corrupted, so I recommend uninstalling and reinstalling SPM/2. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 45. Manuals on SPM/2 APIs ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have a customer who'd like more information on the APIs provided by SPM/2. I can't find anything in the PUBS database and the product appears to ship with minimalist documentation. Please could someone point me in the right direction? The customer wants to use SPM/2 on about 180 servers but would like to add a few bits and pieces of his own. Thanks for any help Andrew Smethurst Workstation System Specialist City of London Business ********************************************************************** Have you looked in the on-line documentation? It seems to describe the API fairly well. James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** [Reference to append at 13:07 on 93/06/24 by 86609212 at EHONE ] James, thanks for your reply re: SPM/2 APIs. The customer would like to know whether there are an REXX based APIs... he does not know how to code in C. Also does PSTAT use any special APIs and if so what are they? If it doesn't use normal API interface then where does it get it's info from and could the customer write an application to do the same type of thing? Thanks Andrew Smethurst Workstation Systems Specialist ********************************************************************** James is out until Thursday (July 1) so I'll answer what I can now. No, there isn't a REXX interface to the SPM/2 API's -- sorry. It's only in 'C'. As for where PSTAT gets it's info... I'm pretty sure it's from somewhere below the published API's. You'll have to wait for James' return for a more complete answer. Laura Adams ********************************************************************** ═══ 46. SPM/2 API support under C SET ++ ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is the new C++ compiler in C SET ++ a supported compiler for application development using the SPM/2 API's? Thanks, Roger Sankey, API Software Ltd ********************************************************************** ═══ 47. NETBIOS resources ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have bumped the NETBIOS counts up quite high and still get SPM errors indicating a lack of NETBIOS resources. Then I went crazy and bumped them up so high NETBEUI.OS2 wouldn't load (couldn't get enough LDT selectors). I have pared back the demo to be just local stuff (although I have to have the LAN Requester running to satisfy some itch about network nodes but no network; when LAN Server is also running I don't have enough NETBIOS resource...) and so am running again. I vaguely recall a tool that figured the various IBMLAN.INI and PROTOCOL.INI resource settings for a given resource need but can't find it. Does any one else know of such a thing? Jim Bennett ********************************************************************** The tool is LAPSDUMP and it is on OS2TOOLS. A couple of things to remember when setting the Netbios resource counts. PROTOCOL.INI defines the total set of resources. IBMLAN.INI defines the resources to be used by the requester and gets its set first. Whatever is left is available to other applications in the box that allocate these resources. This information can be found in \IBMCOM\LANTRAN.LOG. If you have DBM installed with Remote Data Services, it allocates large numbers of these resources. Anyway, LAPSDUMP will help you determine what applications have allocated these resources and how many. Jeff Cates ********************************************************************** ═══ 48. Is SPM/2 2.0 supportedunder OS/2 1.3???? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * hello, I would appreciate if someone could help me in solving this question. SPM/2 2.0 is installed at my customer site in a couple of workstations (one of them with OS/2 2.0 and the other one with OS/2 2.1) I am a new user of SPM/2. The customer wants to know if SPM/2 2.0 can be used with OS/2 1.3 (they have still a lot of OS/21.3 configurations) They know about SPM/2 1.0 and I have also talked them about it. But they want to know if SPM/2 2.0 can be used with OS/2 1.3 (in order to have the same product in all their machines) They are going to change to OS/2 2.0 in the future (so to SPM/2 2.0) in all their machines but know they have OS/2 1.3 in many development workstations. so, can SPM/2 2.0 be installed with OS/2 1.3? Can maybe some of SPM/2 and Theseus2 functions be used with OS/2 1.3? Thank you very much ********************************************************************** SPM/2 2.0 cannot collect data from an OS/2 1.3 machine. THESEUS2 does not run on 1.3, either. SPM/2 1.0 is for OS/2 1.3. SPM/2 2.0 is for OS/2 2.0 & 2.1. James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** [Reference to append at 07:29 on 93/07/01 by 83811246 at SPAVM1 ] No, SPM/2 2.0 does not support OS/2 1.3. The manner in which SPM/2 1.0 and 2.0 get performance information from the underlying OS/2 1.3 and 2.0 operating systems (and the format of the data itself), is totally different between the 2 OS/2 versions, requiring a totally different algorithm for each product. We didn't have the resource to add in a "OS/2 1.3 data handler" to the SPM/2 2.0 product. So the net is: - SPM/2 1.0 only works for OS/2 1.x - SPM/2 2.0 only works for OS/2 2.x Laura Adams ********************************************************************** ═══ 49. User Performance Hooks in .DLL ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Just started using SPM. Doing some testing and evaluation. I am trying to collect data using performance hooks in a .DLL file. The documentation only mentions .EXE files, not .DLLs. 1) Is this supported ? 2) If it is, do I assume where the documentation refers to .EXE, I can substitute .DLL ? PJ Herman - Hursley CATS Code Warriors ********************************************************************** You can use the User Hook stuff from a DLL. The only thing to be aware of is that you need to make sure you DeRegister your DataBlock before the process terminates. (Bad things happen if you don't.) James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** ═══ 50. Very little memory in RAM graph on OEM HW ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We've installed SPM on a OEM machine. When monitoring CPU, DISK and RAM with the graphing function we get only '0.1MB RAM' in the headline of the RAM graph, installed are 12MB. I assume the graph itself is right, it goes up when more load on the system is and down when the load decreases. Thanks for any hints, Harald Kiehle ********************************************************************** Hi there, is nobody interested that SPM/2 V2 will run on OEM hardware? Or am I just in the wrong forum? We have a problem at our customer with that and I'd really apreciate any help. Harald Kiehle ********************************************************************** As I just mentioned in the previous append, we're sorry, but we've been swamped with other stuff and this forum has gotten the short stick. We'll see if we can get some answers this week. Laura Adams ********************************************************************** [Reference to append at 15:02 on 93/07/01 by HAKIEHL at HAMBVM2 ] Is it possible that you are running SPM on OS/2 2.0 GA version? The symptom that you are describing would happen in this case, no matter what hardware you have. SPM/2 2.0 requires at least OS/2 2.0 plus Service Pak. How about it??? Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 51. DT2STR contain NLS ?? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The API supplies several date/time conversion routines which I use extensively. My question is, does the routine DT2STR contain National Language Support? If I pass a FPDATETIME value to DT2STR on a box with country code = Germany, does it return the correct string ??? I believe the answer is no, but wanted verification. Is there a way I can get a correct conversion ?? Jeff Cates ********************************************************************** ═══ 52. SPM and hooks into OS/2... ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am working with a large customer who is planning to install 155 remote OS/2 LAN server machines. They obviously need to be able to monitor performance/capacity of these machines and have been looking at SPM/2 to achieve this. However they feel that SPM/2 does not provide all functionality they require. The customer was wondering whether IBM can provide him with information on the low level hooks that SPM uses into OS/2 to collect it's data so that he could write something to use these and complement the function of SPM. The customer is registered on the Developer Assistance Program in the UK. Thanks for any input. Andrew Smethurst Workstation Specialist - City of Lonodn Business ********************************************************************** Andrew, What exactly does the customer want to do? Perhaps there's another solution than having them tie into the low level hooks. Laura Adams ********************************************************************** [Reference to append at 10:42 on 93/07/07 by 86609212 at EHONE ] Laura, The customer would like to be able to retrieve information about individual processes running on any one of their 200 proposed servers. Currently this would mean loading SPM on each server. What they need is a subset of the SPM function that would enable them to make rapid enquiries as to the processes running. They need to know how much CPU, RAM, threads etc any process may be using. The bottom line is that they can't justify having to pay for a license on every server OR download 2Mb programs everytime they need to make an enquiry. They really need to be able to diagnose problems quickly from the central site...there will be no skill at the remote branch offices. They also would like SPM to provide realtime info on a process level which it does not do at the moment. This would enable them to see what resources a process has used during any given time period. So avoiding having to start the collection facility, create the report, run the report, print the report (at the branch site) and analyse it. They will be using DCAF to access the server from the central site. Thanks for your help. Andrew Smethurst ********************************************************************** These servers would only need the SPM Distributed Feature to be able to extract their data and it is much cheaper than the full SPM product. :plug. Also, LMU/2 has an application that collects SPM/2 data in real-time, monitors for threshold exceptions to generate alerts, helps centralize SPM/2 data and any generated alerts to site management locations, etc. :eplug. Maybe this will help. Jeff Cates ********************************************************************** [Reference to append at 15:48 on 93/07/07 by 86609212 at EHONE ] > The customer would like to be able to retrieve information about > individual processes running on any one of their 200 proposed > servers. Currently this would mean loading SPM on each server. > What they need is a subset of the SPM function that would enable them > make rapid enquiries as to the processes running. They need to know to > how much CPU, RAM, threads etc any process may be using. Andrew, There's a company (Shany, *I think*) that makes a product (AlertView, *I think* - wellspring of knowledge, aren't I) that would do much, if not all, of what your client would like to do. In fact, the client could set resource thresholds down to the application level and AlertView(?) would report exceptions to NetView and/or LAN Network Manager. However, it's not cheap but maybe your large customer could work out a deal. Fwiw... Regards, Lee Butkiewicz - HDR, Inc. (NAAG340 @ HONE8x / USHDR5Y4 @ IBMMAIL) ********************************************************************** ═══ 53. Spmistop problem ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am running SPM/2 2.0 on OS/2 2.0 with SP and collecting data from 5 managed workstations into one log file. Each day I run SPMISTOP.EXE to stop recording, then run a summary report. I issue SPMISTOP /DRAIN filename.LOG After running the report, I copy the log to a backup, then overlay the log with a newly created log for the next days logging. Mostly it works great. Occasionally though, when I try doing the first copy, I get the following error message: SYS0032: The process cannot access the file because it is being used by another process. I then run SPMISTOP.EXE again and get rc=0, but still can't copy the file. Only after issuing SPMISTOP /HALT is the log file free. Any suggestions ? ********************************************************************** ═══ 54. Documentationlem ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have a copy of SPM/2 2.0 but need some documentation. Can someone please tell me how to get/request a copy of the docs? Thanks. Message from: Jorge Nazario 8/773-7464 Lexington KY (606) 276-7464 ********************************************************************** Here's what you do to get a LIST3820 of the on-line reference: REQUEST SPM20DOC FROM LSYSINFO AT AUSVM1 By the way, this same question was answered earlier in the following append: --- OS2SPM20 CFORUM appended at 16:40:02 on 93/02/18 GMT (by SPM2 at... Subject: Update to SPM20REF LIST3820 document Laura Adams ********************************************************************** ═══ 55. Reporting problems with SPM/2 V2.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * What is the correct procedure for a customer to open a problem report with IBM on the SPM/2 V2.0 product? Can we use the support function found here in the OS2BBS, the ETR function in IBMLINK, COMPUSERV. We've looked arround but everything we find looks specific to OS/2 base. Thanks, Roger Sankey, API Software Ltd ********************************************************************** I believe you should be able to use the defect support in OS2BBS. I've never seen it, but I'm told there is support there. If you have problems with that, call the defect support number, 1-800-237-5511. Laura Adams ********************************************************************** Thank you, I'll give that a try. Roger Sankey, API Software Ltd ********************************************************************** ═══ 56. Information on SPM/2? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Where can I get some information on the functions that SPM/2 can perform? We may be interested in the product, but want to read up on its abilities. ********************************************************************** There are two different options: - The spec sheet (order no. G221-3496-00) - The white paper If you're internal to IBM, both of these can be obtained from the MKTTOOLS disk: - spec sheet is in G2213496 PACKAGE - white paper is in LSPAPERS PACKAGE From your userid, it looks like you're external to IBM?? In that case: - I know that the spec sheet can be faxed to you by calling 1-800-IBM-4FAX, and requesting the order number (G221-3496) - As for the white paper, I'd have to send it to you or fax it to you. Let me know. Laura Adams ********************************************************************** Laura, please FAX us a copy of your white paper to: 214-991-1214 SPM/2 is a handy product - it has always bothered me that the best documentation I was able to find on it came out of a third-party book. I'd be interested in information detailing planned enhancements as well. Thanks in advance! Dave Scheffer, ARGO Data Resource Corp, (214) 386-4949 ********************************************************************** Dave, I'll fax you that info. You mentioned a 3rd party book that describes SPM/2... what is that?? (Sorry for being so ignorant!) Laura Adams ********************************************************************** ═══ 57. Lan Network Manager and SPM/2 V2.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Did anyone has both lannetwork manager and spm/2 running on the same machine? It you did, what are the things I need to know before I install both code. My customer has a network with 13 bridges and 15 servers and would like to use lannetwork manager and spm/2 to manage the availability and performance of the network. He may put in a dual backbone so the number of bridges may double. Tai at NYCVMIC1(Chris L. Tai) ********************************************************************** Yes, LNM 1.1 and SPM/2 work fine on the same machine, just install as the products suggest. Many customers use SPM/2 to watch the CPU utilization on their LNM machine. Append LNM questions to LNM10 FORUM on IBMPC, look at LNM11CPU package on MKTTOOLS, and/or contact off line for other questions. Dave LeGrys LNM Performance ********************************************************************** Dave, what information is available to customers like ourselves? Dave Scheffer, ARGO Data Resource Corp, (214) 386-4949 ********************************************************************** I have not written a redbook, or other external document, if this is your question. Working with you and your SE, I can help you with capacity questions about LNM, select the best PS/2 for your network as a LNM host, and help keep you up to date on product maintenance. With the exception of SPM/2, the analysis tools are IBM internal, so I need to work with your SE. Help is provided on a individual basis. Dave LeGrys LNM Performance ********************************************************************** ═══ 58. Can not monitor a DOMAIN Controller ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I can monitor remotely an aditional server, but when I try to monitor The Domain Controller y always go to the off-line status. Putting 'I' in the SPMSNAP line in Config.sys and running SPMSNAPP after trying a remote Monitoring, I get: (ES20250B is the monitoring system) @ 1993-07-20-16.44.16.12 P:9 T:1 L:128 SPMNBL: DCF ES202E061 (PID 69) started on behalf of ES20250BA @ 1993-07-20-16.44.16.59 P:9 T:1 L:128 SPMNBL: DCF started. @ 1993-07-20-16.44.34.78 P:69 T:4 L:8 SPM0387: SPMDCF: Warning OS/2 kernel message queue not active Does Anybody know what I have wrong? Regards Luis Morras ********************************************************************** ═══ 59. Questions from a new spmv2 user... ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello SPM team, first of all I found this product very impressive and useful to "understand" how OS/2 V2 is working | I used SPM/2 V2.0 + theseus (from os2tools) on OS/2 V2.1 514M. My customer developped an application with Cobol microfocus and the Dialog System (from microfocus) and we want to see ressources used by this application. I run lots of scenario with differents reports.... and I have some questions to understand what happens: 1/Is spm able to "see" the printer figures if the spooler is desabled ? In my case the spooler is OFF and the figures seems to be not at all good. 2/What is the description of the following items on a report: Dispatch (Threads level) % util and Rate (Interrupts level) 3/In the same conditions I use SPM and THESEUS to see the needs of the customer application in terms of RAM. In SPM, I have found for instance 5600 in Dispatch value for the application. In Theseus, the working set is 2000 for .EXE and 1800 for system .DLL. Is there any relation between these values ? Sorry for these novice questions| My last one is : IS THERE ANY DOCUMENTATION WHICH IS DESCRIBING ALL THE FIELDS OF A SPM REPORT ? (The SPM documentation helps to install and run SPM but not to describe the fields...) Thanks a lot for any help. Best regards Philippe (France) ********************************************************************** ═══ 60. Netbios requirement for managed workstation ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I need to install the distributed code on my customer's server and would like to find out what the netbios requirements are. The server is pretty much max out in resources so I want to be very careful on this. The server is running LS 3.0. I read through the documentation and it only mentions the requirement for managing machine. ********************************************************************** Check out the header of this forum - there is a description of the Netbios resources needed by SPM at the bottom of the header. Pam Lilly (SPM/2 Development) ********************************************************************** Sorry, I gave you an incomplete answer. The header of this forum lists netbios requirements by component, but not by managed or managing system. The only component that runs on the managed system and uses Netbios is SPMNBL. So the Netbios requirements for a managed system is 1 Name, 2 NCBs, and 2 sessions. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 61. Append from SPM2 forum on remote monitoring setup ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The following append from the SPM2 forum is being moved and answered here: "I have installed SPM/2 v2.0 under OS/2 2.1. I used the /DISTFTR option when installing on a server and did a full SPM install on my network manager console. I then defined a montior description to monitor my server. When I try to monitor the server SPM shows the server as off-line. What am I forgetting to do? The online help has not been to useful so far... " The first part of the answer was already stated in the SPM2 forum and is repeated here: "Is the system that's the "managing system" just a requester? It must either be a peer server or a full server. Please see the following appends in this OS2SPM20 cforum for info on getting remote monitoring up and working: - Pre-reqs and steps needed: append "16:42:09 on 93/03/30" entitled "Remote Monitoring doc confusing" - Correction to Peer Services setup: append "15:16:07 on 93/04/16" entitled "SPM/2 and Peer Services" " The other part of the answer that I didn't think about were fixes needed for running on OS/2 2.1. They aren't specifically related to getting remote monitoring to work, but you'd probably be interested in them: - Theseus2 fix for initial screen to come up correctly. Fix available in THESEUS2 package on OS2TOOLS. - Fix for Query function trapping on OS/2 2.1 Fix available by sending a note to me: LSADAMS @ AUSVM1 - Installation onto OS/2 2.1 system will result in copying a back-level version of OS2DASD.DMD into the \OS2 subdirectory. (SPM install was only supposed to copy this file on the system if it was OS/2 2.0 + the first Service Pak or earlier. However, we goofed and always copy it.) During the copy, we rename the original one to OS2DASD.OLD. The fix is to copy OS2DASD.OLD back to OS2DASD.DMD. Laura Adams ********************************************************************** ═══ 62. LOG File Size Limitations ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Are there any limitations to the size of a LOG file? The customer that I have been working with created a LOG file that reached 45MB by monitoring at the thread level 2 clients and the LAN Server 3.0 system (OS/2 2.1) where SPM was running. Then, the Server got an access violation in DOSCALLS1 that caused SPM to terminate. John Czernuszka ********************************************************************** No, there are no practical limits (to our current knowledge!) on the log file size. It's just an OS/2 file, and we didn't place any restrictions on it. Sounds like you might have a bug of some sort. I suggest that you open a defect. You can do this internally by linking to IBMPC and typing OS2DFECT and then filling in the blanks. Laura Adams ********************************************************************** Is there any way in SPM to put restrictions on the size of the LOG file? What if the recording process is kicked off on a machine and "forgotten" for over a week. If the file continues to grow to the point that there is no available space on the disk, that machine could be brought to its knees. What process is there to prevent this from happening? ********************************************************************** There is no way to restrict the size of the SPM 2.0 log file. There also is no function in SPM to schedule or automatically start and stop logging, though that could be accomplished with REXX command files. The rate of growth of the log file CAN be controlled by varying such parameters as number of workstations, recording frequency, number of resources, and type of resources. By the way, the LAN NetView Monitor product has more automation features for collecting performance data, including 24-hour 7-day scheduling, min/max log file sizes, and automatic daily/weekly/monthly processing, if these functions are important to you. ********************************************************************** ═══ 63. SPM/2 1.x on non-IBM hardware ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I know that this forum deals with SPM/2 2.0, but I need an answer from IBM regarding SPM/2 1.x for OS/2 1.3. I realize there have to be hooks in the disk device driver that SPM/2 must be able to use in order to report performance data. The question is whether only IBM drivers (DISK01.SYS, DISK02.SYS,...) will have these hooks & be recognized by SPM/2, or if it is likely that drivers from other vendors, such as DPT (for the PM2012B EISA SCSI card for example), Adaptec, et cetera could have been programmed with these hooks and therefore function. Does IBM provide the information necessary for those controllers to work with SPM/2? Thanks, Tom Grover Dell Computer Int'l Product Support ********************************************************************** No problem about asking the SPM/2 1.0 question here... since you're external, there's no way for you to get to the 1.0 forum. It's highly unlikely that an OEM device driver would have hooks in it. We really haven't had many folks contact us, other than a couple of IBM groups that were putting out new device drivers. Someone interested in instrumenting a device driver would have to contact us somehow (via this forum would be a start!) to get the information necessary to put hooks in the device driver. Laura Adams ********************************************************************** Laura, Thanks a bunch for the reply. I had already decided from looking at the situation that it was highly unlikely that such a 3rd party driver would exist. I don't intend to pursue it further, since I seriously doubt we have a need that would justify the development effort by the vendors that would be involved. Thanks again, Tom Grover ********************************************************************** ═══ 64. Beginner's questions on SPM/2 2.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None Hello, One of our large corporate customers has installed SPM/2 2.0. They are not thorougly familiar with the product, so, here are a few questions: 1. Is there any documentation about SPM/2 user experiences such as how to to use the monitor data from OS/2 LAN Servers? 2. Are there any fixes for the various bugs in the report area? 3. Why is data like HPFS386 only available in the DUMP report? 4. Will the use of LMU/2 give better reporting capability as to collecting SPM/2 performance data? Obviously somewhat novice questions, but the customer apparently has not fully understood the documentation and I have not got access to the product and its manuals. Thanks for any help... Carl-Magnus Wiberg, PC Competence Center, Sweden ********************************************************************** ═══ 65. SPM/2 2.0 Education Materials ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello, I'll be doing a service to a customer and teach them how to use SPM/2 2.0, can some one be kind enough to send me the education material? Appreciate your help. Manson Yang ********************************************************************** ═══ 66. Dump report ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I had sucessfully run both summary reports and tables. But I was unable to run the dump report. Are there anything different I need to do? When I run the dump report, the window pop up with message saying that it is running but it never finish.The disk light blinked for a while and stopped. ********************************************************************** ═══ 67. SPM2 data collection ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * From SPM2 documentation I see that data collection can be done on workstation level or application/process/thread level. The last one is not recommended because it generates huge data file and creates CPU overhead. Is there a way to collect data on application only level or process only level to avoid the above problems? The workstation level is too general. The summary report shows total CPU utilization only. Thanks. Alex Zlatsin (ZLATSIN@WATSON) ********************************************************************** The way that SPM generates the Application and Process level data is by calculating it from the Thread level data. OS/2 gives us the system totals and the thread level data, but nothing in between. Are you leery of collecting thread level data because of the warning we gave? Or have you tried it and found it to be too intrusive? SPM's intrusiveness will vary depending on how many resources you collect and how frequently you collect. You may be interpreting our caution statement too strongly. It is ok to collect thread level data, but you should use this option when you are trying to narrow down a problem, rather than trying to determine where the problem is. For example, if the system totals in the workstation summary report show that you have high CPU utilization, you would want to collect thread data and run a application or process summary so you could see where your CPU is being used. The caution statement is just to warn you that the detailed information costs a little extra in performance. But it is not a recommendation against using this very useful feature. Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 19:20 on 93/08/04 by ZLATSIN at YKTVMV ] :plug. PERFCAP2 in the LMU/2 package can be used to collect SPM data on one or more specific processes by name. There is no magic - it collects the thread information related to the process and at specific time intervals (determined by the user), analyzes the data or sends it to a database. The data collected during that time period is kept in memory and does not generate a large SPM log file. The same discussion about intrusiveness still applies. :eplug. Jeff Cates ********************************************************************** ═══ 68. Going from Process ID to Process Name ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is there an API call where you can give the process id and get back the process name? Alex ZLatsin ********************************************************************** ═══ 69. Workstation monitoring with Novell Netware ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I read a posting here from February about using SPM/2 2.0 in a Netware environment. It stated had been no testing done with monitoring a Netware server. My question is can a remote workstation, NOT the server, be monitored from another workstation running SPM/2 2.0 where both are running Novell Netware's Netbios? Or can the Named-Pipes be used? And if not now, is IBM planning to introduce compatibility with Novell Netware for SPM/2 2.x? Thanks! Craig Bugg -- United Parcel Service ********************************************************************** I can't speak for the future of SPM/2, but you can use LMU/2 (and the SPM/2 Dist Feat) to collect performance data on a Netware requester today. LMU/2 allows you to collect SPM/2 data and bundle it up for shipment and placement into a DBM database and/or monitor the data for threshold exceptions and generate alerts and/or create summary analysis reports on a time interval. LMU/2 does not provide the data in PM graph form. The collection of data by LMU/2 on a Netware requester is a 'local' collection only. Also, LMU/2 provides applications to collect Netware server performance data (this function does not involve SPM/2). Jeff Cates ********************************************************************** ═══ 70. Query Stations under OS/2 2.1 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When I am setting up a log and press the Query button to find all the stations on the LAN, SPM Control terminates. This was documented as a bug during the OS/2 2.1 beta with a work around of manually entering in the names of the workstations. Is there a fix available for this problem ? Kyle Petersen (ProTools Inc.) ********************************************************************** Did you ever get a response to this append ? I am having the same problem with spm/2. thanks, Per Karlsen ISSC Intergrated Systems Solutions Corporation Dept. CCKA / B671 / Rm LAa02 Regency Park Cary N.C. Cary Tie Line 883-7145 Cary Outside Line 919-469-7145 From: PER at CARY ********************************************************************** Yes, I have the fix for it. I'll send the file to you: SPM EXEBIN. We've finally been freed up from LAN NetView Monitor development (announced yesterday ... see HONE announcement 293-538), so now we have some time to focus on the CSD for SPM/2 2.0. We're starting some testing of fixes next week. It will probably be several weeks before we can get it all pulled together. Laura Adams ********************************************************************** I would like to participate in beta testing of spm/2. If you need assistance please let me know. My team has approx 70 servers in RTP and Cary that we own and approx 150 throughout the Southeast Region. We are currently evaluating SPM/2 as our server monitoring tool. thanks, Per Karlsen ISSC Intergrated Systems Solutions Corporation Dept. CCKA / B671 / Rm LAa02 Regency Park Cary N.C. Cary Tie Line 883-7145 Cary Outside Line 919-469-7145 From: PER at CARY ********************************************************************** [Reference to append at 15:43 on 93/09/30 by LSADAMS at AUSVM1 ] Please send the file to PER AT CARVM3. CARY is my tcpip node. thanks, Per Karlsen ISSC Intergrated Systems Solutions Corporation Dept. CCKA / B671 / Rm LAa02 Regency Park Cary N.C. Cary Tie Line 883-7145 Cary Outside Line 919-469-7145 From: PER at CARY ********************************************************************** [Reference to append at 15:43 on 93/09/30 by LSADAMS at AUSVM1 ] Thanks for the file. It did fix the query problem I was having. fyi. ********************************************************************** ═══ 71. SPM20 docs ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I downloaded the SPM2DOCS file but could not have VM successfully UNPACK the .PAC file. I uploaded the .PAC file to CMS using PCOM30/DOS's "SEND" command w/ no parameters. Is there a special way with which I should upload the .PAC file? Can a LP3820 version of the docs be made available .. pretty please? Lawrence Liu - World Savings ********************************************************************** ═══ 72. CPU Graph problem. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * My CPU graph stays at 100%. When I close it and reopen it, it starts at 0% then immediately jumps to 100% and stays there. I'm running OS/2 2.1, SPM/2 v2.0GA w/ the THESEUS2 updates from OS2tools. Can someone help me understand what the problem is and how to solve it? Thanks. David A. Bringley 313-337-1551 835 Mason, Suite 350 DearBorn, MI, 48124 ********************************************************************** Record the data, then run a report. I think you will find you have a CPU hog running somewhere. Pulse does this and some DOS apps. Also, the 'e' editor can sometimes be opened detached and this does it. James Summers (OS/2 Performance) ********************************************************************** [Reference to append at 19:17 on 93/08/17 by DBRINGLE at DETVMIC1 ] Ref: Append at 13:57:58 on 93/08/18 GMT (by SUMMERS at AUSVM1) James, You were right. Unknown to me someone install a utility that ran detached on the idle thread. Thanks for the tip. David A. Bringley 313-337-1551 835 Mason, Suite 350 DearBorn, MI, 48124 ********************************************************************** ═══ 73. Getting SMP/2 2.0 To Run On v2.1 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have SPM 2.0 running on several OS/2 v2.0 systems. I have been unable to make any progress getting it to run on 2.1. All attempts result in a complete system hang requiring a cold reboot. Any suggestions? I see from other posts here that others are successful where I have not been. Stan Smith ********************************************************************** If you are not using THESEUS2, then it should work on OS/2 2.1. For THESEUS2, you need the fix that is on CompuServe or OS2BBS. Just unzip it over the top of the existing files. James Summers (OS/2 Performance) ********************************************************************** I'm experiencing exactly the same problem on an IBM PS/2 90 (486DX 50/25) with 28MB and OS 2.1 GA US installed. Recording the system might hang, and the only way to recover is to do a Power-On-Reset. Niether Alt-Ctrl-Delete or Aly-Ctrl-NumLock works. Upon POR the LOG-file is reported "badly corrupted". Is there any way to recover a damaged LOG file ?? Theseus2 is NOT active. Morten M║ller ********************************************************************** ═══ 74. Remote monitoring. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Can someone pls help me out, I've been trying to get remote monitoring to work for some time w/ no success. The monitorING ps/2 has 2.1, and LS3.0. The monitorED ps/2 has 2.1 and LAN Req 3.0. Both are loggged on. I can monitor locally ok. The status remains 'pending' when I try to remote monitor. I've read all the appends, searched through the manuals, adjusted protocol.ini and ibmlan.ini. What could be wrong? Thanks.... David A. Bringley 313-337-1551 835 Mason, Suite 350 DearBorn, MI, 48124 ********************************************************************** Did you get a response to this append ? I have just setup spm on a workstation running os2 2.1 and lan server 3.0. I get the pending status forever. fyi. Per Karlsen ISSC Intergrated Systems Solutions Corporation Dept. CCKA / B671 / Rm LAa02 Regency Park Cary N.C. Cary Tie Line 883-7145 Cary Outside Line 919-469-7145 From: PER at CARY ********************************************************************** Per, At set of instructions for getting remote monitoring going was put in the following append: " ----- OS2SPM20 CFORUM appended at 16:42:09 on 93/03/30 GMT " " Subject: Remote Monitoring doc confusing " Look to see if you've done everything listed there. If this still doesn't work, come back and we'll go from there. Laura Adams ********************************************************************** ═══ 75. SPM0529 error. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * After I install LAN Req 3.0, SPM will nolonger run. I get the above error when I start the graph and monitoring functions locally and the monitoring stays in 'pending' status. My ps/2 has OS/2 2.0 xr06055 GRE xr02100 LAPS wr06000 LAN Req 3.0 ip07000 Any suggestions....... David A. Bringley 313-337-1551 835 Mason, Suite 350 DearBorn, MI, 48124 ********************************************************************** ═══ 76. Two appends. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: Append at 11:02:51 on 93/06/28 GMT (by 86609212 at EHONE) I have solved the two problems that I posted. Regarding the SPM0529 error the spm monitorED ps/2 need more ncb, sessions, and names. All seems to be working now. Thanks... David A. Bringley 313-337-1551 835 Mason, Suite 350 DearBorn, MI, 48124 ********************************************************************** ═══ 77. Can LS Peer Server monitor a Lan Requestor 2.0?? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Can a MonitorING PS/2 with Lan Server 3.0 Peer Services installed remotely monitor a OS/2 2.0 w/servicePAK + Lan Requestor 2.0? Message from: Gary R. Martin - Advisory SE (614) 225-3761 T/657 ............: 140 E. Town Street - 12th floor (Dept L58) ............: Columbus, OH. 43215-5195 ********************************************************************** ═══ 78. SPM2V2 Reports ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have questions on the SPM2V2 Tabular report format, specifically when using the "spreadsheet" format with percentages: Memory PageIn/Out % always adds to 100%, but without some indication as to absolute numbers of pages faulted, the percentages on this report are little value to me. A cumbersome workaround I've used is to re-run the report with resource data represented in units, then cut & paste the unit paging numbers into my percentage report. IS there some way to get paging quantified in the percentage report??? Second, with SPM version 1, the delimited report showed us the CPU utilization associated with each PID for each interval. Can this be added to the SPM/2 Version 2 reports? Thanks, Larry Larry Pisko, TheTravelers ********************************************************************** I have questions on the SPM2V2 Tabular report format, specifically when using the "spreadsheet" format with percentages: Memory PageIn/Out % always adds to 100%, but without some indication as to absolute numbers of pages faulted, the percentages on this report are little value to me. A cumbersome workaround I've used is to re-run the report with resource data represented in units, then cut & paste the unit paging numbers into my percentage report. IS there some way to get paging quantified in the percentage report??? Second, with SPM version 1, the delimited report showed us the CPU utilization associated with each PID for each interval. Can this be added to the SPM/2 Version 2 reports? Thanks, Larry Larry Pisko, TheTravelers ********************************************************************** 1) Most folks don't know how to interpret paging rates (myself included) but we wanted to provide some percentage to be consistent with the rest of the percentage report. Page in & out as a percent of total pages was all we could think to do. One workaround is to do a Tab report with Units, Delimited and import the result into a spreadsheet template that you've set up. Then you can calculate any statistic on paging you want in the spreadsheet. 2) CPU utilization by PID can be obtained by collecting Thread data; check the box for Appl/Process/Thread level data in Resources. Then do a report with Format = Summarization and Summarization Level = Process. Unfortunately, the Summarization report is not available in the Delimited format, if that's what you're after. ********************************************************************** ═══ 79. Multiple T/R Adapters ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I'm trying to monitor a Lan Server 3.0 with three T/R adapters from a Lan Requester 3.0 with Peer services and get pending or off-line for the Lan Server. I can not find any information on monitoring a Server with multiple adapters. SPM works fine with a Lan Server and workstations with one adapter. Can a workststion with multiple adapters me monitored? If so, how do I make it happen? Thanks in advance. ********************************************************************** ═══ 80. Monitoring a remote station ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello, our customer has SPM/2 2.0 installed in a LAN environment with LAN Server 3.0 Entry, OS/2 2.1, but not at the server. He wants to monitor with a separate station. We can start the monitor session with one remote station, but it stays in the status pending and the status of the station is offline. The LAN Requester is started at both stations, SPMNBL is active, the Distributed Feature is installed at the remote station. What can we do to change the status? Best regards, Odile M. Ebbers ********************************************************************** Hello, i think i have to verify that the Peer Services are installed correctly and i have to look for the Netbios parameters. Are the maximum values the ones i have to use? Best regards, Odile M. Ebbers ********************************************************************** ═══ 81. Offline message ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We have the following environment: Monitored workstation has OS2 2.1 with LR 2.0 and the distributed fea- ture of SPM2. The monitoring workstation has OS2 2.1 with LR 3.0 and peer services. It has the full SPM2 code. A query shows the monitored workstation name but when we try to collect the data, the monitored workstation shows as offline. We have been struggling with this for a while. Any advice? Both workstations log on to a domain controller running OS2 1.3 and LAN Server 1.3. Regards, Andrew P. Greco APGRECO at NYCVMIC1 ********************************************************************** There are several setup things that you need to be aware of, especially when MonitorING from a Peer Server. These things were addressed in the following earlier appends to this Cforum(which I also sent to you): -OS2SPM20 CFORUM appended at 16:42:09 on 93/03/30 GMT (by SPM2 at AUSVM1) Subject: Remote Monitoring doc confusing Ref: Append at 15:45:20 on 93/03/29 GMT (by CHVY002 at OS2CUST) -OS2SPM20 CFORUM appended at 15:16:07 on 93/04/16 GMT (by SPM2 at AUSVM1) Subject: SPM/2 and Peer Services Let me know if this does not solve the problem. Laura Adams ********************************************************************** ═══ 82. Monitoring a remote station / Creating a report ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello, our installation works, we haven't installed Peer Services correctly| Sometimes we have a problem by creating a report. One example: We monitor a station, the user of this station makes a restart. Our monitor session gets the status pending because the workstation is offline and we decide to stopp the monitor session. Creating the report doesn't work| The error is SPM2014. What can we do to use this LOG-file? What is the problem? Best regards, Odile M. Ebbers ********************************************************************** ═══ 83. Swapper analysis from THESEUS2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The number of disk frames in SWAPPER.DAT reported "used" by THESEUS2 is incorrect. The number reported is (probably) lower than what is actually in use. This will be corrected in the next version of THESEUS2 distributed. James Summers (OS/2 Performance) ********************************************************************** ═══ 84. Mandatory kernel fix ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The fix for APAR PJ09893 is MANDATORY on any machine where data is being collected by SPM/2 2.0. We have seen 2 traps on customer machines in the last 2 days caused by the problem. "Level 2" support in Boca can provide the fix. This requirement applies to both OS/2 2.0 (both ServicePaks) and OS/2 2.1. James Summers (OS/2 Performance and SPM/2 development) ********************************************************************** Can you be a bit more specific about this fix and problem? Are there any particular situations that cause the crash, or can it happen at any time on any system? I am a bit wary of putting individual fixes on a system that is otherwise working smoothly - would prefer to wait until a CSD if possible. Thanks, Dave ********************************************************************** [Reference to append at 18:42 on 93/09/03 by SUMMERS at AUSVM1 ] I called Level 2 support in Boca to get this fix. I was told that it was for HPFS partitions only (I am currently using FAT). I was also told that it was a fix for the Netview Monitoring product. So, it is not clear to me that I need this particular fix, although, since I have had a couple of traps (usually trap 0002) in the last few days while remotely monitoring an HPFS workstation running OS/2 2.1 GA, maybe I do. Could you be more specific as to what kind of traps this fix is supposed to fix, whether it is for HPFS or FAT or both, and whether it is for a Netview product or SPM/2 2.0 ? Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** Trap 2 is generally a hardware failure, typically memory. If these are occuring on PS/2 systems, try getting the SADA diskette to see whether hardware errors have been recored by the PS/2. Dave ********************************************************************** [Reference to append at 19:02 on 93/09/13 by DENNIS at RALVM0 ] I believe that the kernel fix was generated since the bug was found while testing LAN NetView Monitor. However, SPM/2 2.0 accesses the same hooks in the Kernel, so the fix also applies to SPM/2 2.0. Who did you talk to? As for what does it specifically fix? James Summers is out this week so I don't know the detail, but I recall him telling me that it was primarily needed for HPFS, but also for some file and thread situations. So, he just figured everyone would probably need it (most folks collect process data sometime!). We'll have to wait for James to return to get more specific detail. So, later! Laura Adams ********************************************************************** I was able to talk to James Summers and get more detail on the "PJ09893" fix. He said that although the fix was generated in response to an HPFS problem, it actually ended up fixing several other problems that we'd been seeing for about a year, and couldn't pin down. As best I understand, here's what's involved: when SPM (or LAN NetView Monitor) initializes for data collection, there are some performance hook registration calls that are made (to the OS/2 kernel). There was pointer that was being set up wrong in the registration control block. This particular scenario applies for HPFS, File and Thread (which includes Process) data. Since James felt that most folks would collect one of these at some time, he said the fix should be mandatory. Laura Adams ********************************************************************** I applied this mandatory kernel fix (which contained a fix for several APARs including PJ09893; OS2KRNL dated 8-30-93). I have not encountered a trap 000D or 0002 as yet, but, some of the other applications I use to monitor memory availability (such as OS20MEMU, etc.) no longer work ! I have reported the problem to the OS2MEMU FORUM and called the developer of the other application. Since THESEUS2 still works, it must be getting its "memory available" information from a different place than these "other" applications. Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** ═══ 85. SPM2V2 overhead ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * What is the expected/typical overhead in monitoring (graphing only) my 'local' system. I installed SPM20 and am getting noticeable (about 1 second) delay in keyboard response on a period that seems consistent with the sampling rate. Even, for example, typeimatic repeats to this 3270 session show noticeable, periodic delays, Is this normal or what did I do wrong? Environment is spm 2.0 (GA version), 8580-M311 (20Mhz), 16MB RAM, OS/2 2.1 (GA + 21WPSF), LR 3.0, ES 1.0 CM, etc, etc Sam ********************************************************************** Sam, I think this is pretty normal... the overhead is definitely going to be there for the updating of the graph. What is your collection interval set to? Each time data is collected, the graph will be updated. Laura Adams ********************************************************************** ═══ 86. Can SPM/2 Graphing monitor a CD ROM? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Thanks for the help, Lee ********************************************************************** ═══ 87. Installation Problem/Question ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi, I'm trying to install SPM/2 V2.0 on my OS2 2.0 machine. I am using the following command from the C:\ prompt: a:spminst I get a message that directory SPM2V20 have been successfully created and then the install program runs thru diskette 1, prompts me for diskette 2. I then get an error message as follows: SPM0124 Unable to add program SPM.EXE to SPM 2.0 folder (SYS4866). Am I doing something wrong? Is there a bug? Thanks for any and all help!! Al Agacki AGACKI at STLVM6 ********************************************************************** The person who wrote our install program says that "adding the program to the folder" failed due to a return code she's getting from PM. System error "4866" is "Initialization file was corrupted." You didn't do anything wrong ... somehow the .INI file is corrupted. Just as a shot in the dark: do you have enough disk space? Laura Adams ********************************************************************** Hi, I've got 90Meg available on the C: partition. I'm going to run MAKEINI to recreate OS2.INI and OS2SYS.INI to see if SPM/2 will then install. Will keep you posted! ********************************************************************** ═══ 88. System lockup? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I'm seeing a system lockup when accessing the SPM/2 User's Guide and Reference, while a Graphing and Recording session is under way. System is OS/2.1, SPM/2 2.0 with Theseus/2 fix applied. Machine is a model 95-33Mhz with 24MB, lots of free space on HD for swapper, 5% utilized. Any ideas? Dave LeGrys ********************************************************************** ═══ 89. THESEUS2 (2.0t) REXX API SYS0004 problem ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am using the REXX API's of THESEUS2 (version 2.0t from BB) and after several invocations of the REXX CMD in the same session i start getting SYS0004 THE SYSTEM CANNOT OPEN THE FILE. THESEUS shows serveral copyies of THESEUS2$ open. The problem seems to be that either THESEUS2$ is not being closed when the CMD terminates or that RT2LoadFuncs is not reusing THESEUS2$ if one was already open for the process. There does not appear to be a T2Terminate available under REXX, so i cann't close the device myself. If i skip calling RT2LoadFuncs after the first time in the session, things seems to work ok, but i have no way of knowing if RT2LoadFuncs had ever been called in this session and not calling it the first time is very, very fatal. REXX is OS/2 2.0 at XR06050. Roger Goetz / Chubb Group of Insuarance Companies / 908-580-2944 ********************************************************************** The person who wrote this REXX API (James Summers) is currently out on personal business, but I was able to get some answers for you. He said that he hadn't ever thought of someone calling the RT2LoadFuncs multiple times, so he only designed it to be called once. What this does is loads and initializes all that you need, and then there's no problem in re-using things after that. I don't think there's anything "fatal" if you happen to not call it the first time ... will probably just get some errors. You're right -- there's no terminate function... James will generate some sort of fix when he returns. Until then, I don't think you'll have a problem using it as you are. Laura Adams ********************************************************************** ═══ 90. TRAP E with 2.1 + CM/2 + NTS/2 + SPM2 2.0 + Custom apps ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * CONFIG: 8556-059 SN55-7Y28R 12MR 8518-002 SN 55-AKF69 IBM Mouse Bus Master IBM 4/16 MBits/Sec card OS/2 2.1 + CM/2 + NTS/2 + SPM2 2.0 C:лOS2лINSTALLлSYSLEVEL.GR Version 2.10 ID de composant 562107701 Type 0-2 Niveau de modifications en cours : XRF2010 C:лOS2лINSTALLлSYSLEVEL.OS Syst}me d'exploitation de base IBM OS/2 Version 2.10 ID de composant 562107701 Type 0-2 Niveau de modifications en cours : XRF2010 SPM2 2.0 + THESEUS Fix E:лSPM2V2лSYSLEVEL.SP IBM System Performance Monitor/2 Version 2.00 ID de composant 562201000 Niveau de modifications en cours : WR06000 Get a TRAP E TRAP 000e ERRCD=0000 ERACC=**** ERLIM=******** EAX=795f12b4 EBX=7be2be74 ECX=7b6e005d EDX=7b6ec608 ESI=7be055ab EDI=7be02b1c EBP=00005854 FLG=00012246 CS:EIP=0160:fffb4627 CSACC=c09b CSLIM=ffffffff SS:ESP=0030:00005820 SSACC=1097 SSLIM=00004967 DS=0158 DSACC=c093 DSLIM=ffffffff CR0=8000fffd ES=0158 ESACC=c093 ESLIM=ffffffff CR2=fff86f78 FS=03b8 FSACC=0093 FSLIM=00000023 GS=0000 GSACC=**** GSLIM=******** Erreur interne d'ex{cution @ l'adresse ╢╢0160:fff5fbd5. 60000, 9084 038600d1 Internal revision 6.514, 93/04/12 I've got 3 times this trap when i raise the custom apps after loading SPM (PSPM doesn't trap) It seems it happens when the swapper is about 18M and W.S. is 7,5 M What can i do with these trap's information ? Thanks for any help or suggest Is there a useful doc about using SPM ? ********************************************************************** ═══ 91. Documentation ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Some unmentionable person has borrowed my SPM documentation and hasn't returned it. Does any kind soul out there have a LIST3820 version they could send me Please. Thanks John Leech LEECHJA at GFDVM2 ********************************************************************** It's on it's way. The file name is SPMREF20 TERS3820. You'll need to unterse it with the (unpack option. Laura Adams ********************************************************************** Thanks Yet again Laura.. John Leech ********************************************************************** [Reference to append at 15:41 on 93/09/18 by LSADAMS at AUSVM1 ] Laura, please make an LP3820 (OS/2 app) version of the SPM/2 v2.0 docs available for download via OS2BBS. The LIST3820 version that is currently available cannot be "unpack" on my VM host. I get the error message like "the file is not in PACK format." I've tried several times .. but no go! Please advise. Thanks! Lawrence Liu - World Savings ********************************************************************** >> Sorry, I didn't put the document on OS2BBS, and the person who did >> is out on vacation today. Howveer, I don't know if the OS2BBS version >> IS packed. If it has a filetype of LIST3820, it probably is NOT >> packed. If the filetype is TERS3820, I believe it's tersed/packed. I was wrong about some assumptions I made in the first time I appended to this ... I just now logged onto OS2BBS for the first time ... and I see that the owners of that bulletin board packed the document in a different way than I was familiar with. I followed their instructions, and everything turned out fine for me (although the "unpack" option on COPY was a new one to me ... maybe your VM doesn't support it?). I think we'll have to wait until Monday to get to this one. Laura Adams ********************************************************************** I bet I know why you couldn't unpack the file on VM, because it just happened to me. I didn't follow the instructions exactly for uploading the unzipped file (SPM2DOC.PAC). When you upload this file, make sure that you use the correct options on the send command - send spm2doc.pac a:spm2doc pack3820 (recfm f lrecl 1024 I left off the stuff after the open paren and I got the same message that you did. After I uploaded it correctly, then I could use the unpack option on the copy command. Give this a try. Pam Lilly (SPM/2 Development) ********************************************************************** Can I impose on someone to create a LP3820 version of the SPM2DOC on the BBS? We don't have a VM system in house and cannot use this version because its pack ed (I think). I appreciate any assistance from anyone. Wil Kimura (Fireman's Fund Insurance Co) Wil Kimura Fireman's Fund Insurance Co., San Rafael, CA. ********************************************************************** To print the SPM 2.0 hardcopy documentation on a laser printer, you don't need VM. On the OS2BBS there are two (2) versions of the documentation: LIST or list3820 PACK or packed (compressed) list3820 The LIST version can be downloaded (binary), unzipped (I think), and printed on a laser printer using the LP3820 program. Both the ZIP and the LP3820 tools are available on OS2BBS. The PACK version can be downloaded (binary), unzipped (I think), uploaded binary to VM, unpacked, and printed on a 3820 printer. Why are they zipped? To save disk space. Why is the list3820 version packed? Because it won't download correctly in its list3820 form. ********************************************************************** Are the LIST or PACK files available internally (someplace other than the OS2BBS)? It seems silly for me to have to dial the IBM BBS in the US to have to get these. If no, could someone please send them to me? Thanks. Scott Atkinson TeamOS2, IBM Support Centre, Sydney, Australia SCOTTA at SYDVM1 ********************************************************************** [Reference to append at 21:30 on 94/02/16 by HOLCK at AUSVM1 ] I went looking and only saw the LIST3820 version of the SPM/2 documentation, At least that's what the package discription says it is and there is no reference to the files that can be printed using LP3820 ... on the OS2BBS that is. Keith A. Blake (The OS/2 Mole) V: (513) 983-8119 (USPGC9GC/IBMMAIL) ********************************************************************** Somehow, the OS2BBS was at some point restored from a backup and the SPM 2.0 doc files were wiped out. They should be back by Monday 3-7. Thanks for bringing this to our attention. ********************************************************************** Thank you for the quick response, downloaded and printed just fine. Keith Blake (The OS/2 Mole) V: (513) 983-8119 (IBMMAIL/USPGC9GC) ********************************************************************** ═══ 92. Mandatory Kernal Fix ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I was receiving a Trap E on a PS/2 Model 70-A21, 16 MB memory, and 120 MB disk running OS/2 2.1, CM/2, HPFS, DB/2, FFST/2 and LAN Requester with PEER Services (Net Share IPC$). Each time I start up a session monitoring local, I start col lecting cpu, disk 1, hpfs cache, lan requester, memory, and working set. This session always starts fine! Next I try to start a session for our Domain Contr oller, a PS/2 Model 80 running OS/2 2.0 with LAN Server 3.0 Entry. When I sele ct run and machine id, the cursor blinks a couple of times, and then trap E wit h only one or two lines of addresses. I opened a problem ticket to receive the fix, PJ09893, but they are not sending it to me yet. They collected my parameter files to review them, but have not indicated that the kernal fix is the route to go. Can I pull this fix from the bulletin board so I can continue my testing? Lew Poleate CIGNA Hartford,CT ********************************************************************** It looks like the OS/2 support team isn't convinced that your problem will be fixed by apar PJ09893. Without seeing the trap information, I can't tell either. My suggestion to you is to keep working with the OS/2 support folks to resolve this. But you still need the fix for PJ09893 whether it fixes this problem or not. With it you may avoid future problems. There has been some confusion about which products require PJ09893, and we are working to make sure that the confusion is cleared up. So keep pushing to get the OS/2 support people to resolve your problem. Pam Lilly (SPM/2 Development) ********************************************************************** Received the following note: << I just got off the phone, talking to my favourite CE. He informed << me, that unless you've got a severity 1 PMR, the change team will << not deliver ANY fixes for closed apars. This is a new policy they've << implemented for OS/2 2.1. I suspect this will be a problem for the << large part of the forumers, you are addressing with your note. << << Are there any other ways to get this fix? << Is it important enough to create a severity 1 pmr? I've talked to John Graham, who is manager of the department in Boca that manages the fixes (I think), and he said that any of our (SPM/2's) customers that need this fix (PJ09893) can get it by simply telling the support contact. You will need to tell the support person that it is "mission critical" (which probably translates to "severity 1") in order to be be high enough in the queue to be serviced in a timely manner. More information about the problem/fix: This bug can cause you to have a kernel trap if you: 1. Have an IFS statement in you CONFIG.SYS which points to HPFS. Note, this is the default and will be there unless you explicitly edit the CONFIG.SYS and remove the IFS statement. or 2. Collect thread or file data. The problem was found during testing of LAN NetView Monitor, but is applicable to SPM/2 as well. The problem is in the underlying function in the kernel that both of the products use. The location of the trap will vary according to the level of OS/2 2.x that you are running. The fix is in the kernel itself and not in a DLL nor the SPM/2 / Monitor code. I am sorry if we have caused anyone unnecessary work in trying to get the fix. Boca is in the process of changing some of their customer support policies and we don't have all of them "aligned properly", yet. James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** ═══ 93. Measuring .DLL activities for CICS OS/2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We are trying to create a SPM report summarizing by application level. CICS OS/2 uses mainly .DLL modules besides of .EXE. However, definition of the application does not permit any DLL's to be specified, only .EXE are allowed. How can CICS OS/2 activities be measured accurately ? Thanks for any help. Walter Krischker ********************************************************************** ═══ 94. (no subject) ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Lawrence Liu - World Savings ********************************************************************** ═══ 95. Recommendation for # of concurrent monitoring sessions ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Are there ary guidelines/considerations for the maximum number of machines that can be remotely monitored by a single machine? ---------------------------------------- __ Otto Kitchens \/ ISSC Atlanta otto@atlvm2 ********************************************************************** Hello Otto! The limit on the number of monitored machines is defined by the amount of memory on your system. It also depends on whether you are trying to monitor many machines in one monitor session or in multiple monitor sessions. On a 16 meg 386 machine, we found that we couldn't get more than 12 monitor sessions going without bringing the machine to a near halt. But in one monitor session, we could monitor 40+ machines. So the general guideline I can give you is that the fewer monitor sessions you have, the more machines you can monitor. Pam Lilly (SPM/2 Development) ********************************************************************** Hey, Pam. What then are the considerations between monitoring multiple machines in one monitor session versus monitoring machines in concurrent monitoring sessions? It would seem that I must lose something by monitoring multiple machines in one session; otherwise, why would I want to run concurrent monitor sessions? Am I missing something? ---------------------------------------- __ Otto Kitchens \/ ISSC Atlanta otto@atlvm2 ********************************************************************** The only reason I can think of that you would want to have separate monitor sessions for separate monitored machines is if you want to watch a graph of more than one machine at a time. If you group all of your monitored machines into one monitor session, you can only see one at a time on the graph. But if you have multiple monitor sessions, you can start multiple graphs and keep an eye on them all at the same time. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 96. Monitoring Multiple Workstations ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I know that it is possible to monitor multiple workstations via multiple monitor sessions because I have actually done it. It is also possible to place multiple workstation names in a single monitor session .log file, but it appears that it is possible to monitor only one of them at a time (at least that is what seems to take place when any of the included workstations is selected from the SPM/2 graph window). If this is true, what is the purpose of placing multiple workstations in a single monitor session .log file ? Is it true that the only way to monitor multiple workstations at the same time from the same monitoring workstation is to use multiple monitor session .log files ? Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** NOOOOO! You most definitely CAN monitor multiple machines at a time in one monitor session (or log file). You can only see one at a time on the graph, but SPM is still collecting data for all the machines you selected. You can see that by running a report to look at any or all of the machines you monitored during a session. You can also switch the graph between the various machines you are monitoring by clicking on the node pushbutton (the second one from the left). When you switch between nodes, the graph starts graphing from the leftmost edge, but data is still being collected for all nodes in the log file. Makes sense? Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 97. CPU totals not correct? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We reported a measurement and the following is in the report: ----- Processor ------ ------- Threads ------ Summary Busy Avg Time % Util Time Dispatch Timeslice ======== ======== ============ ======== ============ 17:14:56 56.60 00:00:22.651 6049 00:00:00.004 11.63 00:00:04.653 3032 00:00:00.002 46.80 00:00:18.731 5423 00:00:00.003 4.82 00:00:01.929 86 00:00:00.022 2.90 00:00:01.161 540 00:00:00.002 The first line supposes to give the totals and the lines below present the breakdown but the breakdown do not add up to the totals. Is this a bug or am I missing something? Cheers, Johan Stabler PS. For the interupts it says 4.84 % Util, Rate 66, and service time 00:00:01.096 ********************************************************************** (This was diagnosed off-line, but the answer needs to be made public.) The analyst has customized the report's definition of "applications" to add 2 of their own. However, one process name was included in both of the application definitions. The SPM report generator simply place the CPU used by that process' threads in both application's data, which keeps the total from adding up the summary. So, ya'll, be very careful when defining your own "applications" that you only have a process in one of them. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 98. Remote Monitoring doc confusing " ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: Append ated at 16:42:09 on 93/03/30 GMT " Well I check the append that you suggested and also the forum header for the parameter requirements. Things look ok on this end but I still get pending or offline status for any remote machine that I am trying to monitor. Local monitoring works fine. fyi Per Karlsen ISSC Intergrated Systems Solutions Corporation Dept. CCKA / B671 / Rm LAa02 Regency Park Cary N.C. Cary Tie Line 883-7145 Cary Outside Line 919-469-7145 From: PER at CARY ********************************************************************** ═══ 99. <<< hmte.... error >>> msg. ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * while browsing thru the theseus2 tool (V2.01) of SPM/2 (in an attempt to solve memory disorder in the application) we opened "Shared Object Summary" in the Process pull-down menu. there is a list of shared objects belonging to the selected process. the list contained hundreds of lines ending with the following message: >>> hmte (????) does not refer to a MTE -- Error <<< the same list appeared in the "Shared Object List" of other processes too. All the objects at hand were shared giveable segments allocated to my processes, and "Given" to other processes. curiously, after killing one of my processes (the receiver of most of the giveable objects), the messages disappeared from all the lists. Can someone pls explain to me the meaning of this message., since we suspect it has something to do with the memory problem that we have. Baruch (Bubi) Rubin ********************************************************************** There was an error in THESEUS2 that caused problems like this in the Shared Object displays. The 2.0t version on OS2CSD should have them fixed. James Summers (SPM/2 Development & OS/2 Performance) ********************************************************************** The version that we are using is 2.0t and we got the error (there was a typo in the original append stating that the version was 2.01). Is there a chance that it was not fixed? I just saw version 2.0v on OS2TOOLS, is that more advanced? Or is it possible that this points to an error in the application (there are other objects which do get allocated and freed, as seen when running the memory leakage option)? P.S. I was on vacation so I asked one of the other workers to place the original append. Yakov Shasha Tel Aviv, Israel ********************************************************************** Please try the 2.0v version and see if the problem goes away. If not, please re-append and I will see if I can determine what is happening. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 100. THESEUS2 REXX API PROBLEMS ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I'M TRYING TO USE THE THESEUS2 REXX API TO OBTAIN WORKING SET INFORMATION & AM RUNNING INTO SEVERAL PROBLEMS. 1) EACH TIME I CALL RT2WSPROCESSTICK IT APPEARS THAT MEMORY IS ALLOCATED AND COMMITTED BUT NEVER FREED. EVENTUALLY MY SWAPPER FILE GETS HUGE. 2) AFTER RUNNING MY PROGRAM FOR SEVERAL MINUTES (SOMETIMES 10+) I GET A SYS3175:A PROGRAM IN THIS SESSION ENCOUNTER A PROBLEM... (REXX.DLL IS REFERENCED IN THE REGISTER DUMP). IF I TURN REXX TRACING ON AND REDIRECT IT TO A FILE THE LAST ENTRY IN THE FILE IS EXECUTING RT2WSPROCESSTICK. ANY ONE HAVE ANY IDEAS/SIMILAR PROBLEMS?? D STOYER -- MERIDIAN BANCORP X ********************************************************************** Thank you for finding this memory leak. The REXX interface I use to retrieve the values for the handle and interval allocates memory and nobody frees it, because I didn't realize it was doing this. (Another example of a service routine doing something that you need to undo and then the documentation not telling you this.) This will be fixed in the CSD which we are currently building. I will ship you a copy to use as soon as I figure out how I'm supposed to get rid of the memory. | I've got it ready. Now, how can I send it to you? OS2CUST is not recognized | by my file router. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** HOW ABOUT COMPUSERV? MY USERID IS 72740,102 THANKS FOR THE QUICK RESPONSE. D STOYER -- MERIDIAN BANCORP ********************************************************************** Pam Lilly is the one with the CompuServe access and she is on vacation. The only thing I have access to is VM. Is there a branch office that I could send it to and have them get it to you? James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 20:28 on 93/10/06 by MER1 at OS2CUST ] Please append your mailing address. I've been told to send a diskette via overnight mail. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** Overnight may not be necessary since I'll be out of the office until Tuesday (10/12). My address is below: Dave Stoyer Meridian Bancorp, Inc One Meridian Blvd Wyomissing, PA 19610 I can also be reached by phone at (215) 6550532. Thanks. D STOYER -- MERIDIAN BANCORP ********************************************************************** [Reference to append at 22:56 on 93/10/05 by MER1 at OS2CUST ] This APPEND was deleted at 15:46:15 on 10/07/93 EST. ********************************************************************** [Reference to append at 19:50 on 93/10/07 by MER1 at OS2CUST ] It should go out on AirBorne tomorrow. Airbill number is 4806210360. Should get there on Tuesday. | The diskette contains version 2.0v1, which is later than the version that | is being put on OS2TOOLS and the other electronic distrubtions. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 19:50 on 93/10/07 by MER1 at OS2CUST ] Did you get the fix? If so, did it work for you? James Summers (OS/2 Performance & THESEUS2) ********************************************************************** Sorry I took so long to respond. I did receive the fix, and it does appear to fix the memory leak. Thanks D STOYER -- MERIDIAN BANCORP ********************************************************************** ═══ 101. Simulation tools ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Does anyone have any idea whether any tools is available to simulate CPU, RAM, harddisk utilization?? I want to demo SPM/2 to a group of customers but do not have the tools to generate the utilization. Thanks for your help. Manson Yang ********************************************************************** Manson, Yes, we have 2 tools: CT.EXE (CPU and Disk activity) RAMHOG.EXE (RAM, obviously) I'll send them to you. Laura Adams ********************************************************************** Hi, Recently you sent these programs to someone in need of simulation tools. I am also in need of these for some education material on SPM/2 I'm putting together. Could you send them to me please? > Yes, we have 2 tools: CT.EXE (CPU and Disk activity) > RAMHOG.EXE (RAM, obviously) FRED, ALFREDO @ BCRVMMS1 ********************************************************************** [Reference to append at 01:42 on 93/10/07 by YANGM at HKGVM8 ] Do you know of any LAN-based script driven tool (like host TPNS) to simulate different testing conditions? Any hints/tips are welcomed. An D. Nguyen ********************************************************************** Have you considered Workstation Interactive Test Tool (WITT)? Dave ********************************************************************** ═══ 102. HPFS386 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I believe I had SPM/2 monitoring disk activity (graphing and data) on a HPFS386 drive on a server (both locally and remotely), this then appeared to stop working (ie always zero activity) after applying the Lan Server 3.0 servicepak (IPU7001 with new HPFS386 modules). Upon reading appends earlier in this forum it seems that it should never have worked in the first place as there are no performance hooks in HPFS386! Was I imagining things originally? Can anybody confirm if I should be able to monitor HPFS386 activity? Thanks Simon ********************************************************************** ═══ 103. version 2.0v of THESEUS2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Version 2.0v of THESEUS2 is on OS2TOOLS. It will be going onto OS2BBS and CompuServe next week. It has gone back to OS2TOOLS as a result of a letter sent by Lou Gerstner about sharing tools within IBM. The update file has been changed from a .zip file which required you to have PKUNZIP to a "self-exploding" .exe file. This way you do not have to have PKUNZIP. This version has a new device driver. This is the same level of device driver shipped in LAN NetView Monitor. If you are planning on installing LAN NetView, you need to update your THESEUS2 files BEFORE installing LAN NetView. This is the last update prior to the SPM/2 2.0 CSD. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 104. Remote monitoring problem ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have followed the instructions carefully and did the following: On the OS/2 2.1 with Lan Server 2.0, I installed the SPM 2.0. Next, I increased the commands, names, & sessions in protocol.ini and also checked the maxconnections, maxusers, and numreqbuf. On the OS/2 2.1 with Lan Requester 3.0, I installed the distributed feature. I checked on the lan connections by checking on lantran.log and accessing the server's files. I started SPM 2.0 Control on the server and could see that it could monitor itself, but it could not monitor the remote system. The remote system has the following errors in SPMSNAP.ERR ---------------------------------------- SPMSNAPP: Print Utility Version 2.0.0 SPMSNAPP: The file to be formatted is spmsnap.err @ 1993-10-09-22.15.20.62 P:10 T:1 L:128 SPMNBL: Listen requested @ 1993-10-09-22.15.21.18 P:10 T:1 L:128 SPMNBL: Network available. Network List ner spmnet.exe (PID 14) started on behalf of LOCAL0 @ 1993-10-09-22.15.21.22 P:10 T:1 L:128 SPMNBL: NBL listening for DCF start req ests @ 1993-10-09-22.17.25.56 P:14 T:1 L:32 SPM0049: SPMISTRT: Listener unable to re ister with NETBIOS ->0047:7c72: 08 00 00 00 ................ ---------------------------------------- SPM0049 indicates something is wrong with our network. We are using NetBios and I am communicating with other systems. So what is possibly wrong. Help... Help... and the requester's name is core-r02. Eric Louie ********************************************************************** Whenever I click on the button Query from the Monitor...Setup Option, the SPM 2.0 would close on itself. At first, I thought reinstalling SPM 2.0 would fix the problem. But the problem still occurs. Does that symptom be related to SPM's problems talking other stations? Eric Louie ********************************************************************** I bet you're running on OS/2 2.1, right? We know about that bug. The workaround is to manually enter the names of the workstations that you want to monitor. And watch for our CSD, coming soon... Pam Lilly (SPM/2 Development) ********************************************************************** Yup, Pam, I am running OS/2 2.1 and I will wait for the CSD to come out. In the mean while, I have tried entering the machine name and it still says the remote machine is off-line or pending. Eric Louie ********************************************************************** I know what this problem is too! We also have a problem where we don't recognize machine names with a dash ('-') in them. In your first append you said that your requester's name is core-r02, so that's the problem. The workaround is to change the requester name (not very nice, I know) and to check out the CSD when it is available. Pam Lilly (SPM/2 Development) ********************************************************************** Thanks, PAM, for the help. Eric Louie ********************************************************************** ═══ 105. SPM/2 Code Availability - Lou Gerstner Memo ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * As per Lou Gerstner's Memo, how can obtain a copy of SPM/2 for marketing and demo purposes only. Thanks. ********************************************************************** Our policy has always been to provide copies of SPM/2 for marketing, demo, or education purposes. I'll send you a copy. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 106. SPM/2 and LU6.2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi, are there any plans to allow SPM/2 to gather data from a remote station attached in an SNA network? That is: ------ I know that NetBIOS is only supported now. However, I have a client who is using DCAF's LU6.2 to LU6.2 connection and is thinking of using SPM/2 as a LAN tool to watch the file server's metrics. If the answer is "yes" then another tool my client would add is LMU/2 for alerts. Thanks, Kong. ********************************************************************** ═══ 107. Disk Graph ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * How can I get SPM/2 to graph disk activity for multiple harddrives? I have two physical drives installed in a Mod57. The physical drives are divided into five logical drives. All that the disk graph shows is "Disk 1". Larry Armstrong ********************************************************************** When you create your log file, you need to modify your resources to include Disk 1 & 2. You do this by selecting the Preferences pulldown, then the Resources menuitem. This takes you to the Resources panel. On that panel, select the checkbox to Expand available resources. Then select Disk 2 from the Available list. Now you're ready to collect data from both physical disks. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 108. Pentium ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Are there any known problems when SPM/2 runs on a pentium machine? We are running an application which runs fine without SPM, however, the application is fully utilizing the machine. When the application is started and SPM is active, SPM graphs show after 30 secs when the application is active, no more CPU actitvity and no more disk activity, in other words the application stops working without any reason Since SPM is the only 'extra' code running compared to the normal situation we are suspecting SPM. Any suggestions or known problems?(Machine configuration: Mod 9595-OPT, Pentium chip, 64 M memory, 1G and 2G drive) Thanks, cheers, Johan Stabler ********************************************************************** ═══ 109. Theseus2 API (more function needed) ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * In the older (full screen) version of Theseus2 I seem to recall that we could create a script and every once in a while generate a report of any function that we needed from Theseus and have it output to a file. This function is not available in the PM version (the reports can only be generated from the user interface). I tried to use the REXX API (the C API does not seem to have more function), but it does not give everything that is available in the user interface. Our scenario is that we suspect that a certain process is causing problems (same one I appended about with the HMTE error, whatever that is). On an intermittent basis the PS hangs due to an out of storage condition. We would like to run Theseus and every once in a while output data (we suspect the shared objects), since once the PS gets to the state that it is out of storage Theseus can't get the information we need (for obvious reasons -) ). That is why we want to collect information up to (and before) the problem occurring. Is there a way to do this with the current product, or do we need additional API's, or will we need to write a PM program that will feed the necessary keystrokes (in place of the user) every once in a while? Yakov Shasha Tel Aviv, Israel ********************************************************************** I'm sorry. There is no way to do what you want to do with THESEUS2. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 110. SPMDIR ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Does SPMDIR space allocation calculations include the EAs associated with the files? Rich Hall ********************************************************************** SPMDIR does not seem to know anything about EAs. Sorry. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** In your up coming SPM/2 2.0 PTF/SP you might want to point this out in the doc/readme. Rich Hall ********************************************************************** ═══ 111. when is the csd out? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The title says it all thanks mike prendergast ********************************************************************** ═══ 112. Memory Leak in SPMDCF.EXE ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * About a month ago, I received an updated SPMDCF.EXE (SPMDCF EXE 73539 9-09-93 9:16a) that was supposed to resolve a memory leak problem. I've installed the update and after running for several days with PERFCAP2, I think there is still a problem, though less pronounced. Attached are two snapshots from Theseus2 taken two days apart that show a significant increase in memory utilization for SPMDCF. Also, 4.3MB RAM for SPMDCF sounds excessive. Our intent is to monitor our server performance 24x7 using PERFCAP2, and this seems to be causing our servers to run out of RAM resources, forcing an outage! Any help would be appreciated... Larry Memory Utilization for Process with PID = 002D, name = 'SPMDCF': bytes bytes number bytes bytes allocated committed present each present description 00000824 00000824 1 0824 00000824 PTDA 00000A08 00000A08 6 01AC 00000A08 TCBs 00006000 00006000 6 1000 00006000 TSDs 00010000 00006000 6 1000 00006000 LDT 00000200 00000200 1 0200 00000200 Process Page Directory 00080000 00014000 20 1000 00014000 Page Tables 04000000 005F5000 508 1000 001FC000 Accessible Shared memory 00880000 00018000 13 1000 0000D000 Originated Shared memory 01350000 003A9000 746 1000 002EA000 Private memory -------- -------- 0009742C 0002142C 0002142C Total System 00880000 00018000 0000D000 Total Shared originated 01350000 003A9000 002EA000 Total Private -------- -------- 01C6742C 003E242C 0031842C Total RAM for the Process 29085 3977 3169 (in Kbytes) 28.404 3.884 3.095 (in Mbytes) TWO DAYS LATER=========================================================== Memory Utilization for Process with PID = 002D, name = 'SPMDCF': bytes bytes number bytes bytes allocated committed present each present description 00000824 00000824 1 0824 00000824 PTDA 00000A08 00000A08 6 01AC 00000A08 TCBs 00006000 00006000 6 1000 00006000 TSDs 00010000 00006000 6 1000 00006000 LDT 00000200 00000200 1 0200 00000200 Process Page Directory 00080000 00015000 21 1000 00015000 Page Tables 04000000 005FE000 508 1000 001FC000 Accessible Shared memory 00880000 00018000 13 1000 0000D000 Originated Shared memory 014E0000 004E2000 1058 1000 00422000 Private memory -------- -------- 0009742C 0002242C 0002242C Total System 00880000 00018000 0000D000 Total Shared originated 014E0000 004E2000 00422000 Total Private -------- -------- 01DF742C 0051C42C 0045142C Total RAM for the Process 30685 5233 4421 (in Kbytes) 29.966 5.111 4.318 (in Mbytes) Larry Pisko, TheTravelers ********************************************************************** The version you got was patched to fix a trap, which also fixed one of several memory leaks that SPMDCF had. The other leaks were not fixed by this patch. We are testing a version with a number of fixes for memory leaks; however, this version will not be available until our full CSD is ready (there are dependencies with other modules, which are not yet ready). Our CSD will not be ready before 11/15 and maybe not then. Our plan just "fell apart" and we are trying to get another one together. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 113. CD-ROM DRIVERS FOR 2.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I HAVE INSTALLED OS/2 VER 2.0 ON MY SYSTEM, WICH INCLUDES A CD-ROM DRIVE. AFTER INSTALLATION WAS COMPLETE THE SYSTEM COULD'NT RECOGNIZE THE CD-RO M GIVING ME SOME ERROR MASSAGES. THE DRIVE IS SONY MADE. MY QUESTION IS: WHAT H AVE I DONE WRONG DURING INSTALLATION AND ARE THERE ANY NEW DEVICE DRIVERS AVAIL ABLE FOR SONY CD-ROMS?. THANX ISAAC H. ********************************************************************** Two questions: 1) SPM/2 2.0 requires that either Service Pak 1 or 2 be installed. Do you have either of these? 2) Do you know if the CD-ROM player has it's own device driver? If it does, (rather than using some standard device driver installed with OS/2), you won't get any performance information on physical activity for that drive. SPM/2 gets its info from performance hooks that are embedded in the system -- and for physical drive activity, those hooks are embedded in the device drivers. Note however, that you should be able to get "file" activity for this drive. These hooks are installed in the file system, which is always used independent of what physical device you eventually go to. Laura Adams ********************************************************************** ═══ 114. Installation Question ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None I am attempting to install SPM2 version 2 on my OS2/ 2.1 system. at the end of the installation, after notification that CONFIG.SYS is being modified, I receive the following: **************************************** SPM0129: Adding programs to SPM 2.0 folder. SPM0123: Unable to add folder for SPM/2 2.0. (SYS5391) SPM0124: Unable to add program SPM.EXE to SPM 2.0 folder. (SYS5391) SPM0124: Unable to add program SPM.EXE to SPM 2.0 folder. (SYS5391) **************************************** Can anyone offer an explanation as to what the matter might be? Charles F. Rouh Executive and Decision Support Systems Southlake, TX ********************************************************************** I talked to the person who wrote installation and she said the following: " This error happened while SPM20 installation was trying to add spm icon to the spm2 folder. The PM error code for #5391 is: 'pmerr_read_only_file.' I cannot find help for this error. The only guess is that for some reason, the user's OS2.INI or OS2SYS.INI file has been changed to the read-only attribute. The result of this error will be that the SPM/2 folder creation is incomplete and the user will no be able to use the icons. If one of the .INI files really is 'read-only,' its attribute can be changed by typing: 'attrib -r filename'. " Laura Adams ********************************************************************** Thanks! That was the problem. I can't figure out what is setting the OS2.INI and OS2SYS.INI files to read only, but everytime I reboot my machine, thats the way the files come up. Charles F. Rouh Executive and Decision Support Systems Southlake, TX ********************************************************************** ═══ 115. Capacity Planning in a Client/Server Environment ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I was wondering if anyone had successfully used SPM/2 to come up with a version of a capacity plan on servers. Cynthia McCallum Union Gas Limited ********************************************************************** At IBM we tried very hard to create a formal capacity planning tool that used SPM 1.0 and SPM 2.0 performance data as input. This task proved to be very hard and ultimately we did not release a tool to the field/customer. I believe that the best way to approach the capacity planning issue is to do capacity TREND ANALYSIS for installed servers. Collect OS/2 resource information such as CPU utilization, Disk utilization, and Memory working set to see how much of each resource is used. Collect over weeks or months. Collect during peak hours. Put the data into a spreadsheet template that will graph the resource usage for each of your LAN Servers. Stare real hard at it and decide when your servers will run out of capacity. Then upgrade to the biggest fastest machine you can buy so you don't have to worry about it again for awhile. ********************************************************************** ═══ 116. SPM/2 V2.0 hang at config.sys time ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We have a PS/2 9595 running OS/2 2.1 GA, CM/2 1.01, DB2/2 V1.0 and SPM/2 V2.0 When the machine boots and we load RUN=C:\SPM2V2\SPMNBL.EXE /LISTEN and SET SPMSNAP=E C:\SPM2V2\SPMSNAP.ERR from the config.sys the machine hangs (meaning the machine beeps and the cursor goes to the upper left hand corner of the screen and STOPS, never getting to a PM desktop). If we take out the above two statements from the config.sys it works fine, machine comes up as it should. We can then issue the SPMNBL.EXE LISTEN from an OS/2 prompt and it loads O.K. and is LISTENing, we cannot load it from the config.sys. Any thought would be appreciated..... Thanks in advance, Tom Talone (8/577-4148) ********************************************************************** ═══ 117. Logon attempt Unsuccessful ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When SPM/2 2.0 is monitoring a server and I try to logon to that server from the servers requester I get logon attemp unsuccessful. When I issue a SPMNBL /STOPLISTEN command I can logon. There is no problem logging onto the server from a remote requester when SPM/2 is monitoring. Is this normall operation? Thank you in advance. Mike Pocchiari, rhqvm19(mikejp), 8-566-4820, White Plains NY ********************************************************************** Is the server being monitored from a peer server? There is a known problem in LAN Server 3.0 - APAR IC05298 - where if a peer server is monitoring a server and the server isn't logged on to a domain, and then you try to log on to a domain, you can't. You get the "logon attempt failed" message. The workaround is to log on at the server before you start remote monitoring with SPM. The fix that you can get for IC05298 will temporarily disconnect the remote SPM session while you logon. SPM will restart by itself after about 5 minutes. Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 20:41 on 93/11/02 by MIKEJP at RHQVM19 ] Mike, Do you have enough NETBIOS Sessions, Commands and Names in your IBMLAN.INI amd PROTOCOL.INI. Its possible that you when have SPMNBL loaded you have exhausted all you sessions, command or names. This would then cause your LAN logon to fail. Try increasing these parameters. Thanks, Tom Talone (916/641-4148) ********************************************************************** [Reference to append at 23:14 on 93/11/02 by SPM2 at AUSVM1 ] Pam, Thank you for the response. I am monitoring a server from a peer server. On the server I have IP07001 that includes APAR IC05298 installed. One correction. The command I issued at the server was SPMNBL /HALTALL. This put the server off-line to SPM/2 and it allowed me to logon to the server. Mike Pocchiari ********************************************************************** To review my setup, I have the following Monitoring Station Monitored station ------------------- ----------------- OS/2 2.1 OS/2 2.1 LS 3.0 peer server LS 3.0 advanced server w/APAR IC05298 SPM/2 2.0 I have increased the Netbios resorcess for Sessions, Commands, and names in the monitored station. But still can not log onto the server from its local requester unless I do a SPMNBL /HALTALL. Any help Mike Pocchiari ********************************************************************** [Reference to append at 15:36 on 93/11/03 by MIKEJP at RHQVM19 ] * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * ═══ 118. SPM/2 2.0 and IBM 3514 Disk Array ???? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Does SPM/2 V2.0 support the IBM 3514 Disk Array? We are monitoring a remote server collecting file data on both the C and D drives. The SPM/2 report does not show any reads or writes for ANY file on the D (3514) drive, but we know that we are writing to the physical drive. We are using the OS2DASD.DMD that comes with OS/2 2.1. Is the 3514 suported, or are we doing something wrong or is there some type of a fix for this. We are also using LS 3.0 but with just HPFS on the monitored machine. Thanks, Tom Talone ********************************************************************** Tom, I'm assuming that your remote server is an LS 3.0 Advanced server with HPFS386. If that is true, than the following will explain what you are seeing. In OS/2, there are performance hooks in the HPFS file system that allow SPM to collect file information. When you install LAN Server with HPFS386, the original OS/2 HPFS file system is replaced with HPFS386. There are no performance hooks for file information in HPFS386. So when you perform reads & writes on that server, the dosreads & doswrites are issued to the HPFS386 file system, where there are no performance hooks. So SPM can't see that file activity. Can you see Disk activity? OS2DASD.DMD is where the performance hooks for Disk activity are, but it has nothing to do with the File activity hooks. There is a workaround which will let you see the file activity on the server. Monitor the server's file activity on a requester's redirected drive. In other words, on the managing machine, define one or more redirected drives which map to the server's logical drives. Then monitor the file activity on the managing machine. You'll be able to see file activity on those redirected drives. What happens in this case is that dosreads & doswrites are issued to the requester's (managing machine's) file system. There are performance hooks in the requester's file system, so the reads and writes are seen by SPM. The reads and writes are sent over the network to the server, but they go to the HPFS386 file system, where there are no performance hooks. So SPM doesn't see that file activity. Pam Lilly (SPM/2 Development) ********************************************************************** [Reference to append at 07:35 on 93/11/03 by TATALONE at SFOVMIC1 ] Pam, Thanks for the information. I will check to see if their server is running 386HPFS and implement you recommendation. Thanks, Tom Talone ********************************************************************** File I/O statistics on a LAN Server Advanced ARE in fact supported by SPM 2.0 for LOCAL I/O. Local I/O is I/O requests issued by applications that are running on the server machine. The reason that remote I/O requests issued by LAN Requesters aren't collected/reported at the server is because those requests are handled directly by the Ring 0 LAN Server Advanced code and bypass the OS/2 file system for performance reasons. File I/O statistics on a LAN Server ENTRY are supported for both remote- and local-I/O requests. LS Entry is a Ring 3 application and ALL I/O requests go through the OS/2 file system where the file I/O hooks are. ********************************************************************** ═══ 119. (no subject) ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Thank you for the response. I do have IP07001 that includes the apar IC05298. ********************************************************************** ═══ 120. SPM/2 2.0 and DB2/2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We have run several tests using SPM/2 2.0 and DB2/2. SPM/2 is not picking up the file activity for files that DB2/2 reads and writes to. But if we use the OS/2 E editor SPM/2 will pick up the reads and writes of files that are on the same physical drive as the DB2/2 files. Do you know if DB2/2 is supported by SPM/2 for collecting file level data (reads and writes) Thanks, Tom Talone ********************************************************************** SPM 2.0 does support DB/2 file I/O. I ran some tests doing SQL queries from a DB/2 database and collecting file activity with SPM 2.0. SPM 2.0 did report file activity for a number of DB/2 files, including .DAT and .INX files. Some files had 0 requests, which means they were open when SPM started collecting but had no reads or writes. Are you getting absolutely no files reported for DB/2, or just some? Are you sure you actually have file reads/writes for the DB/2 files? Maybe DB/2 has an internal cache from which it satisfies some I/O requests so that it doesn't have to go to the OS/2 file system. ********************************************************************** ═══ 121. (no subject) ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * OS2SPM20 CFORUM appended at 10:40:55 on 93/11/04 EST (by MIKEJP at RHQVM19 Subject: Logon attempt Unsuccessful Ref: Append at 15:03:36 on 93/11/03 GMT (by MIKEJP at RHQVM19) To review my setup, I have the following Monitoring Station Monitored station ------------------- ----------------- OS/2 2.1 OS/2 2.1 LS 3.0 peer server LS 3.0 advanced server w/APAR IC05298 SPM/2 2.0 I have increased the Netbios resorcess for Sessions, Commands, and names in the monitored station. But still can not log onto the server from its local requester unless I do a SPMNBL /HALTALL. Any help Mike Pocchiari ********************************************************************** ═══ 122. Disk graph for ESDI drives ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When we start graphing on PS/2 Mod 80's with ESDI drives the disk line stays consistently at 75%. Is SPM/2 2.0 supported for ESDI? What device drivers must be in CONFIG.SYS for ESDI? Since these machines are "cloned" that might be the problem. Has anyone else been having this problem or know of a solution? Thanks. Lucie Wilson - Skill Dynamics ********************************************************************** SPM 2.0 does support ESDI drives. Older IBM Mod 80s have ESDI drives and we are able to collect Disk statistics on them in our lab. The SPM disk hooks for SCSI or ESDI are in the OS2DASD.DMD driver, which is loaded with CONFIG.SYS. Do you have IBM Mod 80s with OEM disk drives? If those disk drives use a different driver than OS2DASD.DMD then there would be no disk hooks (unless the OEM vendor implemented some using SPM's User Hook API). I don't know why you see disk utilization consistently at 75%. Maybe it's a LAN Server with disk mirroring (?) on, continuously copying data to/from another LAN Server's disk. 75% is quite a workout for a disk. ********************************************************************** ═══ 123. CSD availability?? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * What is the prognosis for the availability of the CSD? Thanks, Dave ********************************************************************** ═══ 124. resource group ordinals ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * | lunch time rush corrections I'm updating a .cmd file that some workstation performance testers use to build the SPM/2 2.0 log file. They pass in the file name and the .cmd builds and issues an SPMLOGF cmd. I don't have immediate access to an OS/2 2.x machine but last week I did, and noticed, in the list of resources to monitor, LAN Requester and LAN server were listed (from memory). I do not see these in the list of resource group ordinals in the list3820 guide. | That list shows 3-7,9,10,15-38,42-49. Could some post other resource group ordinals if they exist. Thanks, Rich Hall ********************************************************************** Rich, You know, you're the first person that's noticed this, including me! The ordinals for the missing resource groups are: IBM LS 3.0 Advanced - 200 IBM Lan Requester - 201 You can see this for yourself by looking at the file SPMCTRGP.H which gets installed in the \api subdirectory if you installed SPM with the /api option. Something to keep in mind - you can only see the data for these resource groups in a dump report - not in summary or tabular reports. These two resource groups are the only ones that are missing from the list in the on-line docs. The gaps that you see are for groups that existed once upon a time and were removed or combined with others for various reasons. Pam Lilly (SPM/2 Development) ********************************************************************** ═══ 125. Stopping a Monitor Session Automaticly ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * There is no provision in SPM 2.0 to stop/start a session at a specific time of day, or to run for a specific length of time. Some users have written REXX programs to do session scheduling using SPM 2.0 Command Line's SPMISTRT and SPMISTOP along with a Sleep command. ********************************************************************** Ref: none Is there any way to stop a monitor session at a given time or after a specified time period? Thanks, Norm ======================================== NOTICE: The entire universe, including this append, may one day collapse back into an infinitely small space. Should another universe later emerge, the validity of this information can not be guaranteed. ======================================== ********************************************************************** Note that the answer to this subject append is on the forum disk BEFORE the question. We are now so responsive that we can, incredibly, anticipate precise questions and actually answer them before they are asked. The answer to the next question: "It depends." ********************************************************************** Tim, I hope the "answer" to this question is not correct ("It Depends")! I am in need of the most recent code for some internal personal education before I perform this work at a client location on their production machine. I looked around the tools disks but could not find the product. I need the Distributed Feature (if possible) for 2.0 and 2.1 machines, as well as any fixes/patches for both. Any help would be greatly appreciated. Tom Friends - Account Systems Engineer - Services Co. Boston (T/L) 362-1245 (617) 895-1245 ********************************************************************** ═══ 126. Three questions from Raleigh ITSO ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Three questions handled off-line will be repeated here: 1. Why does SPM 2.0 report considerable PageIn and PageOut activity when a File report does not show any file I/O to/from the SWAPPER.DAT file? Answer: The SWAPPER.DAT file is not monitored for file I/O. When the performance hooks were instrumented in OS/2, it was decided not to maintain statistics on this one vital operating system file to avoid any performance degradation of OS/2. 2. In Reports, why is the difference between the Start and Stop times (which = Sum Interval if default report options are used) not the actual period of time that data is collected/available? Answer: Start time is the time the log file is opened. There is often a time lag until the first data snapshot is actually recorded to the log file. Unfortunately the period of time for which the data represents the performance of the system is not shown on the reports. One workaround is to use CPU statistics: TimeBusy/PercentBusy*100 will give the actual time over which data is collected. 3. Why are some of the rates in the Summary report such as paging rate or disk access rate truncated to integer values? Answer: This is a bug. The CSD will probably provide some decimal places and round off instead of truncate. ********************************************************************** ═══ 127. High CPU Again ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I've just read all of the appends which discuss high cpu utilization and feel confident that I suffer from none of their symptons. I am currently running SPM/2 on OS/2 2.1 and the only application also running are the clock, System Resources (from OS2BB), PASPORT/2, and PCPRINT. Right now Systems Resources shows 99-100% CPU. When I bring up SPM/2 he also reports the same high CPU, but I have no idea why. From what I've read, you say that it could be because of a DOS program. The PCPRINT program I'm running is for OS/2 and PASPORT/2 is for OS/2, so who is eating up the CPU. After I finish with this append I will take down PAS/2 and PCPRINT and see if that affects the utilization. G. Akridge ********************************************************************** Just finished checking and CPU remains at 98%. I'm not running THESEUS, but I did bring up MemoryU and according to it and the DESKTOP, virtually noting else was running. Virtually means that I've got the Minimized Windows up and also the TCP/IP minimized window up. That's it. G. Akridge ********************************************************************** Well, the suspense was killing me so I brought up THESEUS2 and discovered that I am indeed running MEMSIZE. I've followed the questions in the FORUM on MEMSIZE and am somewhat satisfied. Wasn't aware it was a DOS based program.. Is there any way of getting it to release cycles or must I close the window? G. Akridge ********************************************************************** ═══ 128. Using SPM 1.0 to monitor SCSI devices ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: From: Terry A. Steilen - Advantis Subject: Using SPM 1.0 to test performance problems on store systems Maybe you can help me out with a hardware question. Using SPM/2 on 1.3 we are testing our store systems on our register simulators. We can turn up and down the transaction rates, and vary the mixture of commands that are being sent to the ISP processor. This is very much like TPNS in the SNA world. The visual results that we see are: CPU utilization in the mid 50% range, varying +/- 7% or so. ( We have 5 SCSI hard drives....) The DBM database drive sometimes approaches 90+% utilization. The drive having the transaction software runs 20-30% or so. The oddity here is that it seems that the sum of all drive activity never exceeds 100% Could this possibly mean that we have reached an I/O limitation on the single SCSI card that controls all the hard drives? At a predefined transaction rate input from the simulators, when we add a process that just continually copies files around on the C drive (a dummy I/O task), the scsi utilization of that device goes up, and the utilization of some other drive...the database drive.. goes down. CPU utilization does not increase at all. There is little or no swapper activity. Eventually the database transactions seem to get behind...our registers time-out waiting for a response. How can I possibly tell if and when we reach an I/O limitation of the SCSI card itself? The graphs may be telling it to me, but I need some support from your hardware folks. For multiple hard drives on one SCSI card, should I see SPM/2 drive utilizations that when added together, approach some number? Any idea what that number is? Is there a way to deterine it in the lab. Could I expect to see drive utilizations at 90% simultaneously on 2 separate hard drives? ********************************************************************** SPM/2 Version 1.0 for OS/2 1.3 reports disk utilization on a per-physical disk basis. Logical volumes and SCSI adpaters are transparent to SPM. The physical drive requests are each timed from start to finish. The times are summed to represent the total time the disk must have been busy. If many requests to one disk are outstanding at one time (over- lapped), the Sum could be more than 100%, so SPM 1.0 uses Max ( Sum, 100% ) I don't know how SPM could be used to determine bottlenecks at the SCSI card. You could instead add another SCSI card and see if performance improves, or increase the speed of the CPU and disk and see if performance doesn't improve. ********************************************************************** ═══ 129. When is information logged to .LOG files? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: When is the data recorded in the .LOG files during recording? Yesterday I thought the .LOG file was being over-written because the size was very small after a period of time. When I stopped monitoring, the size of the log file jumped. Is the information held in memory for as long as possible? The machine was a server with 32M of memory. Tom Friends - Account Systems Engineer - Services Co. Boston (T/L) 362-1245 (617) 895-1245 Tom Friends - Account Systems Engineer - Services Co. Boston (T/L) 362-1245 (617) 895-1245 ********************************************************************** The file system buffers the data written to the log file. My observation is that a lot of data can be buffered before it is written out to disk. ********************************************************************** ═══ 130. Install frustation ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None I've made several attempts now to get SPM/2 V2.0 installed with the API and the distributed feature and have run out of things to try. I think its safe to say that the installation instructions in the Quick Reference card -- this is the only documentation that comes in the box -- is completely, totally, woefully inadequate. This is where I've been (OS/2 is V2.1): 1) At the OS/2 window, I entered SPMINST /T:D:\SPM2V20 /API /DISTFTR. This ran without calling for diskette 2 (somewhat strange, but ?) 2) Shutdown and rebooted, as instructed. 3) Came up without any references to SPM/2 on the desktop. Conclusion is that the install was didn't run correctly. But it did create an API subdirectory under the SPM2V20 directory. My conclusion was that I'll have to run the install three times, once for the base product and an additional time for each feature. 4) Ran SPMINST /T:D:\SPM2V20 /UNINSTALL. Supposedly the product is uninstalled. But the directories, SPM2V20 and the API subdirectory, were left on D. 5) Shutdown and reboot. 6) Ran SPMINST /T:D:\SPM2V20. This time it called for diskette 2. 7) Shutdown and reboot, as instructed. Got the SPM/2 program folder on the desktop. Much better. Now I just need to install the distributed feature and the API. 8) Ran SPMINST /T:D:\SPM2V20 /DISTFTR. No dice. I got message SPM0138, telling me that a previous version os SPM/2 2.0 exists. I should uninstall the previous product. Now we already discovered in step 1) that the install won't do multiple things in one pass. So where do we go from here? There were a few other trials and failures but I think one should be able to get the idea of what's going on. At this point, I think I have the base installed. I would like to have the distributed feature and the API also installed. But frankly, I've already invested more time in this than ought to be necessary and I'm going to wait for some advice from development before I waste any more. I would appreciate it if SPM/2 support could give me an exact procedure to get from where I am to where I want to be. Tom Sander ********************************************************************** When I installed the distributed feature, there was no folder created. The Distributed Feature is a subset of the base system and only requires the first diskette. I have not tried to install the API so I cannot comment on that. Otherwise, it appears that everything you tried worked correctly. I am not sure if UNINSTALL should be removing directories or not. It probably would not in case you added other files to these directories, but this is just a guess. ********************************************************************** [Reference to append at 22:01 on 93/11/19 by SANDERT at STLVM1 ] I also could not install the API with SPMINST. I ended up doing it manually D: MD D:\SPM2V2\API CD D:\SPM2V2\API UNPACK A:\API\*.* The header files are readable, I suppose it is a good bypass. tanti saluti knuti ********************************************************************** [Reference to append at 22:01 on 93/11/19 by SANDERT at STLVM1 ] Sorry your Install was so frustrating. Here's some exaplanations. 1) SPMINST /T:D:SPM2V20 /API /DISTFTR worked and all the files are on the first diskette. Note that SPMINST /DISTFTR installs just the Distributed Feature SPMINST installs the full-function product including the DistFtr SPMINST /API installs full-function including DistFtr plus the API 2) 3) The Distributed Feature has no user interface, hence no desktop presence. 4) SPMINST /T:D:SPM2V20 /UNINSTALL uninstalls the full-function product but not the API. I guess you have to explicitly uninstall the api by specifying the /API option when you uninstall. 5) 6) SPMINST /T:D:SPM2V20 installs the full-function product, but not the API. However, your API is still installed! 7) 8) The DistFtr got installed in step 6, so trying to install it again fails. I think you have everything installed, so no additional action is necessary. If you install another license, just do: SPMINST /T:D:SPM2V2 /API ********************************************************************** Thanks for your response. I've now read some of the documentation and I tend to agree with you that I now, probably, have all three functions, base, distributed, and API, installed. Haven't tried to use it yet but we'll see. I sorta wonder about your response to the first item. I don't really think that the first try worked. My intent was to get all three functions installed in one pass. The reason I think that was unsuccessfull is that the program folder did not appear on the desktop after re-boot. And when I installed the base function alone (item 6), it did require the second diskette. As for your response to item 4, /API is not shown as an option on uninstall. Again, thanks for your response. But you folks do need to work on the install instructions in the quick reference card. Particularly if you insist on not shipping hardcopy pubs (which, by the way, is not universally accepted as goodness). Tom Sander ********************************************************************** [Reference to append at 16:25 on 93/11/22 by 61809747 at VIEVMA ] I had a system that had SPM/2 v2 previously installed and when attempting to install the API it requested that I uninstall first. I ended up following the same route as you did and just used the unpack command on the A:\API\ files and I got what I needed. Not pretty but it sure was easier on the mind. Is there a way to just install the API at a later date without uninstalling the product? Am I now missing some syslevel updates? Rich Hall ********************************************************************** The API option cannot be installed separately. It must be installed with the rest of the product. As noted in previous appends, either 1) Uninstall and re-install with /API, or 2) Unpack the API files from the installation diskette directly. In the CSD we'll either fix or document the fact that the API must be explicitly uninstalled. ********************************************************************** ═══ 131. Hardware dependencies ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * This is a generic question in order to understand whether SPM/2 and Theseus2 run properly for clones? Does SPM/2 get all its information from OS/2? In other words, can one assume that SPM/2 will work once the machine boots up under OS/2? John Tong Sydney, Australia ********************************************************************** SPM 2.0 is supported on all hardware that OS/2 is supported on, though we haven't actually tested much other than PS/2s. All the data comes from OS/2, none from the hardware. In some cases a new device driver will be installed to support various peripherals, and the DD will not have the performance hooks that standard OS/2 DDs have. For example, the OASES disk array requires a new DD to be installed that does not have SPM hooks, so SPM cannot provide Disk I/O performance data for it. ********************************************************************** ═══ 132. Cross Domain Monitoring ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I've a customer asking if he can *concurrrently* remote monitor all the additional servers defined to five different domains all at once, from the monitoring workstation which is on the another domain (the sixth domain). At the moment, he is successfull with remote monitoring within a single domain at a time but not cross domains. Is it correct that he has to setup a monitoring workstation to remote monitor the servers for each domain.. and therefore collecting the data concurrently! Customer environment incl. OS/2 2.1, LS 3.0A, and SPM/2 2.0. Lena Smith - IBM Australia. ********************************************************************** Cross-domain monitoring is supported by SPM 2.0. See the on-line reference for how to do it. It's in section SPM/2: Overview Monitoring a Network: Overview ********************************************************************** ═══ 133. quick questions ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 1. On my SPM report for my os/2 2.1 server (32Mb) I get the following values under PAGE : Demand loads=123, discards=3, faults=2849 and fault rate=0.18 On a 16Mb PS/2 I get values Demand loads=18,5k discards=50.9k, faults=80k and fault rate=4.9 and page in=34.4k and page out=17.6k (these were 0 on 32Mb server) Can someone briefly tell me what this means and what is an acceptable range of values. Should I be putting more memory in???? How do I know when I should put more memory in????? 2. The docm says increase max netbios session by 2 for each additional instance of SPMILOG. How do I know how many of these I have?? I'm just using SPM/2 to monitor a server. Thanks very much (Hafiz Ali) ********************************************************************** An old question that may not have gotten answered: |1. On my SPM report for my os/2 2.1 server (32Mb) I get the following | values under PAGE : | Demand loads=123, discards=3, faults=2849 and fault rate=0.18 | On a 16Mb PS/2 I get values | Demand loads=18,5k discards=50.9k, faults=80k and fault rate=4.9 | and page in=34.4k and page out=17.6k (these were 0 on 32Mb server) |Can someone briefly tell me what this means and what is an acceptable |range of values. Should I be putting more memory in???? How do I know |when I should put more memory in????? The PAGE stats in a summary report show the amount of work OS/2 is doing just loading programs and data. Your server has a low fault rate and no paging, indicating that most everything is in memory and 32 MB is plenty of memory. The 16 MB system on the other hand is paging heavily and may need more memory. The best way to determine if more memory is needed is to collect Working Set stats with SPM. If the Working Set for the 16 MB system is 14 MB or more, consider installing more RAM. |2. The docm says increase max netbios session by 2 for each additional | instance of SPMILOG. How do I know how many of these I have?? I'm jus | using SPM/2 to monitor a server. Each Log you start will create another instance of SPMILOG. A log can be defined to collect from 1 or more remote systems. If you're just monitoring one server then you only have one ILOG. ********************************************************************** ═══ 134. Netware Requester Coexistence ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am pleased to announce that there is now a fix available that allows SPM 2.0 to coexist with a Netware Requester on an SPM Monitored or Monitoring workstation. The fix is for OS/2, not SPM. Request the fix for OS/2 APAR IC05445 (or Defect 6670). ********************************************************************** Correction: For Netware Requester coexistence with SPM 2.0, the APAR number for OS/2 is PJ09970. Furthermore, it is not quite available yet. I will append a notice when it's available. ********************************************************************** >Correction: >For Netware Requester coexistence with SPM 2.0, the APAR number >for OS/2 is PJ09970. Furthermore, it is not quite available yet. >I will append a notice when it's available. With this be available with the service pak? Thanks! Matt Hickman Internet: bhic@chevron.com (713) 754-2389 ********************************************************************** ═══ 135. Writing page-fault detector ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I would like to write a small program to analyze the page-fault performance of other running processes under OS/2 2.x. I have used the SPM/2 (System Performance Monitor for OS/2) package successfully in the past. However, I'd like to write a utility which can be shipped to customers which only does a small subset of what SPM/2 does (e.g. page-fault detection). What would my best alternative here? I could write to the SPM API or write my own program. If writing to the SPM API means that some portion of the SPM package needs to be shipped with my utility (e.g. device driver, DLLs), then I would prefer the latter alternative. If it is the latter alternative, can someone point me to the right set of OS/2 control program functions that I may use (e.g. DosXXX) to implement a page-fault detector or perhaps to the correct fora? Thanks, Tony Lee ********************************************************************** The best option is to use the "OEM" API for the DCF that is shipped with LAN NetView Monitor/Enable. The LNV DCF is somewhat different from the SPM 2.0 DCF. It has it's own API that 3rd party vendors can and do use to collect the same performance statistics that SPM and LNV collect. I will send you the package for your evaluation. The LNV DCF must be included with your application, and there is a royalty fee for non-internal sales. ********************************************************************** ═══ 136. Remote Monitoring With OS/2 2.1 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * We have several machines running os/2 2.1 with Lanserver 3.0. SPM seems to work when run on the local machine. We're seeing two strange behaviors when trying to run against another machine on the Lan. 1. Choose Monitor->Setup->New. Hit the Query button. The 'Please wait....' dialog comes up for 20 seconds or so, and then SPM exits. Restarting SPM does an updated list of workstations in the Monitor->Setup->New->Available listbox. 2. After creating a LOG file for a remote station, try to make it active using Monitor->Start. The Sessionappears, but the session status never changes from Pending. The workstation status remains Off-line. DCF was installed on the remote system, the NetBios listener is us, and the remote system is found by the Monitor->Setup->New->Query. Thanks, Rick Rvalenti @ Wasvmic1 ********************************************************************** 1) Known bug. On OS/2 2.1 using the Query will bring down SPM Control. Get the SPM 2.0 CSD to fix this. 2) Is the remote system logged on? Are there enough NetBios resources on BOTH the remote system and the managing system? Just because the Query function locates a machine does not mean SPM can monitor it. The Query function simply sends out a NetBios broadcast message "anybody out there?". Any system that has NetBios loaded may respond if they happen to catch the broadcast. But SPM requires that remote systems also be logged on and the SPMNBL (NetBios Listener) be up and running/listening before the system can be remotely monitored. ********************************************************************** ═══ 137. CSD ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am pleased to announce the availability of a ServicePak (Corrective Service Diskette) for SPM 2.0. The ServicePak WR06075 is available (or will be very soon) through the following distribution points: OS2CSD OS2BBS(Talklink) CompuServe Internet PSSBBS Canadian BBS Australian BBS German BBS EMEA Hone DialIBM Diskette orders will also be made available from ISMC Boulder for CCSS orders. The package is available in the U.S. Version only. Everyone with an SPM 2.0 license is encouraged to take advantage of this free offering. More than 60 bugs have been fixed. New SPM 2.0 orders will receive a refreshed product that includes these fixes as well. ********************************************************************** The name of the SPM 2.0 CSD package on OS2CSD is WR06075A. ********************************************************************** [Reference to append at 17:59 on 93/12/10 by HOLCK at AUSVM1 ] I have just downloaded package WR06075A from OS2CSD. If I try to install it in my SPM/2 2.0 system I find these the steps needed: From (package WR06075A) first diskette 1) SPMINST /U 2) SPMINST It seems to me that these 2 diskettes could be used in a system with no SPM/2 2.0 installed before. Am I right?. Are the two diskettes of this package a replacement of SPM/2 2.0 original ones? Rafael Tesoro Madrid, IBM Spain ********************************************************************** When the SPM 2.0 CSD was first put on OS2CSD and other bulletin boards it was mistakenly the full product refresh version, not just the CSD. That mistake has now been corrected. Since you have already purchased a product license, you may replace it with the version you received. ********************************************************************** ═══ 138. Page fault definition ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * What is meant by a 'page fault' in SPM/2? Is it when memory is swapped or when a segment is not found in physical memory and needs to be brought in from disk (by the OS)? Thanks, Tony ********************************************************************** A page fault occurs WHENEVER a page is needed by an OS/2 process and it's not in memory. This includes: 1. Pages that have never been referenced before 2. Pages that have been moved from active pages in memory to the "idle list" in memory 3. Pages that have been swapped out to the SWAPPER.DAT file 4. Pages that have been discarded, such as code segments When a page fault occurs, the operating system takes the appropriate action to resolve it. This action may or may not result in a disk I/O. ********************************************************************** ═══ 139. Can SPM/2 monitor a 295? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I appended to the 295 forum for input on what alternate performace monitoring tools are availabe (other than MASS/2 which the customer I am helping is not satisfied with). A support person for the 295 said that SPM/2 will work on the 295 but that there is a big performace problem running it. If SPM/2 can work on the 295 could someone address how it sees the dual processors, and how SPM sees a RAID-5 disk array. The customer wants to know if the graphing would represent the combined performance of both processors and whether the disk array is seen as a single disk (my assumption). I would think that someone has already tested this or knows the answer off- hand. Thanks, Lee ********************************************************************** The PS/2 295 is an asymetrical multiprocessor machine. In this system, an OS/2 kernel is running on both processors in a master/slave relationship. The main processor is the Application Processor and the other is the Co-processor. My understanding is that SPM will collect performance statistics on the application processor only. LAN Server 3.0 Advanced is supported on the 295. LS Advanced has hooks for SPM 1.0 support and hooks for SPM 2.0 support. But since LS runs on the co-processor, the hooks are not automatically accessible to the main processor (the application processor). The hooks are still in LS but aren't accessible to the SPM data collection facility. The Boca and Austin development teams' focus will be to support OS/2 and SPM on SYMETRIC multiprocessor systems. Our design would track CPU utilization by processor, and track thread execution by processor. To determine how SPM would monitor a disk array, run FDISKPM. Each PHYSICAL disk reported by FDISK will be tracked separately by SPM. ********************************************************************** ═══ 140. Kill Process ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is there any way to kill a process via SPM/2 or THESEUS2? Ken Kahn IBM - Poughkeepsie,NY Staff of IBM Fellow: IBM Personal/370 Internet: KENKAHN@VNET.IBM.COM IBMMAIL: USIB25P9 at IBMMAIL CIS:71333,2526 ********************************************************************** There is no way to kill a process with Theseus2. We know it would be a nice function to have. You can do it with the internal tool PSPM2 on OS2TOOLS. ********************************************************************** Without getting too much into internal tools, I can't kill a certain process via PSPM2 because it doesn't properly show a background child process that is still active. However, Theseus/2 does show this child process. Ken Kahn IBM - Poughkeepsie,NY Staff of IBM Fellow: IBM Personal/370 Internet: KENKAHN@VNET.IBM.COM IBMMAIL: USIB25P9 at IBMMAIL CIS:71333,2526 ********************************************************************** ═══ 141. THESEUS2 - Working Set Display explanations ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I just ran working set for the system on my PS/2. I did this for 12 intervals with time between intervals of 5 seconds. I ran this for approximately 9 minutes. I am really confused at what the information really means. There is a 'now' , 'ws', and 'accessed' fields that I really don't know how to interpret. The documentation says the 'now' field is the number of pages referenced in the current interval, the 'ws' is the number of pages referenced in the past n intervals and the 'accessed' is the total number of pages accessed since the start of the collection run. The numbers are in the format of xx.xxxM. I am a little confused because I thought a page in OS/2 was 4K in size. So if I multiple the number of pages specified in the numbers given in these values by 4096...then I get an astronomical figure. For instance perhaps the values are as follows: 'now' = 10.688M' 'ws'=10.688M, 'accessed' = 10.688M. If I multiply these numbers by 4096 then I get 4096 X 10688000 = the actual number of bytes. So, what am I really looking at?? Should the 'now' value of 10.688M really be 10.688Mb?? Also, the final results I got were: Absolute minimum amount of memory: 12.621M Recommened amount of memory: 15.031M Total amount of accessed memory: 15.555M Analysis of SWAPPER.DAT after the same period of time was: Analysis of the SWAPPER.DAT file: There are 4864 disk frames in SWAPPER.DAT (each is 4K bytes). There are 3773 frames used. (15092K => 14.738M) There are 1091 frames free. (4364K => 4.262M) There are 4864 frames total. (19456K => 19.000M) < End of THESEUS2 (v 2.0.1a) output @ 13:17:38 on 12-15-1993 > Does this indicate that we have a problem?? How should I be reading this information? Thanks, Cynthia Berry 816-340-7973 ********************************************************************** To determine the working set for a steady-state system, use the "ws" column for the memory accessed over the last n intervals. The units can be pages, KB, or MB. You specify which you want under "parameters" in the working set pulldown. No problems are indicated by the swapper.dat info. OS/2 will shrink the swapper.dat file by removing the free pages. It does this when it determines the system is idle enough. ********************************************************************** [Reference to append at 19:00 on 93/12/16 by STW1CYC at OS2CUST ] Oh, dear! I fixed the "System Working Set" panel but didn't fix the "Process Working Set" panel when I added the option for displaying in Meg or Kilo bytes. Each of them should say "amount of memory" instead of "number of pages". So, the 10.68M is actually 10.68 megabytes of memory, not 10.68 millions of pages. (This will be fixed in the 2.0.1c version, which will be sent to OS2TOOLS after the first of the year, as this is my last day and I won't get through with this today.) You didn't say how much physical RAM you have on the machine. You also didn't say if the summary is a "Process Working Set" or a "System Working Set". However, a general rule that we have been trying to verify is: if (working_set_for_the_whole_system > 80%_of_RAM) you're_in_trouble; else probably_ok; Simply having a large swapper.dat file is not a sign of trouble. I have a 24M model 90. Right now I have a 22M swapper.dat file because I've loaded a lot of stuff, swapped a bunch out, then terminated the "ramhog". However, I'm running about 60% working set. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** The amount of RAM we have is 16MB. Actually, it is less than that because of the image adapter device driver we have loaded. It is probably closer to 15.5MB. So...are we really in trouble??? I need to take this to management, in an attempt to provide some capacity planning. Thanks, Cynthia Berry 816-340-7973 ********************************************************************** I would say (from the data you provided) that the working set is greater than your physical RAM. I don't know how much bigger. (Once it goes over about 85% of physical RAM then it is difficult to even guess.) If this is one of several machines, then take one of the machines to (at least) 24M and see what the numbers say. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 142. Summary reports using multiple times in same log file ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * BUG! I am having a terrible time with spm creating summary reports. I have over 50 summaries to run, but in 10 files. Averaging 5 timestamps per log file, it would be nice to be able to enter more than one timestamp per testcase. However, since that is not currently the case, I have been creating one RDF file, and selecting the first time, which works fine. Then, I open the RDF again, put in the new time and run the summary. It completes, but the file is twice the size of the first, and has all the data in there twice, with incorrect CPU usages (double) for individual events. The only way I have found to correct this problem is to open a new RDF file for each and every timestamp in the same log, and redifine my parameters. Is there any other way to fix this? I have installed the latest CSD (12/9), but it did not correct the problem Thanks, Derek C. Sanderson ********************************************************************** Answer to an old append, about setting up reports for multiple log files and multiple collections. I assume by "timestamps" you mean what we call "collection periods", each with a start and stop time. The SPM 2.0 Reports can support log files with multiple collection periods and can merge multiple log files. If your 10 log files don't overlap in time you could even do a single report for all of them. (If they do overlap in time, they were probably collected on different systems and a report would still work.) If you want to report each snapshot in a Summary report you can set the TimePeriods-SummarizationInterval to the the collection interval. If each collection period in the log is a different length of time, and you want to summarize for the length of the collection period, then separate reports is required. ********************************************************************** ═══ 143. TRAP 00E on monitored LS30A Server when IBM AV/2 starts ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have IBM LAN Server 3.0 Advanced Server with IP07001 and WR07020 applied. It has IBM AV/2 (1.03 I think) which is scheduled to run at 1:00 a.m. There is no other activity on the machine at this time (not a single user, no other scheduled activity other than DCDB replication). Only when I am monitoring that machine, at precisely the time AV/2 kicks off, the machine get the TRAP 00E error. It doesn't happen when I kick of AV/2 manually. Any ideas? Thanks, Jim Burnham BURNHAM at RHQVM01 ********************************************************************** ═══ 144. SPM and OS/2 SMP ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Will SPM work with the OS/2 SMP code ? We would like to do some performance comparisions between single and multi-processor systems. We currently have the SMP code that was delivered with the DEVCON CDROM up and running. Thanks Roger Faulkner, Edify Corp. ********************************************************************** ═══ 145. Overwriting .log file ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * When starting SPM/2 with the "SPMISTRT" command, is there a way to specify overwriting the .log file? When automatically monitoring a server, on a daily basis, seems the .log file will get very large, very quickly without the ability to overwrite. When starting SPM/2 from the PM interface, append/overwrite options are available, but how about with the command line? Thanks in advance, Rich Boehlke Dept:77MF TL:347-3372 Boulder CO ********************************************************************** Unfortunately there is no way to overwrite the data in the log file when starting it from the command line with SPMISTRT. The workaround is to save a backup of your new log file definition. Then whenever you want to start fresh, just copy the backup .log to your logfile before starting. ********************************************************************** ═══ 146. SPM/2 2.0 Hangs ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I just started to use SPM/2 and can not get it working. I am running OS/2 2.1 and am trying to use the SPM/2 2.0 Control monitor utility, however it keeps hanging in a way that requires a cold start. I noticed that another person had this problem as well, was his problem solved ? Note: I am not currently using THESEUS2. Karen Appleby ********************************************************************** ═══ 147. Help identifying processes in SPM/2 Report ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Can someone tell me what the following processes, from a SPM/2 summary report, are: EPWMP, EPWMUX, EPWPSI, EPWROUT, and sysinit (pid 001). Thanks. Dave LeGrys ********************************************************************** sysinit is the "system process". Things like the "page ager", etc. run there. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 17:56 on 94/01/06 by LEGRYS at RALVM14 ] EPW is FFST, or First Failure Support Technology. According to the OS/2 Glossary, FFST is "The OS/2 LAN Server component that OS/2 LAN Server or other applications can use for problem determination. Functions include logging and displaying errors, formatting and routing generic alerts, and generating data dumps." ********************************************************************** ═══ 148. Workstation status is always off-line ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am a new user of SPM/2 and have installed on two OS/2 2.1 workstations. Each has Lan Requester 3.0. I set up one as the monitor and created a LOG for the other. When I set up the LOG, I used the query button to show the active workstations and then selected the one I wanted from the resulting list. So why when I start the monitor does the workstation status stay off-line. What do I need to do to get it on-line? Another question, does it matter which workstation I bring up first, the monitor or the monitored? Mark Raine ImagePlus Development ********************************************************************** For System A to monitor System B, A must be more than a LAN Reqeuster. It must have at least Peer Services or be a LAN Entry or Advanced Server. The Query will display systems that have NetBios loaded and the SPM NetBios Listener (SPMNBL) loaded. But to be monitored they must also: be logged on to a domain, and probably have their netbios configuration upgraded. See the online Reference Guide for more information on Network Monitoring. ********************************************************************** ═══ 149. How to install Theseus2 if SPM 2.0 dist feature already installed ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello, We've got SPM2.0 DISTRIBUTED FEATURE installed on an OS/2 2.1 GA machine, working fine. Now I need Theseus2, (I don't need to run or monitor Theseus2 remotely... I have physical access to the PS/2). Is there some way to install just Theseus2 "on top of" a PS/2 which already has SPM2.0 distributed feature on it, without undoing, or resinstalling the distributed feature? How? Thanks! Message from: Gary R. Martin - Advisory SE (614) 225-3761 T/657 ............: 140 E. Town Street - 12th floor (Location JJA) ............: Columbus, OH. 43215-5195 ********************************************************************** Is this a customer machine or an IBM internal machine? James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 21:32 on 94/01/13 by GMARTIN3 at DETVMIC1 ] Ref: Append at 22:57:18 on 94/01/13 GMT (by SUMMERS at AUSVM1) James, We've got BOTH (IBM internal machine, and CUSTOMER machine scenarios). Message from: Gary R. Martin - Advisory SE (614) 225-3761 T/657 ............: 140 E. Town Street - 12th floor (Location JJA) ............: Columbus, OH. 43215-5195 ********************************************************************** [Reference to append at 21:32 on 94/01/13 by GMARTIN3 at DETVMIC1 ] You can't install just Theseus2, and you can't install it "over" the Distributed Feature. You must uninstall the Distributed Feature and then install the full product. (Don't bother to re-boot in between.) The full product includes Theseus2. You must have a license for the full product. Uninstall/Install doesn't take long, so just do it! ********************************************************************** ═══ 150. Dump report problem (with user metrics) ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Subject: Dump report problem (with user metrics) From: Mani Arjunamani (VTMANI at SYDVM1) Hello, I am a new SPM/2 user. I am trying to use it to measure the performance of an application under development. I followed the example in the on-line reference and coded a sample program with a counter, a bulk counter and 3 timers.Then I started the SPM/2 monitor session and then the test application and recorded data for sufficiently long time. The timer is just timing a huge for-loop. Then I terminated the test application which de-registers before terminating and then created a Dump report. The dump report shows correct values for 'counter' and 'bulkcounter' but the timers(I have 3 of them in all) are all having ZEROs only. When I check the value throug the debugger, each timer has a large value in its QWORD struct. from the IPMD debugger, it shows a large value in the QWORD variable. I tried SPMStartTimer() - SPMStopTimer as well as SPMQueryTime() with SPMAddQword(), SPMAubQword() combinations to compute the time difference. The time difference is calculated correctly. I do not understand why the timer values are all appearing as 000:00:00:000.( I am sure the timer values calculated by the application are atleast several hundred milli-seconds) Could any one tell me what I am doing wrong. Note: I am using the ordinal for the group as 202 in the metric definition file With thanks in advance. Mani Arjunamani ********************************************************************** From the symptoms you describe, everything seems to correct except the final dump report. The timers should have large values in them. An SPMStartTimer will subtract the current time of day from the timer, and an SPMStopTimer will add the current time of day to the timer. The SPM Data Collection Facility will snapshot the values of the timers any time it wants, and correctly calculate average time between Start and Stop. The Dump report should have your timer names as a column headings. Send us your source code (HOLCK at AUSVM1) and we'll check it out. ********************************************************************** ═══ 151. Dump report problem ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi Timothy Thank you for your quick response. I have alreay sent the files to your reader(HOLCK AT AUSVM1) Looking forward to hear from you. Thanks. Mani Arjunamani ********************************************************************** I've reviewed your code and have a couple of comments although I'm still not exactly sure why the timers are all zeros but try this and see how it goes. 1. In the code you sent us, timer2 and timer3 do not appear to be in use. 2. Just after registering your user hooks, you do a SPMSetSem. This mutexs the user hooks and prevents us from collecting data about them while you modify them. You do not release the semaphore with a SPMClearSem until just before deregistering your group. To get more than one entry in the log file, use these routines only around where you actually modify your hooks NOT around the entire program. I understand that our sample program does this but it is a short running program. More sample points could give us more information about what is happening to your timers. -Mark Molnar SPM/2 development ********************************************************************** ═══ 152. Measuring response time ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi, We are trying to find a way to measure the response time of an application on OS/2 2.1. Is there a way to do that with SPM/2 ? If not, does anyone know of another product to do that ? Daniel Larouche, ISAS, Quebec City, Canada. DLAROUCH at TOROVM1 ********************************************************************** SPM 2.0 can measure AVERAGE response time. Only an event-based or trace tool can measure individual response time. A program with calls to DosGetDateTime can measure program response times, but that OS/2 call only has a resolution of 32 milliseconds. A REXX program with calls to Time could measure more end-user-like response times. To use SPM 2.0, you must program "user hooks" into the application using SPM'2 User Hook API, and then recompile. For each periodic interval, say 60 seconds, SPM will report the average response time. ********************************************************************** ═══ 153. Setup for LOCAL, CMDline args, No Page Table, RAM Graph ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 1. First-time startup for LOCAL (which I suspect would be the most-used method) is pretty complicated what with the need to define (and understand) LOG files and everything; I recommend a fast-path for this simple usage. 2. Is there a way to bypass the first window for starting LOCAL and going straight to the graph? Perhaps some commandline arguments? 3. When I start Theseus2 and click on RAM USAGE BY PROCESS, it takes a really long time to start and I get a window with thousands upon thousands of lines all identical to these: RAM Usage by Process: >>> Linear address: no Page Table, linear=00000000. >>> Linear address: no Page Table, linear=00000000. >>> Linear address: no Page Table, linear=00000000. I have to scroll all the way to the end to see anything interesting. What gives? 4. On SPM/2 Graph, my RAM Graph shows 31.6MB on the title bar, but it shows Used Memory, Working Set, and Resident Memory all way up there near 100%, even if I start SPM right up before doing anything else interesting in OS/2. My (admittedly pathetic) understanding of what I'm seeing in Theseus leads me to believe that I shouldn't be near 100% "full" in any sense. Here's the bottom lines from Theseus Ram Usage by Process... 00000000 0 0.000 00005000 20 0.020 CMD 0005E000 376 0.367 00009000 36 0.035 THESEUS2 -------- ------ ------ -------- ------ ------ 00CD6000 13144 12.836 012C9000 19236 18.785 total RAM in use 00000000 0 0.000 free RAM -------- ------ ------ 01F9F000 32380 31.621 total of all RAM pages found (Pvt + Shr + Free < End of THESEUS2 (v 2.0r) output @ 14:48:50 on 1-20-1994 > I am hoping to use SPM/2 graphs to give me feedback to help me determine when it's time to shut down some apps if I'm about to fire up a mips/ram- hungry application. But if the guages are always around 100% "full", it doesn't really help me in this way. ********************************************************************** 3. You have a VERY OLD copy of THESEUS2. Please get the current version off OS2TOOLS. It has this problem fixed. 4. Your THESEUS2 output says all your RAM is in use; therefore, the "Used Memory" line should be at 100%. (It is normal for OS/2 to "use" anywhere from 20M up, depending on what all is loaded.) The other things could be artifacts of your old programs. See what it looks like after applying the new version. BTW, you ought to get the SPM/2 CSD, also. If you get it, put it on BEFORE applying the THESEUS2 update, as the T2 update is later than what is in the CSD. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 20:45 on 94/01/20 by KELLEY at YKTVMH ] > >2. Is there a way to bypass the first window for starting LOCAL and going >straight to the graph? Perhaps some commandline arguments? > Please see the SPMISTRT command and the SPMMON command in the online documentation. -Mark Molnar SPM/2 development ********************************************************************** ═══ 154. SPM2 Stops collecting data at lan server startup time ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * SYSTEM: OS/2 base XR02010 CM 1.1 WR06000 NTS/2 WR07020 SPM/2 WR06075 Lan Serv IP07001 (ENTRY LEVEL) I start SPM graphing and recording as the first task in my system startup. The last function that I start is Lan Server Entry Level. As soon as the server completes its statup, SPM/2 stops collecting any data. The spm control menu still shows that everything is graphing/recording, yet the collected data and the graph show nothing after the server start time. When you stop recording and graphing, the control panel go to a pending message. If you enter stop, you get a message about some process that is supposed to time out...but is evidently not! Any ideas whats happening. Is there some function/tuning I may need to do to make SPM work with the server. IF you think it's a defect, let me know and I'll open one up. Terry A. Steilen - Advantis ********************************************************************** In <940121145647.os2spm20.cforum.ibmpc.tsteile@carvm8.vnet.ibm.com> tsteile@carvm8.rscs writes: >SYSTEM: OS/2 base XR02010 > CM 1.1 WR06000 > NTS/2 WR07020 > SPM/2 WR06075 > Lan Serv IP07001 (ENTRY LEVEL) > >I start SPM graphing and recording as the first task in my system startup. >The last function that I start is Lan Server Entry Level. >As soon as the server completes its statup, SPM/2 stops collecting any data. >The spm control menu still shows that everything is graphing/recording, >yet the collected data and the graph show nothing after the server start >time. When you stop recording and graphing, the control panel go to a >pending message. If you enter stop, you get a message about some process >that is supposed to time out...but is evidently not! > >Any ideas whats happening. Is there some function/tuning I may need to >do to make SPM work with the server. IF you think it's a defect, let >me know and I'll open one up. > >Terry A. Steilen - Advantis > Terry, This is a known issue. I say an issue instead of problem because I think we document this phenomenan somewhere but anyways the answer is you must start SPM after LAN Server or LAN Requester. Without alot of detail what happens is that when SPM/2 is started we don't find a LAN environment so we configure essentially local only. Then you start the LAN and we become confused. As I said, just move us after starting the LAN and you should be set. -Mark Molnar SPM/2 development ********************************************************************** ═══ 155. PACK utility to enable customer to build SPM install disk ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * One of my customers wants to create a pre-defined SPM/2 install disk. At the moment it doesnt work with OS/2 V2.1 because of the problem described previously in the forum - the OS2DASD.DMD gets replaced by a back level file version during the install. He wants a slicker way out of the problem (better than having to rename OS2DASD.OLD back to OS2DASD.DMD). He asked me if I could get hold of the PACK utility so he can replace OS2DASD.DM@ on the SPM install disk with the correct one. I cant find the PACK utility anywhere. Can someone supply it to me please? Thanks - Jerry Client Systems Consultant IBM UK Insruance Business, North West Unit ********************************************************************** There is a PACK.EXE supplied with the OS/2 2.X toolkit. This is the one OS/2 uses. I'm not sure if it is the one SPM/2 uses (I haven't got the disks in front of me at the moment to try). I'll send it to you to try... Scott Atkinson TeamOS2, IBM Support Centre, Sydney, Australia ********************************************************************** ═══ 156. SPM/2 documentation ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I saw a previous append where someone was sent the SPM/2 "manual" electronicaly. I have an SE who needs a soft-copy. Please send it to SKALTER at ATLVMIC1. Does this soft-copy have a different format than if you printed out the on-line documentation (maybe page numbers?). I thought that SPM/2 did not ship with a manual. Is this true? If the soft-copy "manual" is any different than a print-out of the on-line help then could I also get a copy: INTC7B15 at DALVM41B. Also is there any other means by which a customer or SE can order or down-load the documentation? Please send SPM/2 and Theseus/2 docs. Thanks, Lee ********************************************************************** You can request LIST3820 versions of all the SPM 2.0 documentation by typing the following on VM: REQUEST SPM20DOC FROM LSYSINFO AT AUSVM1 ********************************************************************** I tried this from the IBMLink prompt and was told REQUEST is an unknown command. How can we IBMLink customers obtain these files? Also, is the format compatible with LP3820? Dave Scheffer -*- VHA, Inc. -*- Information Systems -*- 214-830-0330 ********************************************************************** ═══ 157. Data Collection in a Wide Area Network ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * How can I collect data in a wide area network? The customer has 10 remote LAN's with a lot of servers. All of this LAN's are connectet via a SNA/APPN network. There is no LTLW installed to have a NETBIOS connection to the remote LAN's. The customer would like to have a APPC connection to collect the performance data? Is there any way to do that? Thank you for your help. Hans J. Sarny ********************************************************************** In <940127143937.os2spm20.cforum.ibmpc.61811381@vievma.vnet.ibm.com> 61811381@vievma.rscs writes: >How can I collect data in a wide area network? The customer has 10 >remote LAN's with a lot of servers. All of this LAN's are connectet >via a SNA/APPN network. There is no LTLW installed to have a NETBIOS >connection to the remote LAN's. >The customer would like to have a APPC connection to collect the >performance data? Is there any way to do that? >Thank you for your help. >Hans J. Sarny See the lannetvu cforum. There are good suggestions to solve this problem using LAN NetView, LAN NetView Monitor, and AnyNet. Look back about 2 months. -Mark ********************************************************************** ═══ 158. SPM vs Pegasus ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * A company names C.O.L. Systems Inc. sells a program called Pegasus, which appears to be yet another performance monitoring tool for OS/2. It interfaces with IBM's own Theseus and the company claims to have accordingly acquired the right to distribute the Theseus library (this seems to be only an option, though). I wondered whether anybody has heard of that tool and might be able to draw a comparison between the two packages. The version # for Pegasus as currently advertised is 1.9, which means it has been around for some time already. Thanks beforehand Olivier Lefevre ********************************************************************** Pegasus from C.O.L. is a standalone shareware product available on Compuserve and other bulletin boards. It is currently still under development. It collects and reports the same OS/2 metrics as SPM 2.0. C.O.L.'s flagship product is called OSRM2, network-enabled performance management software that also collects the OS/2 metrics. C.O.L. currently sub-licenses the IBM SPM 2.0 Data Collection Facility for OSRM2. Pegasus uses its own dcf, and eventually so will OSRM2. Pegasus requires Theseus2 executables for some memory analysis; users can purchase Theseus2 executables from C.O.L. for about $75, or get it by purchasing the full SPM 2.0 product. -- Information obtained from C.O.L. Corporation. ********************************************************************** ═══ 159. SPM Log File Space ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is there any to estimate the amount of Log File disk space that may be used ...for any given configuration? based upon time.... ...that is not based on any other external conditions. I'm looking for an answer that is not: The more files that an applications open, the bigger the log file gets! or more disk activity, more log file space! Is there any way to estimate the amount of log space needed for a 24 hour monitor run? Thanks, Terry A. Steilen - Advantis ********************************************************************** To estimate the log file space required for a 24-hour monitoring session with the Thread and File resources, I suggest monitoring for an hour beforehand, closing the session, looking at the log file size, and multiplying it by 24. (The true size isn't apparent until you stop the session.) The log file size depends on many things, including (in order probably): Resources -- threads and files will both increase log file size a lot. Collection Interval -- a 10 sec interval will grow the log file about 6 times faster than a 60 sec interval. The activity on the system when collecting threads/files -- Each thread creation/deletion and file open/close will grow the log file. The number of resources -- Collecting all resources logs more data than just a few. Please remember to keep a fairly short collection interval if collecting threads or files on a powerful busy system, since the thread/file birth/death rate can overwhelm SPM DCF. See the SPM 2.0 CSD Readme file for more info. Try 15 seconds. ********************************************************************** ═══ 160. Memory problem after install of WR07006 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have a 9577 with 32 Mb of memory. I am running OS/2 V2.1, Lan Requester 3.0, Lan Network Manager 1.1, and 4 monitoring sessions of SPM/2. I have all known fixes on it (less the current SPM/2 fixes which I think I might have seen) I have applied 7001, 7006 and the Theseus2 fixes. My problem is memory. The machine has gotten very slow lately (since 7006???). I looked at Theseus2 and it says I am using 25.6Mb of memory for 'system' 'private' and another 1.2Mb for 'system' 'owned shared'. WOW. My swapper.dat is 16 Mb as well. Looking at the details for this memory shows 6402 entries of: (Address) ######## is assigned as SYSTEM (private) and 328 entries of: ######## is assigned as SYSTEM DLL. What is hogging all my memory? Does this sound like a bug and should I contact Defect Support? Is there a way to collect more information? Is this a problem with LMU? Thanks for your help. ********************************************************************** ═══ 161. Capturing CICS OS/2 Performance Data ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am trying to do performance monitoring and benchmarking for a CICS OS/2 based client/server application. To include the CICS OS/2 generated data I have followed the procedure documented on page 37 of the recently published CICS OS/2 V2.0 Sizing and Performance (GG24-4190) book. I see that CICS OS/2 appears as a resource on the AVAILABLE list of SPM/2 Monitor Session Resources which I select and ADD to the Selected Box. However none of the report formats show any CICS OS/2 data. I have also included the /P option on CICS OS/2 startup command. These should have resulted in capturing and recording CICS OS/2 performance data. What else needs to be done? The other problem I am experiencing is that the printed output of the summary format truncated half of the information; there does not seem to be options on SPM2 Print panel to allow me to ask for a landscape format or reduced font. Your help in resolving these issues is greatly appreciated. ********************************************************************** I have placed this append on the CICSOS2 conference disk also. If you installed the SPM 2.0 CSD, did you wipe out (replace) the SPMAPP.INI file? If so, you need to recreate it for the CICS hooks by re-running SPMAPPIN. I will send you a test program called SPMLOGPR, which you can use to format your log file and see if any CICS data ever even got collected. Redirect the output of SPMLOGPR to a file. For printing reports, SPM Reports-Run automatically inserts a 12-pitch printer control character at the beginning of reports. Since all reports are 132 columns wide or less, they should fit on 8.5x11 paper. Maybe your printer doesn't support 12-pitch. You can also: 1) Use an editor to insert printer pitch control characters in the report. Try 17-pitch. 2) Try printing from a word processing program that allows printer pitch manipulation. ********************************************************************** The 12 pitch printer setting is for IBM/Epson type printers. If your going to HP or Postscript, or something else, then you will have to change the report before printing it. Terry A. Steilen - Advantis ********************************************************************** ═══ 162. SPM0296 with SPM/2 on OS/2 2.1 with LS 3 (adv) and DB2/2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None Help! I am trying to get SPM/2 2.00.1 to monitor a DB2/2 server (running OS/2 2.1 with LS 3.0 advanced) but with no success. This machine is an "additional server" and we want to do the monitoring on the machine itself (ie: monitoring & recording on the same machine). When we configure SPM/2 and use the local name for the workstation we get an SPM0296 error when we try to start monitoring! This error says that the network is not working -- yet we know it is! Any help would be much appreciated! BTW we tried using LOCAL for the workstation name, but this caused other errors (SPM0531 RunDCF: unable to read logon pipe). David Twyerould ********************************************************************** 1 - Is LR or LS started? 2 - Logged on to a domain? 3 - Enough netbios resources? My guess for this one is that you aren't logged on to a domain. Being logged on locally isn't good enough. ********************************************************************** Thanks for your response, in answer to your questions: 1 - Is LR or LS started? YES 2 - Logged on to a domain? YES 3 - Enough netbios resources? How can I tell?? We are Lan Server and requester is started and we are logged on both locally and on to the domain. How can I tell if I have enough netbios resources? There seems to be plenty from what is shown when the machine starts up (ie: Netbios resources available for applications...). David Twyerould ********************************************************************** The Netbious requirements for both the monitoring and monitored systems are stated in the SPM 2.0 online reference. See: Overview Monitoring a Network Network Monitoring Requirements For this problem and others like it, the general method for debugging the problem is to use the SPM error log. Error messages are logged in SPMSNAP.ERR. There are three levels of error logging, Error, Warning, or Informational. The default is Error and has only critical errors; Informational will cause a lot of stuff to be logged. The level is controlled with an OS/2 SET command, which is put in CONFIG.SYS at install time (an * is a backslash): SET SPMSNAP=E C:*SPM2V2*SPMSNAP.ERR W I The error log can be formatted into something readable by using the SPMSNAPP command: SPMSNAPP > SNAP.OUT For this specific problem, check the snap log. If there are no clues change the error level to Informational and check it again. ********************************************************************** ═══ 163. SPM/2 seeing OS/2 Memory on Compaqs ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Machine #1: Compaq Prolinea with 28mb installed. SPM/2 and Theseus/2 report seeing: 15.6mb. Machine #2: Compaq (Deskpro 4/66i) with 32mb installed. SPM/2 and Theseus/2 report seeing: 15.6mb. OS20MEMU reports seeing: 25mb virtual machine size. SPM/2 is at service level: WR06075. 21GT16 fix for APAR PJ08805 not applied (do not think it applies)? OS/2 2.1 GA no CSD applied. Machine #1 is running: OS/2 2.1, CM/2 1.1, and LAN Mail Gateway code with a swapper.dat size of 20mb. Is OS/2 using all the physical memory on the machines? Is there a problem with SPM/2? Suggested course of Action? Thanks for the help, Lee ********************************************************************** I believe that SPM/2 is reporting the amount of RAM being used by OS/2. You probably need the 21GT16 patch. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 164. QUERY is greyed out ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have just installed SPM/2 on a machine that is running LAN SERVER 3.0 ADVANCED. When I try running SPM, the QUERY button is greyed out. If I enter a workstation address and start monitoring, a connection is never made. Any help would be appreciated! Rick Suntag (!rjsuntag@attmail.com) AT&T EasyLink Services ********************************************************************** ═══ 165. Remote LAN segment ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I can run SPM/2 remotely on the local LAN segment, across a local bridge, but NOT across a remote bridge. After "sniffing" this with Sniffer, I find that I never received the initial "NETBIOS Name Recognized". Any ideas? Help! Bob Franklin Illinois Power ********************************************************************** I don't know why your SPM won't work over a remote bridge, but it sounds more like a problem with the Netbious setup on the bridge. Try Query, and make sure you can share a server resource over the bridge. ********************************************************************** ═══ 166. SPM/2 questions ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Having SPM/2 quite extensive in the past two months at a customer project, there are two things that I hope someone can explain to me. (1) How is that the total CPU under summary not always equal to the sum of the components? (2) Working Set is defined to the amount of memory accessed in a given period of time. Why is never greated than the amount of available memory? John Tong Sydney, Australia ********************************************************************** (1) The reason that the CPU Utilization by process or thread doesn't add up to the system total is that the measurements come from two different places. There is a data block in OS/2 for each thread which contains the performance information. SPM/2 looks at each data block to get the individual thread CPU Utilization. Then it looks at the system data block to get the system CPU Util. But time is ticking while SPM/2 is looking at all these data blocks. On OS/2 2.1, SPM/2 looks at the threads first and then the system total. So by the time SPM/2 looks at the system total, more time has gone by since the first thread value. So the system total should be slightly bigger than the sum of the thread totals. On OS/2 2.0, this is reversed - SPM/2 looks at the system total first, so the system total is slightly less than the thread totals. (2) Working set is the number of physical pages touched during a given period of time. So it has to be less than or equal to the total memory. You can't touch pages you don't have! ********************************************************************** ═══ 167. LS 3.0 CSD7001 fails after SPM/2 V2 installed! ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * HELP! After installing SPM/2 V2.0 plus CSD WR06075A the LS fails with error NET3060 or NET3087. Backing out to previous CONFIG.SYS allows Server Services to be invoked. Has anyone experienced this problem? Is there a fix? Environment Summary: HDW: PS/2 MOD 95,16MB RAM, 2 400MB HD MC32 LAN Streamer card (latest device drivers) SFTW: SPM2/2.0 CSD06075, LS 3.0 CSD7001 (ADV), OS/2 2.0 CSD6055 NTS/2 LAPS NOTE: For IBM internal see Problem # 1X076, PSP Thanks in advance! Tim Parker ********************************************************************** We found another 2.0 box with slightly different sftw mix and found out that without SABER and Lexmark software the problem does not occur! We can install SPM2/2 just fine and LS 3.0 (ADV) comes right up! So what should one look for? Thoughts and suggestions are appreciated since our customer expects to use SPM/2 to measure the performance of their 100+ servers which are presently at the 2.0 level with this extra sftw! - Tim Parker ********************************************************************** We have discovered a workaround to this problem! Move "START SPMNBL" to STARTUP.CMD, then LAN SERVER fully initializes and the LEXMARK REDIRECT.EXE task starts prior to SPM initialization. An incompatability appears among LAN SERVER/REDIRECT.EXE/SPM.EXE, if REDIRECT.EXE and SPMNBL.EXE are both started from CONFIG.SYS. Hope this helps anyone else falling into the same trap! - Tim Parker ********************************************************************** ═══ 168. Swapper information needed ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: New Append My system has 32 meg memory. My Working set interval is 5 minutes. My working set is holding at 16.3 meg or so. Pagein and out as well as DASD performance to my swap drive is showing some activity. If my working set is not changing, then why am I apparently paging? I'd really like to know where I can get educated in trying to understand this problem myself. Can anyone offer any good books and or courses that may help me in this quest? Thanks in advance, Terry A. Steilen - Advantis ********************************************************************** Do you mean the "Workin Set Period" is 5 minutes? Or the sample interval? Even though the working set is less than your total RAM, you will still swap if the TOTAL memory requirements is greater than 32M. My machine typically runs 50%-75% of total RAM and swaps occasionally. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** The interval is 30 seconds, and the WS period is 5 minutes (10 intervals) Is there any good documentation on how to analyze this information and put it to good use? Terry A. Steilen - Advantis ********************************************************************** The shorter the sample interval (but not too short because of the overhead caused by the data collection/reduction) the better the results are. Pages allocated and freed within the sample interval are not seen. This could be part of the reason why your numbers seem smaller than expected. However, remember that some swapping is normal and to be expected. The TOTAL memory in use by most system is more than the physical RAM on the system. For instance, my machine has 24M of RAM. However, I'm currently using a total of 27.4M (if I add it all up correctly). But my working set is running about 11M (based on the SPM/2 graph). James Summers (OS/2 Performance & THESEUS2) ********************************************************************** James, I'll be a bit more specific in my request this time. Could you tell me the definition of the ordinal # 5 counters, and how they relate to the displayed PAGEIN and PAGEOUT counters. I think that I may need the ordinal 5 data to analyze my situation, but unless I know what it means, It's difficult to determine what,if,or how, I should use it! Is this stuff documented anywhere? Discarded, Recovered, Idled, Reassigned,Freed, Used....? I think only you could help! Are some subsets of others? thanks, Terry A. Steilen - Advantis ********************************************************************** Ordinal #5 is the OS/2 Memory Group. The individual metrics are defined in the online Reference, under Using the Data Report Output Using the Metrics Dump Report Metrics. There is a lot of information about OS/2 memory and how it works in the online Theseus Reference, under Memory Explanations Memory States Virtual... It's true, the relationship of the memory metrics has not been fully explored/explained. Maybe I'll try to do that in the future. Use the following diagram as you go through the definitions. OS/2 MEMORY MODEL |------------------ | Free | | |------------------ | Used | | | !========! | !Resident! | | ! ! | !========! |------------------ ctIdled ctReassigned - - - - -> Idle - - - - -> List ctRecovered - - - - -> ********************************************************************** ═══ 169. Novice--subscribe ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: -none- I tried to subscribe to OS2SPM20 CFORUM via a TOOLS request here in Boca, but TOOLS said they couldn't do it. What do I need to do to subscribe?? Dick Geisler Zip=1214 ********************************************************************** To subscribe, type TOOLS. Change the tools disk to IBMPC. The Request is SUBSCRIBE. The filename is OS2SPM20 CFORUM. ********************************************************************** ═══ 170. Can 2 sep. SPM/2's ea. meas. ea. other @ the same time? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Have SPM/2 2.0 on M90 and M85, 1 is OS/2 2.0 other is OS/2 2.1 for addl. hdw/sftw refer to earlier append re: LS 3.0 (ADV) fails etc. When we start graph/report of box-b using SPM on box-a we can not start similar process on box-a to measure box-b! Why would anyone want to do this you ask! Well our assumptions are; 1) We should be able to check on any server from any other server. 2) We plan to run LMU/2 along with SPM/2 and if this does not work for SPM/2 it may not work for LMU/2! Thanks, Tim Parker ********************************************************************** You can definitely have two systems monitoring each other. Make sure they aren't also monitoring themselves, that you have enough Netbios resources for both sessions at each system, and that you can monitor a->b only and b->a only. ********************************************************************** ═══ 171. SPMV2 with OS/2 V2.1 and LAN Server Adv. 3.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None From: Tim Line To: Ibm * Original Area: GENERAL * Original To : Spm2v2 @ Ausvm1 (87:44/256) I like many others am having extreme difficulty in starting Monitor sessions with SPMV2. I have scanned all the messages in the SPM2V2 Quick Message area, but although the symptoms match, the error messages do not. My Monitor session goes first PENDING then disappears altogether with messages as follows:- "WAITING FOR ILOG" in the Monitor window SPM0162 API return code 1 function code 20 function return code 29 SPM0224 SPMILOG: unable to locate SPMILOG instance area for LOCAL.LOG I have read all the "Requirements for Monitoring a Network", and have set number of NETBIOS sessions, commands and manes to 200 in PROTOCOL.INI and MAXCONNECTIONS, MAXUSERS and NUMREQBUF to recommended values in IBMLAN.INI. IBM LAN Server Adv is at CSD level IP07005, and the OS2KRNL is sized at 736074 bytes dated 28/09/93. The SPM2V2 is at level 2.0.1 with CSD WR6075 downloaded from CompuServe applied. THESEUS2 works fine. What is wrong with SPM2V2???? What more fixes do I have to apply??? * EMEA OS/2 Developer Assistance Program BBS * gHostNet(0057002C01000000 2d652264) reply to (0057002C01000000 2d650fbe) ********************************************************************** In <940218151714.os2spm20.cforum.ibmpc.emeadap@winvmb.vnet.ibm.com> emeadap@winvmb.rscs writes: >Ref: None > >From: Tim Line >To: Ibm > >* Original Area: GENERAL >* Original To : Spm2v2 @ Ausvm1 (87:44/256) > >I like many others am having extreme difficulty in starting Monitor sessions >with SPMV2. I have scanned all the messages in the SPM2V2 Quick Message >area, but although the symptoms match, the error messages do not. My Monitor >session goes first PENDING then disappears altogether with messages as >follows:- >"WAITING FOR ILOG" in the Monitor window >SPM0162 API return code 1 function code 20 function return code 29 >SPM0224 SPMILOG: unable to locate SPMILOG instance area for LOCAL.LOG >I have read all the "Requirements for Monitoring a Network", and have set >number of NETBIOS sessions, commands and manes to 200 in PROTOCOL.INI >and MAXCONNECTIONS, MAXUSERS and NUMREQBUF to recommended values in >IBMLAN.INI. IBM LAN Server Adv is at CSD level IP07005, and the OS2KRNL is >sized at 736074 bytes dated 28/09/93. The SPM2V2 is at level 2.0.1 with CSD >WR6075 downloaded from CompuServe applied. THESEUS2 works fine. What is >wrong with SPM2V2???? What more fixes do I have to apply??? > Tim, What appears to be happening is that were having a semaphore timeout (30 seconds). If the semaphore times out before SPMILOG.EXE can be started and initialized, The SPMISTRT.EXE program removes the entry from the control panel and generates errors 147 and 162. Then when SPMILOG.EXE looks for the entry in the control panel and does not find the entry, it generates the 224 error and quits. The SPM.EXE (control panel) kicks off SPMISTRT.EXE which kicks of SPMILOG.EXE. There is no way for you to adjust this timeout value so you may want to try and reduce the load on your system. One thing you might try is starting recording and graphing separately. I've recreated this in our lab by trying to start 4 log files for both graphing and recording on a mod 80, 386, 20 Mz with not enough memory. We will look at adjusting or removing the timeout value for the future. If this does not sound like your problem please re-append with more information about your hardware and the timing of the events that you describe. -Mark Molnar SPM/2 development ********************************************************************** FYI: If your OS/2 system has more than one LAN adapter, and you are using SPM 2.0, you should be aware of a restriction. On either the monitored system or the monitoring system, SPM will use only the first LAN adapter that has Netbios support. If adapter 0 has Netbios support, SPM will use it, and cannot use adapter 1. ********************************************************************** [Reference to append at 15:14 on 94/02/18 by EMEADAP at WINVMB ] I have received EXACTLY the same errors ... as follows : 1) Running on OS2 2.1 BUT MONITORING LOCAL (I'm running on a LAN, but just want to monitor my own PC) 2) I'm running on a Model 70 with 12 MB memory. The system is not too heavily loaded...with a 8 to 10 MB swap file. 3) I start a monitor session for LOCAL measuring the extended info (application/process/thread) every 10 seconds... 4) I get a pending (I JUST start recording) and then SPMISTRT: File D:\SPM2V2\LOCAL.LOG verified but collection may not have started SPM0162 API return code 1 function code 20 function return code 29 SPM0224 SPMILOG: unable to locate SPMILOG instance area for LOCAL.LOG I saw the append regarding semaphore timeout...but I'm only running the minimum number of processes...and am only recording. I have tried this with WR06075 applied and without - with same results. David Kerr - Advisory S.E. - Insurance Services and Solutions IBM Canada - Toronto, Ontario - (905) 316-6664 ********************************************************************** ═══ 172. Starting/Stopping SPM/2 functions automatically ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I've read a few entries from the 93/11/09 timeframe with reagard to starting and stopping SPM/2 functions automatically. I'm familiar with the line commands and can intuitively see how to control the functions with a REXX program. Q: has anyone already written these "control pgms" that would like to share'em? Time is money and I'd like to save both. thanks, Matt Marinak voice:407-796-2506 email:marinak@pwfl.com Renee A. Nys Pratt & Whitney 407-796-3692 ********************************************************************** ═══ 173. Paging ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi, does the paging rate include programs pages which are not swapped to swapper but just trashed and reread in from DISK? Thanks very much. Ted Higginbotham Software Specialist PSP 1-905-316-4832 ********************************************************************** When more memory is needed by OS/2, it makes room by swapping data pages to swapper.dat and trashing or discarding code pages. Code pages can always be re-loaded from the original EXE/DLL file. Page Out = data pages written to swapper.dat Page In = data pages read from swapper.dat + code pages re-loaded from EXE/DLLs Discards include code pages that are thrown away. ********************************************************************** ═══ 174. Automating SPM2/V2 reporting: .RDF contents inflexible ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Well, I've written a REXX pgm that starts SPMISTRT logging. Once a day the pgm SPMISTOPs the log process, starts SPMRDF to generate a report file, copies a "skeleton" .LOG file over the old (full) logfile, and restarts SPMISTRT. I'd intended to take the existing report name .SUM, for instance, rename it to J.SUM (ie J94365.SUM) in order to maintain an archive of server performance reports. Unfortunately, the "Time Periods to be Included in Report" appear to stay fixed from the time the .RDF file was created (with one instance of a .LOG file). Does anyone know how to use one .RDF file upon same-named .LOG files with differing contents/collection-periods ? thanks, Matt Marinak / Pratt & Whitney (407)796-2506 ********************************************************************** Set the time periods to 0 in rdf. This will tell it to use the time periods in the .log file. (I use the RDF file creater to verify that the RDF file will always be usable for my automatic data collects!) ________________________________________ | ADVANTIS: "A NETWORKING TECHNOLOGY COMPANY" | |TERRY A. STEILEN-8B TSTEILE ON CARVM8 USD6N76N ON IBMMMAIL | |231 N. MARTINGALE RD PHONE:708-240-4774 FAX:3859 | |SCHAUMBURG, IL 60173 | |_______________________________________ ********************************************************************** ═══ 175. Local Monitor ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Trying to monitor a local node. LAN server installed and am logged on to a server. CPUNAME in IBMLAN.INI=BT1GG999. Added this name to the stations to be monitored and started monitoring. Got SPM0296 - no NETBIOS installed. ???? Ron Morrill - Capable Contractor ********************************************************************** >Trying to monitor a local node. LAN server installed and am logged on >to a server. CPUNAME in IBMLAN.INI=BT1GG999. Added this name to >the stations to be monitored and started monitoring. Got SPM0296 - >no NETBIOS installed. ???? >Ron Morrill - Capable Contractor This problem turned out to be LAN Server not being able to successfully start all the LAN Server services. SPM considered LAN Server to be started, used Netbios to communicate, but then failed. ********************************************************************** ═══ 176. SPM0302, 0511, 0513, 0514 on LS 3.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Customer frequently sees above SPM errors, but obviously no harm is done to running graphing/monitoring sessions. Any ideas welcome. Thanks. Martin Murhammer, FSC Austria. 61821731 at VIEVMA murli@vnet.ibm.com ********************************************************************** >Customer frequently sees above SPM errors, but obviously no harm is >done to running graphing/monitoring sessions. >Any ideas welcome. >Thanks. > >Martin Murhammer, FSC Austria. Martin these errors are usually displayed when the DCF(Data Collection Facility) goes down on the machine being monitored for an unexpected reason, e.g. the machine is shutdown. You'll have to be alot more descriptive of the customer scenario for me to be of further help. -Mark Molnar SPM/2 Development ********************************************************************** The machine being monitored is a LAN Server 3.0 advanced, and it is not at all going down when these errors occur since it can well be monitored before, during, and after that. Strange, isn't it? But that's at least what I've been told. Thanks. Martin Murhammer, FSC Austria. 61821731 at VIEVMA murli@vnet.ibm.com ********************************************************************** ═══ 177. (no subject) ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * From: Tony Bosselait - IBM Corporation - TL 262-5409, Client/Server Marketing; Washington, DC Subject: Is there NFS PM/2 & Distributed Feature code available? Sorry to append this note - I wanted to send it to ATIG1 at AUSVM1 but that ID no longer works.... I have an opportunity to break into FANNIE MAE - big Windows/Sun type shop. They have some Lotus notes servers running on OS/2 and have bought off on my recommendation to use LMU2 and SPM2 to manage them. I need to get them code to test in their lab. If they keep it, they'll buy it - they are good that way. Do you know who can help me out? A brand manager? In advance, thanks for your help. Also do we know of any references running Lotus notes with SPM/2? ********************************************************************** ═══ 178. SPM2 and DB2/2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have the same problem described by Tom Talone. It seems that SPM2 doesn't pick up the files activity for the files that DB2/2 reads and writes to. I run a CICS OS/2 (2.0) Application that access DB2/2 tables, specifying the Files Resource in the Monitor SPM definition. The result is that I have in the SPM report the detail of all the SQL files (DAT and INX) used by my application (all these files are related by SPM to CICS OS/2 Application) but all the counters are set to 0. It is impossible that the activity is performed only in the Buffer (previous answer by HOLK ) because I have run a lot of tests (also immediatly after an IPL) and the result is always the same. Is it necessary to do in SPM any particular definition in order to collect these DB2/2 data ? ********************************************************************** ═══ 179. What means paging rate ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The summary Report Heads - Memory description explains paging rate as swapped in/out/loaded pages PER TIME INTERVAL What is TIME INTERVAL ? Is the time interval the WORKING SET PERIOD ? My report says PAGING RATE 13 but what is it? 13 pages/sec or 13 pages/working set period. Hanspeter ********************************************************************** Paging rate is Pages/Second. It's the page-ins + page-outs / collection interval. ********************************************************************** ═══ 180. SPM/2 2.0 With OS/2 2.1 Service pack XR06200 ref OS2DASD.DMD ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello, I installed SPM/2 2.0 on OS/2 2.1 with the service pack installed. The service pack has a later OS2DASD.DMD than the one SPM copies in the C:\OS2 directory. I suspect it is not save to use the SPM version of the OS2DASD.DMD. Am I correct in this assumption and is there a later version with the SPM/2 hooks available? Tom Principaal, software support VM/VSE, OS/2 ********************************************************************** You are right; the OS2DASD.DMD supplied with SPM 2.0 was for OS/2 2.0 only. Don't use it with OS/2 2.0+SPII or OS/2 2.1+. Anyway, you shouldn't have to worry about it, because the SPM install procedure and the SPM CSD install procedure should automatically check if you need a new OS2DASD. ********************************************************************** Hello, Could clarify something for me. If I read this reply correctly, you are saying the SPM2 and SPM2 CSD should detect if you require a new OS2DASD.DMD. My customer tells me that SPM2 has REPLACED the OS2DASD.DMD on an OS/2 2.1 (GA) system. I have told him to copy the originaly OS2DASD.DMD from 2.1 over any older copies. Why would it be replacing the orignial ? Shouldn't the OS2DASD.DMD from 2.1 be more recent than the one from SPM2 and the SPM2 CSD ? Thanks, Rob Allan IBM Canada ********************************************************************** Correction! I have learned that in fact there is a bug in SPM 2.0 and that OS2DASD.DMD is replaced at installation time regardless. The bug is fixed in SPM 2.00.1. So, here's the story. For SPM 2.0: OS/2 2.0 Service Pak 1: Just install SPM. OS/2 2.0 Service Pak 2: After SPM install, copy back the original .DMD OS/2 2.1 After SPM install, copy back the original .DMD For SPM 2.00.1 (CSD version): Just install SPM. ********************************************************************** ═══ 181. OS/2 XR06200 Level ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have installed the OS/2 CSD and then tried to install SPM and the CSD. SPM hangs and the system stops. Does this latest SPM CSD support this level of OS/2 ?? Larry Hillis ********************************************************************** You didn't say if the is the CSD for 2.0 or 2.1. If it is 2.1, then THESEUS2 and the Working Set computations don't work on it. No target when they will. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** Your append DOES sound pretty clear, but just to make sure - REALLY sure - are you saying that SPM/2 does NOT work on OS/2 2.11 or OS/2 2.1 + SP ? If so, (it does NOT work), is it only THESEUS which does not calculate correctly, or does it trap the system as reported in an earlier append. As well you know, this is REAL important for a lot of customers. When customers call into OS/2 support, let's say to get a fix for an OS/2 problem with Lotus Notes, they no longer get just the patch, they are told to apply the CSD. I imagine this is true for all fixes incorporated into the CSD. If they apply it, and SPM traps the system, this is an immediate SEV1. Can you elaborate a little more on status of compatibility with OS/2 2.1 + SP ? Jim Cioffi ********************************************************************** The parts that are broke are: 1) All THESEUS2 functions (it doesn't initialize) 2) Working Set computations for graphing or recording in SPM/2 In both cases, there is a trap in the THESEUS2.SYS device driver which stops the whole system. I understand that this needs to be fixed; however, I have been unable to figure out what Boca changed in the PTDA, so that I can fix my code. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** Out of curiosity having read the 'Theseus2+XR06200 traps out' appends, I just kicked Theseus2 into life. It appears to work fine. My OS/2 is at XR06200 (GA+SP), SPM/2 is WR06075 and Theseus2 came from OS2TOOLS (EXE is 1-12-94 at 3:11pm, SYS is 11-08-93 at 5:17pm. Ian Glass, WSDL, Warwick ********************************************************************** [Reference to append at 15:47 on 94/03/25 by SUMMERS at AUSVM1 ] I'm really confused now. I have the CSD installed, with 1 additional pmgre.dll fix. SPM works fine. It displays working set size. It produces working set reports. I just thought that the computed values were wrong. Even swapmon/2, which uses theseus/2 works fine. What am I doing wrong? Where did I go wrong. Or am I assuming too much? Terry A. Steilen - Advantis ********************************************************************** [Reference to append at 15:47 on 94/03/25 by SUMMERS at AUSVM1 ] Thanks for the status, James. But can you comment on the more recent appends in which people claim to have it working. I have the OS/2 2.11 refresh installed and I get the same traps as you. Is it possible that the Theseus on OS2OOLS is different and works and that it might be a workaround until you get the "official" version fixed? Or are they the same code? Thanks, Jim Cioffi ********************************************************************** The folks that claim that it works should look VERY carefully at the version they are running. It DOES NOT work with the so called "MR1/SP" for 2.1. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** I applied the SP to my system, and Theseus2 works. Maybe it shouldn't but it does!!! It's tracking the working set of ACS3EINI for me even as I type this - 0.047M while my host window is idle, 0.426M (!!!!!) while I'm typing. Following is SYSLEVEL from my system, and DIR of THESEUS2 bits. C:\CMLIB\SYSLEVEL.ACS IBM Communications Manager/2 Version 1.00.1 Component ID 562125400 Current CSD level: WR05999 Prior CSD level: WR00000 C:\IBMCOM\SYSLEVEL.TRP IBM OS/2 LAN Adapter and Protocol Support Version 2.11 Component ID 562125303 Current CSD level: WR06000 Prior CSD level: WR06000 C:\IBMLAN\SYSLEVEL.LSR IBM OS/2 LAN Server/Requester Product Version 3.00 Component ID 562125305 Current CSD level: IP07000 Prior CSD level: IP07000 C:\IBMLAN\SYSLEVEL.REQ IBM OS/2 LAN Requester Version 3.00 Component ID 562125301 Current CSD level: IP07000 Prior CSD level: IP07000 C:\MUGLIB\SYSLEVEL.MUG IBM OS/2 User Profile Management Version 3.00 Component ID 562125302 Current CSD level: WR07000 Prior CSD level: WR07000 C:\MUGLIB\SYSLEVEL.UPE IBM OS/2 User Profile Management - Extended Version 3.00 Component ID 562125306 Current CSD level: IP07000 Prior CSD level: IP07000 IBM;C:\OS2\SYSLEVEL.EPW IBM OS/2 First Failure Support Technology/2 Version 1.20 Component ID 562119400 Current CSD level: WR00275 Prior CSD level: WR00250 C:\OS2\INSTALL\SYSLEVEL.GRE IBM OS/2 32-bit Graphics Engine Version 2.11 Component ID 562107701 Type 0 Current CSD level: XR06200 Prior CSD level: XR06200 C:\OS2\INSTALL\SYSLEVEL.OS2 IBM OS/2 Base Operating System Version 2.11 Component ID 562107701 Type 0 Current CSD level: XR06200 Prior CSD level: XR02010 F:\CORE\COREDATA\SYSLEVEL.COU IBM Network Door/2 Version 1.00 Component ID 560447200 Type ND Current CSD level: IP00000 Prior CSD level: IP00000 F:\IBMCPP\SYSLEVEL.CT3 IBM C/C++ Tools (compiler) Version 2.00 Component ID 562201703 Current CSD level: CT00008 Prior CSD level: CT00002 F:\IBMCPP\SYSLEVEL.CT4 IBM C/C++ Tools (UTILITIES) Version 2.00 Component ID 562201704 Current CSD level: CT00002 Prior CSD level: CT00002 F:\IBMCPP\SYSLEVEL.CT6 IBM C/C++ Tools (CLASS LIBRARIES) Version 2.01 Component ID 562201706 Current CSD level: CT00004 Prior CSD level: CT00002 IBM;F:\SPM2\SYSLEVEL.SPM IBM System Performance Monitor/2 Version 2.00 Component ID 562201000 Current CSD level: WR06075 Prior CSD level: WR06000 F:\TOOLKT20\OS2BIN\SYSLEVEL.TLK IBM OS/2 Programmer's Toolkit Version 2.00 Component ID 562107801 Current CSD level: XR00000 Prior CSD level: XR00000 The volume label in drive F is APPLS■DISK. The Volume Serial Number is 65BC:E014 Directory of F:\spm2 12-01-94 3:11p 281672 0 THESEUS2.EXE 7-01-94 4:24p 378321 0 THESEUS2.HLP 9-08-91 4:02p 288 0 THESEUS2.ICO 7-01-94 10:41a 362504 0 THESEUS2.INF 8-11-93 5:17p 3019 0 THESEUS2.SYS 9 file(s) 1109690 bytes used 61449728 bytes free Ian Glass, WSDL, Warwick ********************************************************************** [Reference to append at 19:41 on 94/03/28 by SUMMERS at AUSVM1 ] Just for your information: I am using Theseus 2.0.1b with 2.1 SP withput any problems. HW: 9577, 16MB, XGA-2, 400MB+180MB SW: OS/2 2.1 US 6.514L, 2.1 SP Theseus 2.0.1b Michael S. Steiger (PS Technical Support Vienna) on 29/03/94 at 11:45:41 ATIBMWML╣IBMMAIL steigerm╣vnet.ibm.com ********************************************************************** I think what James said was that it doesn't work with the manufacturing refresh (OS/2 2.11). As you all have noticed, it seems to work fine with the CSD. And I thought 2.11 was only 2.1 with the CSD applied, silly me. Ron Morrill - Capable Contractor ********************************************************************** Now I'm thouroughly confuzzed. I've applied the CSD to 2.1, which yielded 2.11 with a syslevel of XR06200. I'm running THESEUS2 and SPM's Working Set stuff without any problems. Buttttttt, I've seen other systems where it didn't run right. I've gotta find out what the other systems REALLY were. (THESEUS2 thinks it is "2.0 with ServicePak2". Don't yet know why. Not only am I confused, but so is the program.) I'll be back here when I find out what's going on. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** James, now that you've had some time to get over your cxonfusion, could you let us know if the data that SMP/2 is producing with the CSD installed is even correct. It appears to be working, but is it valid ? Thanks in advance, Terry A. Steilen - Advantis ********************************************************************** I'm running SPM/2 CSD with 2.11. Everything seems to be OK. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 182. Ordering SPM 2.0: Information for IBMers ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: SPM 2.0 is available electronically to IBMers. For Marketing demo or educational purposes it is free. For any other purpose there is a fee. The current published fee is a transfer fee of about $126. This policy is being revised, and the the new cost recovery fee is or will be about $40. The electronic version is sent the next day for orders received before 1pm CST. In the US, you can get an order form for the electron version on OS2PROD. Search for "SPM". Outside the US, use the information appended below. To order a shrink-wrapped version of SPM 2.0, I use PCWARES on VM. Please don't let $40 stand between you and SPM. To me, the only hassle is filling in the accounting information so that the right area gets billed. *************************************** * Revised 08/11/93 *************************************** * IBM I N T E R N A L - E L E C T R O N I C DISTRIBUTION * * AUSTIN - LAN SYSTEMS - ORDER FORM * *************************************** *Q U E S T I O N S -> * *************************************** * If you have any questions about this order information or the order * * process, PLEASE send a note to LSYSINFO at AUSVM1. Please do not * * use the 1-800 number. This number is only for placing the order * * itself. * * *************************************** * * DEMO and EDUCATION USE: * * * We have certain contractual obligations to outside companies when * * using software internally. There are legal implications if the * * "free-of-charge" software is not used for its intended purpose * * (i.e. demo or education). Please refer to the information PACKAGE * * to ensure that you are in compliance with the Demo / Education usage * * requirements. * * * After you determine that you are in compliance......... * * please reference the PACKAGE file for order information on * * Demo and Education use. Simply provide the required information * * (as described in the information PACKAGE file). You do NOT need * * to complete the LANSYS ORDER form for Demo / Education purposes. * * Please do not use the 1-800# to order products for these purposes. * * *************************************** * * P R O D U C T I O N U S E * * *************************************** *O R D E R I N G I N F O R M A T I O N -> * *************************************** * * To order electronic distribution of a LAN Systems product - * * 1. CALL 1-800-IBM-3040- * US/Canada IBM Internals should call (1-800-IBM-3040) and talk * to a telemarketing operator who will ask you basically the SAME * questions that are included in this form. * * ============ OR =============== * * 2. FAX THIS COMPLETED ORDER FORM TO: * * 1-800-826-5399 (US/Canada FAX #) * 1-512-834-8901 (Other/WW FAX #) * * (All WW orders should be sent via FAX) * * ** IMPORTANT ** COMPLETE Payment Information must be provided when * you place your order. The order cannot be accepted * without this information. Please reference the * Payment Information section below for details. * *************************************** * E M P L O Y E E I N F O R M A T I O N -> *************************************** * * NAME: ___________________________________________ * * EXTERNAL TELEPHONE #: ( )____________________ * * IBM EMPLOYEE SERIAL #: _______________ * * NODE: __________ * * ID: __________ * * LOCATION / COUNTRY: __________________________ * * EXTERNAL MAILING ADDRESS: ________________________________ * ________________________________ * ________________________________ * ________________________________ * *************************************** *************************************** * R E Q U I R E D I N F O R M A T I O N -> *************************************** * 1. FOR WHAT PURPOSE WILL YOU BE USING THE CODE (INDICATE WITH X'S): * * _____ PRODUCTION * * _____ APPLICATION DEVELOPMENT * * _____ TESTING / COMPATIBILITY TESTING * * _____ TECHNICAL SUPPORT * * _____ OTHER * * * Please be sure and indicate the purpose for which you are * ordering the code. Production use will be assumed if no choice * is indicated. * * Each IBM employee is subject to the IBM Business Conduct * Guidelines regarding the use of these products. * * Also remember that this code is for your (internal) use ONLY and may * not be sold or given to any customer. * *************************************** *************************************** * P R O D U C T I N F O R M A T I O N -> *************************************** * PRODUCT FORMAT # LICENSES *************************************** * * * ____ LAN Server Version 3.0 (Entry) _____ * ____ 3.5 (default) * ____ 5.25 * COST ->__Prod./Other : $ 193.00 * Total _________ * *************************** * ____ LAN Server Version 3.0 (Advanced) _____ * ____ 3.5 (default) * ____ 5.25 * COST ->__Prod./Other : $ 589.00 * Total _________ *************************** * ____ LAN Server Version 2.0 (Entry) _____ * ____ 3.5 (default) * ____ 5.25 * COST ->__Prod./Other : $ 193.00 * Total _________ *************************** * ____ LAN Server Version 2.0 (Advanced) _____ * ____ 3.5 (default) * ____ 5.25 * COST ->__Production Use: $ 589.00 * Total _________ *************************************** * ____ NTS/2 _____ * ____ 3.5 (default) * COST -> Prod./Other : $ 16.00 ____ 5.25 * Total _________ *************************************** *************************************** * P R O D U C T I N F O R M A T I O N -> *************************************** * PRODUCT FORMAT # LICENSES *************************************** * ____ SPM 2.0 _____ * ____ 3.5 (default) * ____ 5.25 * COST -> Prod./Other : $ 126.00 * * Additional Dist. Features: 20.00 _____ * * **** Reference the SPM2A PACKAGE for information * **** on the SPM 2.0 Distributed Feature. * Total _________ *************************************** * ____ LAN NetView Start 1.1 _____ Continue on the next panel. * ____ 3.5 (default) * ____ 5.25 * * COST -> Prod./Other : $ 147.00 * * Total _________ *************************************** * ____ OS/2 LAN Server for Macintosh _____ * ____ 3.5 (default) * ____ 5.25 * * COST -> Prod./Other : $ 255.00 * * Total _________ *************************************** * ____ Network SignON Coordinator V1.1 _____ * ____ 3.5 (default) * * COST -> Prod./Other : $ 9.45 * * Total _________ *************************************** * *PUBLICATIONS: Reference the original information PACKAGE to determine * the availability of documentation. * *************************************** * P A Y M E N T I N F O R M A T I O N -> *************************************** * You will be billed within IBM for your order. Please contact your * Financial Advisor to obtain the following payment information: * * * Division: xx * * Project/Box: xxxx-xx * * Major: xxx * * Minor: xxxx * * CIB's location: xxx * * * NOTE: This Payment Information must be provided when you place your * order. No code distribution will occur until this information * is complete. * *************************************** *END *************************************** ********************************************************************** ═══ 183. Where can I get the SPMRDFI.EXE ? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * After installation of SPM/2, I miss the SPMRDFI.EXE, which is needed to avoid the PM-Interface. Who can send it to me ? Thank you in advance. Werner Emich ********************************************************************** ═══ 184. Interpreting the Summary Report ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am trying to tune the client side of a client-server application. I do not understand the SPM paging summary reports. Two reports follow. The application is the same in both reports. With 16MEG, the application ran in 96 seconds. With 32MEG, the application ran in 71 seconds. My goal is 20 seconds. I have the following questions: 1. What do the individual headings mean? Paging rate Page-in Page-out Demand loads Discards Faults Fault rate 2. How are the computations made? I am a 370 dinasaur. I think paging rate=fault rate=pageins/time. I think pageouts are unimportant. What is inportant in this report? I cannot make the report compute. 3. When I added the 16 MEG, I expected paging to drop to ZERO. What happened here? Thanks. Bill Vonderhaar (8-261-2443) 16 MEG 01:36 (96 seconds) ------------------------------- Page ------------------------------- Paging Demand Fault Rate Page-In Page-Out Loads Discards Faults Rate ======== ======== ======== ======== 33 455 1220 665 1489 3332 47.51 32 MEG 01:11 (71 seconds) ------------------------------- Page ------------------------------- Paging Demand Fault Rate Page-In Page-Out Loads Discards Faults Rate ======== ======== ======== ======== 33 63 853 420 968 1887 47.20 ********************************************************************** To tune your application to run in 20 seconds, I suggest you use Working Set measurements, not memory measurements. The problem is, the first time you run this application it must load whatever it needs into memory. This could take just as much time whether you have 16 MB or 32 MB. Subsequent executions in a "steady state" environment would test its execution time without the loading variable, in which case the paging could very well drop to zero. To measure the working set of the application or of the whole system with the application, you must have enough physical RAM installed to contain the working set. So 32 MB should work, but 16 MB might not for this measurement. Note that SPM isn't very good at measuring performance from Time t1 to Time t2. However, a short collection interval and a short report summarization interval (even summ interval = collection interval) then you can figure it out. As far as the specifics of the data in you Summary report for the Memory group, I have several comments. 1. Even though the collection time is reported as 96 sec and 71 sec, that's just when the SPM log file was opened and closed. It's not the time period that the data represents. There is a lag from when the log file is opened until the first snapshot is recorded. 2. Paging rate = pages/second for page-in + page-out + demand loads Some page faults are satisfied from memory actually, since OS/2 keeps an "idle list" of quickly dispensable pages. Page outs are important only because it takes system time and and resources to do it. Hope this helps. ********************************************************************** [Reference to append at 00:56 on 94/03/22 by WJVONDE at CHGVMIC1 ] Thank you HOLCK. Some further clarification is needed. Here are my questions: 1. Page-in Is this a physical input I/O from SWAPPER? Page-out Is this a physical output I/O to SWAPPER? Demand loads Is this a physical input from the DLL? In other words, does these three make up all paging physical I/O? 2. Discards Is this a stolen page not requiring swap? Faults Is this any reference to a page not in RAM? Fault rate What is this? In other words, it is true that discards and faults cause no I/O? 3. I did not include the DISK I/O summary. Does the DISK I/O summary include or exclude pagein, pageout, and demand loads? The counts I see appear to be all paging related (ie - the DISK I/O summary includes pagein, etc.) Thanks. Bill Vonderhaar (8-261-2443) ********************************************************************** ═══ 185. SPM/2 CSD on Images ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have to CID install SPM/2 onto a Bank's Server and in the RIPL images. (They do this manually today). I would like to know if I can apply the CSD to my "images" or copy of SPM/2 on my codeserver. This way I will send an install to the Server and to the RIPL clients only once, instead of sending SPM/2 then the CSD to the server; SPM/2 to the RIPL clients then the CSD; Can someone please comment asap? andris ********************************************************************** >I have to CID install SPM/2 onto a Bank's Server and in >the RIPL images. (They do this manually today). I would >like to know if I can apply the CSD to my "images" or copy >of SPM/2 on my codeserver. This way I will send an install >to the Server and to the RIPL clients only once, instead of >sending SPM/2 then the CSD to the server; SPM/2 to the RIPL >clients then the CSD; Can someone please comment asap? This issue is being handled offline, but in general it is a non-standard method of installing. There are several variations possible, such as: -- Copying the CSD install files over the SPM install files -- Copying the CSD installed files over the SPM installed files The only officially supported methods are: -- SPM install followed by SPM CSD install A CSD install is via SERVICE or FSERVICE. -- SPM "manufacturing refresh" (full SPM with CSD fixes included) install Other methods are on a use-at-your-own-risk basis. ********************************************************************** ═══ 186. Accessing "archived" appends from this CFORUM ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The header append for this CFORUM states that the following files are available somewhere: See OS2SPM20 931CFORU on OS2BBS for appends to 92/10/21 (4945 lines) See OS2SPM20 932CFORU on OS2BBS for appends to 93/02/01 (4436 lines) See OS2SPM20 933CFORU on OS2BBS for appends to 93/04/27 (4417 lines) What is "OS2BBS" ? How may I gain access to it - from either IBM Internal or External locations ? Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** The IBM Internal Tools request is: TOOLS SENDTO HONE80 OS2 OS2BBS LIST OS2SPM20 * This should get you a list of files that you can use the GET verb on. Let me know if there are problems; should work OK. Ken Hopkins, TalkLink/OS2BBS Administration ********************************************************************** Thanks ! I was able to retrieve the archived appends using the IBM Internal Tools request that you indicated. Out of curiosity, how do External customers retrieve these archived appends ? Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** ═══ 187. Summary report ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I'd like some help on interpretation of an SPM/2 summary report. I have a report that shows 74.70% total CPU utilization. The report was done at the process level, with unassigned processes reported. The total for all the processes identified is 60.34%. I'm wondering where the rest of the CPU utilization went? I suspect that a device driver is getting a lot of work in this test. Are there alternative collection or reporting options that I can use to get the rest of the information? Also, what causes a program to be designated "unassigned"? Thanks. Dave LeGrys ********************************************************************** 1. I have come to the startling conclusion that the reason thread CPU% doesn't add up to total CPU% is that IT DOESN'T INCLUDE INTERRUPT PROCESSING! So, now we realize that: Total CPU% = (Sum of all threads) + (Interrupt processing) Even this formula is off a bit for busy systems where an SPM snapshot can stretch over several seconds, but that anamoly is explained in an earlier append. Some OS/2 applications such as LAN Server Advanced and LAN Requester run primariy at the interrupt level, so the time spent processing interrupts becomes very important. There are two factors that make it a bit difficult to work with Interrupt processing. a) Interrupt processing is shown on a Summary report to the right of Total CPU, not below it like the threads. b) Interrupt Busy time is presented as a percentage of CPU Busy percentage (a % of a %). 2. "Unassigned" is processes that are not in any SPM-user-defined- application. If a process (.exe) is part of an application, it is reported with that application. If not, it is reported under "unassigned". 3. Device drivers that run at the interrupt level are counted in CPU Interrupt processing. Device drivers that run on the thread of a regular Ring 3 process are attributed to that thread. ********************************************************************** First, thank you very much for your answer, I really appreciated your help. Let me ask a memory question...same "busy driver test". Below is the last portion of a working set report. Please help me understand how to interpret the difference between the working set and accessed memory. My reading of the on-line docs make me wonder why the working set would not be closer to the accessed memory. Can driver storage be counted differently, like we saw for the CPU numbers? I'm sure that this is a misunderstanding on my part of the report and would appreciate your help in understanding what this information means. Thanks. current time --------- Total ---------- ------ ------ ---- Process - hh:mm:ss.tt now ws accessed Free Idle contributed t 15:02:52.25 6.059M 6.199M 12.879M 18.363M 0.063M 9 15:02:53.32 4.676M 6.199M 12.879M 18.363M 0.063M 9 15:02:58.32 5.602M 6.270M 12.949M 18.293M 0.063M 9 1163 samples collected. Total Absolute minimum amount of memory: 7.648M Recommended amount of memory: 9.773M Total amount of accessed memory: 12.949M < End of THESEUS2 (v 2.0.1c) output @ 15:03:00 on 4-20-1994 > Dave LeGrys ********************************************************************** "accessed" is the amount of physical RAM which has been touched sometime during the working set sampling run. The "working set" is amount of physical RAM touched during the "sliding window" (last 'n' samples). James Summers (OS/2 Performance & THESEUS2) ********************************************************************** Thanks for the answer James, as well as your appends on this topic in the THESEUS2 forum...this cleared it up. I'm looking at the WS by Page and WS by Object reports next. Dave LeGrys ********************************************************************** ═══ 188. Theseus and SPM/2 2.0 -- the same thing? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Are these the same thing? Thanks! Lou Rivas 8-391-4771 ********************************************************************** The Theseus memory analysis tool is part of the SPM/2 2.0 product. SPM 2.0 also has a performance monitor with graphs and reports. ********************************************************************** ═══ 189. Where did I put my USER METRIC output (or do I have any ?) ? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * (!!! LONG APPEND !!!!) Using SPM/2 Version 2.00 CSD Level WR06075 OS/2 Version 2.10 CSD Level XR02010 > > Added portions of output from SPMSNAPP to the end of this append > I am attempting to collect results from some user metrics which measure code path lengths and cannot get any results to appear in my SPM/2 dump reports. I assume that, unless I write a separate application to retrieve the results, the dump report is the proper way to view the results. Since I have not been able to retrieve user metric data results on any levels of SPM/2 and OS/2, I do not suspect that my problem resides with them. I am sure there is a basic problem with my approach, but I cannot determine where to go from here ! My metric definition file is as follows: *************************************** sftwnme=DLPSPM/2 prgmnme=SPM_DLP groupnme="DLP_Timer" groupord=512 grouptag=DLPCPU fieldnme="DLP Counter Ticks" fieldtag=counter fieldtyp=counter fielddsc="DLP Counter Ticks" fieldhlp= fieldnme="DLP Timer" fieldtag=timer fieldtyp=timer fielddsc="DLP Timer" fieldhlp= fieldnme="DLP Queue" fieldtag=queue fieldtyp=queue fielddsc="DLP Queue" fieldhlp= fieldnme="DLP Bulk Counter" fieldtag=bulkCounter fieldtyp=bulkCounter fielddsc="DLP Bulk Counter" fieldhlp= fieldnme="DLP State Counter" fieldtag=stateCounter fieldtyp=stateCounter fielddsc="DLP State Counter" fieldhlp= fieldnme="DLP State Timer" fieldtag=stateTimer fieldtyp=stateTimer fielddsc="DLP State Timer" fieldhlp= fieldnme="DLP Bulk State Counter" fieldtag=bulkstateCounter fieldtyp=bulkstateCounter fielddsc="DLP Bulk State Counter" fieldhlp= *************************************** When I use SPMAPPIN to include the above in SPMAPP.INI, the following is displayed: *************************************** Data from text file 'C:\SPM2V2\api\spm_dlp.SPM' will be stored in 'C:\SPM2V2\SPMAPP.INI' Phase 01 started Phase 02 started Phase 02 input parse Phase 02 table build SPM3020: Ordinal 512 replaced Phase 02 file update Phase 02 finished Phase 01 finished Counter group mapping file 'C:\SPM2V2\api\SPMCTRGP.H' will be generated from 'C:\SPM2V2\SPMAPP.INI' Phase 03 started Phase 03 finished SPMAPPIN ended *************************************** I start a graphing/recording monitoring session which includes the "groupnme" of DLP_Timer from my metric definition file as one of the resources to be collected. My sampling interval is 10 seconds. The source of my application which registers and updates my metric data values follows: *************************************** #define INCL_DOSPROCESS #define INCL_32 #include #include #include #include "SPMUH.h" struct _test_db { DBH dbh; CTR counter; TMR timer; QL queueLength; BULKCTR bulkCounter; CTR_STATE counter_state; TMR_STATE timer_state; BULKCTR_STATE bulkcnt_state; } test_db; NAMES names{}={ SPMDataItem( test_db.counter, SPM_T_CTR, "DLP Counter Ticks"), SPMDataItem( test_db.timer, SPM_T_TMR, "DLP Timer"), SPMDataItem( test_db.queueLength, SPM_T_QL, "DLP Queue"), SPMDataItem( test_db.bulkCounter, SPM_T_BULKCTR, "DLP Bulk Counter"), SPMDataItem( test_db.counter_state, SPM_T_CTR_STATE, "DLP State Counter"), SPMDataItem( test_db.timer_state, SPM_T_TMR_STATE, "DLP State Timer"), SPMDataItem( test_db.bulkcnt_state, SPM_T_BULKCTR_STATE, "DLP Bulk State Counter"), SPMEnd }; VOID main( int argc, char *argv{}) { ULONG rc; ULONG SPMhandle; QWORD qwCurrentTime; int i,j,k; char c; printf("The SPM_DLP program has started ! .........\n"); rc = SPMRegister( "DLPSPM/2", "SPM_DLP", "WindowCompat", "commEnt, and more, and still more", "DLPCPU", (PDBH) &test_db, (PNAMES) &names, &SPMhandle); printf("SPMRegister RC = %9ld\n", rc); if (rc == 0) { /* Let's update them. */ for (k=1; k < 3; k++ ) { /* The timer is started and then stopped. */ printf("Start timing .........(iteration #) %d \n", k); /* The semaphore must be set before we start updating the values. */ SPMSetSem( SPMhandle); SPMStartTimer( &test_db.timer); for (j=0; j < 500; j++) { for (i=0; i<10001; i++) ; /* count to 10,000 */ } /* endfor "j" loop */ SPMStopTimer( &test_db.timer); printf("Stop timing .........\n"); /* Query current time value. *. */ SPMQueryTime( &qwCurrentTime); /* Now that we are done, clear the semaphore. */ SPMClearSem( SPMhandle); Continue on the next panel. /* Generate a pause so that SPM/2 has a chance to grab the semaphore */ printf("Enter any character to end pause in program ...\n"); c = getchar(); } /* endfor "k" loop */ /* Can now deregister the group. */ rc = SPMDeregister( SPMhandle); printf("SPMDeregister RC = %9ld\n", rc); } } *************************************** The output for this application (which is running while SPM/2 is graphing and recording data) follows. Note that I introduced a pause in my program (by requesting user input after a timing interval) to ensure that my application was still active in memory while SPM/2 was attempting to access my application's memory to retrieve my user metric data values): *************************************** The SPM_DLP program has started ! ......... SPMRegister RC = 0 Start timing .........(iteration #) 1 Stop timing ......... Enter any character to end pause in program ... Start timing .........(iteration #) 2 Stop timing ......... Enter any character to end pause in program ... SPMDeregister RC = 0 *************************************** When I run an SPM/2 dump report which includes DLP_Timer, I see nothing related to any of my user metrics. Am I following the correct procedure to use the SPM/2 user metric functions to measure code path lengths ? I attempted to "time" the running of the program to ensure that SPM/2 had plenty of time to gather my user metric data values before my application issued the set and clear semaphore functions. Any suggestions ? > > Portions of output from SPMSNAPP > SPMSNAPP: Print Utility Version 2.0.0 SPM2540: SPMSNAPP: The log is currently active SPMSNAPP: Dump of active area: ->a20f:0000: 01 00 fc 00 04 00 00 00 00 00 00 00 27 00 03 00 ............'... 00016(0010): 00 00 43 3a 5c 53 50 4d 32 56 32 5c 53 50 4d 53 ..C:\SPM2V2\SPMS 00032(0020): 4e 41 50 2e 45 52 52 00 53 45 54 49 4e 46 4f 00 NAP.ERR.SETINFO. 00048(0030): 53 5f 41 4c 49 41 53 45 4e 55 4d 00 53 5f 55 53 S_ALIASENUM.S_US 00064(0040): 45 52 47 45 54 4c 4f 47 4f 4e 41 53 4e 00 53 5f ERGETLOGONASN.S_ 00080(0050): 55 53 45 52 53 45 54 4c 4f 47 4f 4e 41 53 4e 00 USERSETLOGONASN. 00096(0060): 53 5f 55 53 45 52 47 45 54 41 50 50 53 45 4c 00 S_USERGETAPPSEL. 00112(0070): 53 5f 55 53 45 52 53 45 54 41 50 50 53 45 4c 00 S_USERSETAPPSEL. 00128(0080): 53 5f 41 50 50 41 44 44 00 53 5f 41 50 50 44 45 S_APPADD.S_APPDE 00144(0090): 4c 00 53 5f 41 50 50 47 45 54 49 4e 46 4f 00 53 L.S_APPGETINFO.S 00160(00a0): 5f 41 50 50 53 45 54 49 4e 46 4f 00 53 5f 41 50 _APPSETINFO.S_AP 00176(00b0): 50 45 4e 55 4d 00 53 5f 55 53 45 52 44 43 44 42 PENUM.S_USERDCDB 00192(00c0): 49 4e 49 54 00 00 00 00 00 00 00 00 00 00 00 00 INIT............ 00208(00d0): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00224(00e0): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00240(00f0): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00256(0100): 00 00 00 00 37 a7 ca 00 37 a7 0b 00 37 a7 ce 00 ....7...7...7... 00272(0110): 37 a7 ff 01 28 00 00 00 2b 00 02 00 63 00 5f 00 7...(...+...c._. 00288(0120): 40 20 31 39 39 34 2d 30 34 2d 30 34 2d 31 36 2e @ 1994-04-04-16. 00304(0130): 33 36 2e 30 32 2e 33 37 20 50 3a 31 38 34 20 54 36.02.37 P:184 T 00320(0140): 3a 31 20 4c 3a 34 20 52 65 70 6f 72 74 20 73 74 :1 L:4 Report st 00336(0150): 61 72 74 65 64 20 66 6f 72 20 43 3a 5c 53 50 4d arted for C:\SPM 00352(0160): 32 56 32 5c 53 50 4d 5f 44 4c 50 2e 44 4d 50 0d 2V2\SPM_DLP.DMP. 00368(0170): 0a 32 5c 53 50 4d 5f 44 4c 50 2e 4c 4f 47 00 00 .2\SPM_DLP.LOG.. 00384(0180): 00 6e 64 20 62 72 6f 6b 65 6e 20 74 68 65 20 70 .nd broken the p 00400(0190): 69 70 65 20 74 6f 20 69 74 21 00 00 00 00 00 00 ipe to it!...... 00416(01a0): 43 3a 5c 53 50 4d 32 56 32 5c 53 50 4d 44 43 46 C:\SPM2V2\SPMDCF 00432(01b0): 2e 45 58 45 00 00 00 00 00 00 00 00 00 4e 45 54 .EXE.........NET 00448(01c0): 42 49 4f 53 47 45 54 49 4e 46 4f 00 4e 45 54 53 BIOSGETINFO.NETS 00464(01d0): 55 42 00 4e 45 54 42 49 4f 53 4f 50 45 4e 00 00 UB.NETBIOSOPEN.. 00480(01e0): 4c 41 4e 00 00 00 00 00 e8 01 37 a7 4e 45 54 53 LAN.......7.NETS 00496(01f0): 55 42 00 4e 45 54 42 49 4f 53 43 4c 4f 53 45 00 UB.NETBIOSCLOSE. 00512(0200): 4c 41 4e 00 1e 3c 01 00 00 00 00 00 00 00 00 00 LAN..<.......... 00528(0210): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00544(0220): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00560(0230): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00576(0240): 00 00 00 00 01 00 00 00 00 00 4a 02 37 a7 5c 4d ..........J.7.\M 00592(0250): 41 49 4c 53 4c 4f 54 5c 4c 41 4e 4d 41 4e 00 5c AILSLOT\LANMAN.\ 00608(0260): 5c 25 46 73 25 46 73 00 68 6c 65 6e 75 6d 2e 63 \%Fs%Fs.hlenum.c 00624(0270): 00 41 73 73 65 72 74 69 6f 6e 20 66 61 69 6c 65 .Assertion faile 00640(0280): 64 3a 20 66 69 6c 65 20 25 46 73 2c 20 6c 69 6e d: file %Fs, lin 00656(0290): 65 20 25 64 0a 00 5c 53 48 41 52 45 4d 45 4d 5c e %d..\SHAREMEM\ 00672(02a0): 4e 45 54 5c 48 4f 53 54 4c 49 53 54 2e 53 45 47 NET\HOSTLIST.SEG 00688(02b0): 00 68 6f 73 74 6c 69 73 74 2e 63 00 41 73 73 65 .hostlist.c.Asse 00704(02c0): 72 74 69 6f 6e 20 66 61 69 6c 65 64 3a 20 66 69 rtion failed: fi 00720(02d0): 6c 65 20 25 46 73 2c 20 6c 69 6e 65 20 25 64 0a le %Fs, line %d. 00736(02e0): 00 68 6f 73 74 6c 69 73 74 2e 63 00 41 73 73 65 .hostlist.c.Asse 00752(02f0): 72 74 69 6f 6e 20 66 61 69 6c 65 64 3a 20 66 69 rtion failed: fi 00768(0300): 6c 65 20 25 46 73 2c 20 6c 69 6e 65 20 25 64 0a le %Fs, line %d. 00784(0310): 00 00 5c 53 48 41 52 45 4d 45 4d 5c 4e 45 54 5c ..\SHAREMEM\NET\ SPMSNAPP: The file to be formatted is C:\SPM2V2\SPMSNAP.ERR @ 1994-04-04-16.28.57.31 P:176 T:1 L:4 SPMISTRT: File C:\SPM2V2\SPM_DLP.LOG verified and collection initiated @ 1994-04-04-16.28.57.50 P:176 T:1 L:4 SPMISTRT: SPMISTRT Ended return code 0 ->0037:298c: 00 00 ................ @ 1994-04-04-16.29.05.22 P:177 T:4 L:4 SPM1625: NBQIQMgrApdg - Poll response received but poll list inactive. ->8bff:0050: 26 00 02 00 42 49 45 47 41 30 00 41 23 00 00 00 &...BIEGA0.A#... 00016(0010): 00 00 00 00 00 4d 38 37 34 37 32 39 41 23 00 00 .....M874729A#.. 00032(0020): 00 00 00 00 00 00 ................ @ 1994-04-04-16.29.07.50 P:177 T:4 L:4 SPM1625: NBQIQMgrApdg - Poll response received but poll list inactive. ->8bff:0050: 26 00 02 00 50 43 43 50 53 32 30 32 30 00 00 00 &...PCCPS2020... 00016(0010): 00 00 00 00 00 4d 38 37 34 37 32 39 41 23 00 00 .....M874729A#.. 00032(0020): 00 00 00 00 00 00 ................ @ 1994-04-04-16.29.08.03 P:21 T:1 L:128 SPMNBL: DCF M874729A1 (PID 179) started on behalf of M874729AA @ 1994-04-04-16.29.16.12 P:21 T:1 L:128 SPMNBL: DCF started. @ 1994-04-04-16.29.16.72 P:178 T:10 L:4 SPMILOG: Logon in progress for node M874729A @ 1994-04-04-16.29.19.06 P:179 T:4 L:8 SPM0374: SPMDCF: Group name DLP_Timer not found @ 1994-04-04-16.29.19.09 P:179 T:4 L:8 SPM0066: SPMDCF: Unable to resolve OS/2 kernel symbolics - return code 8 @ 1994-04-04-16.29.19.12 P:179 T:4 L:4 SPMDCF: OS/2 kernel message queue processing type enabled by ordinal 4 @ 1994-04-04-16.29.19.12 P:179 T:4 L:8 SPM0375: SPMDCF: Text Block name tmReady in group THD not found @ 1994-04-04-16.29.30.68 P:179 T:4 L:4 SPMDCF: !!Xchg c-time ->02d7:ffe8: d1 9c 19 85 04 00 00 00 ................ @ 1994-04-04-16.29.30.68 P:179 T:4 L:4 SPMDCF: !!Xchg e-time Continue on the next panel. ->02d7:ffa8: fb 8d 19 85 04 00 00 00 ................ @ 1994-04-04-16.29.30.72 P:179 T:4 L:4 SPMDCF: !!DcfPrevRawData ->030f:55d4: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00016(0010): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00032(0020): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00048(0030): 00 00 00 00 00 00 00 00 00 00 00 00 ................ @ 1994-04-04-16.29.30.72 P:179 T:4 L:4 SPMDCF: !!DcfCurrRawData ->030f:5584: 3c 00 00 00 04 00 22 00 0d 00 00 00 96 00 00 00 <....."......... 00016(0010): 5c a0 de 7b 00 00 00 00 11 00 00 00 40 89 f7 7a \..{........@..z 00032(0020): fb ff ff ff 00 01 00 00 2b 00 00 00 e3 9c 05 00 ........+....... 00048(0030): 00 00 00 00 6a 0d 04 00 00 00 00 00 ....j........... @ 1994-04-04-16.29.30.72 P:179 T:4 L:4 SPMDCF: !!DcfCurrOutData ->030f:5624: 50 00 04 00 b3 00 00 00 04 00 00 00 1e 00 00 00 P............... 00016(0010): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00032(0020): 00 00 00 00 00 00 00 00 00 00 26 a8 ba f7 0d 1c ..........&..... 00048(0030): d0 41 00 00 00 00 00 00 00 00 43 3a 5c 53 50 4d .A........C:\SPM 00064(0040): 32 56 32 5c 53 50 4d 44 43 46 2e 45 58 45 00 00 2V2\SPMDCF.EXE.. @ 1994-04-04-16.29.30.75 P:179 T:4 L:4 SPMDCF: !!DcfCalcOutData ->030f:5624: 50 00 04 00 b3 00 00 00 04 00 00 00 1e 00 11 00 P............... 00016(0010): 00 00 12 21 71 3f 00 00 00 00 00 00 00 00 2b 00 ...!q?........+. 00032(0020): 00 00 5b d8 9d 3e a5 ec 63 3e 26 a8 ba f7 0d 1c ..{..>..c>&..... 00048(0030): d0 41 00 00 00 00 00 00 00 00 43 3a 5c 53 50 4d .A........C:\SPM 00064(0040): 32 56 32 5c 53 50 4d 44 43 46 2e 45 58 45 00 00 2V2\SPMDCF.EXE.. @ 1994-04-04-16.29.30.81 P:179 T:4 L:4 SPMDCF: !!Xchg c-time ->02d7:ffe8: df 3a c3 85 04 00 00 00 .:.............. @ 1994-04-04-16.29.30.81 P:179 T:4 L:4 SPMDCF: !!Xchg e-time ->02d7:ffa8: ac 2b c3 85 04 00 00 00 .+.............. ***** ***** ***** Deleted about 550 lines of "repetitive" data ! ***** ***** ***** 00064(0040): 32 56 32 5c 53 50 4d 44 43 46 2e 45 58 45 00 00 2V2\SPMDCF.EXE.. @ 1994-04-04-16.33.03.47 P:179 T:4 L:4 SPMDCF: !!Xchg c-time ->02d7:ffe8: 49 65 af 94 04 00 00 00 Ie.............. @ 1994-04-04-16.33.03.50 P:179 T:4 L:4 SPMDCF: !!Xchg e-time ->02d7:ffa8: 2a 4f af 94 04 00 00 00 *O.............. @ 1994-04-04-16.33.03.51 P:179 T:4 L:4 SPMDCF: !!DcfPrevRawData ->030f:55d4: 3c 00 00 00 04 00 22 00 0d 00 00 00 a2 05 00 00 <....."......... 00016(0010): 5c a0 de 7b 00 00 00 00 b9 01 00 00 35 9d 32 6c \..{........5.2l 00032(0020): fb ff ff ff 00 01 00 00 2f 00 00 00 e5 5f 06 00 ......../...._.. 00048(0030): 00 00 00 00 4b d5 04 00 00 00 00 00 ....K........... @ 1994-04-04-16.33.03.53 P:179 T:4 L:4 SPMDCF: !!DcfCurrRawData ->030f:5584: 3c 00 00 00 04 00 22 00 0d 00 00 00 d2 05 00 00 <....."......... 00016(0010): 5c a0 de 7b 00 00 00 00 c9 01 00 00 d7 c0 7d 6b \..{..........}k 00032(0020): fb ff ff ff 00 01 00 00 2f 00 00 00 e5 5f 06 00 ......../...._.. 00048(0030): 00 00 00 00 4b d5 04 00 00 00 00 00 ....K........... @ 1994-04-04-16.33.03.54 P:179 T:4 L:4 SPMDCF: !!DcfCurrOutData ->030f:5624: 50 00 04 00 b3 00 00 00 04 00 00 00 1e 00 11 00 P............... 00016(0010): 00 00 a1 00 9a 3d 00 00 00 00 00 00 00 00 00 00 .....=.......... 00032(0020): 00 00 00 00 00 00 00 00 00 00 26 a8 ba f7 0d 1c ..........&..... 00048(0030): d0 41 00 00 00 00 00 00 00 00 43 3a 5c 53 50 4d .A........C:\SPM 00064(0040): 32 56 32 5c 53 50 4d 44 43 46 2e 45 58 45 00 00 2V2\SPMDCF.EXE.. @ 1994-04-04-16.33.03.54 P:179 T:4 L:4 SPMDCF: !!DcfCalcOutData ->030f:5624: 50 00 04 00 b3 00 00 00 04 00 00 00 1e 00 10 00 P............... 00016(0010): 00 00 1d e5 ae 3d 00 00 00 00 00 00 00 00 00 00 .....=.......... 00032(0020): 00 00 00 00 00 00 00 00 00 00 26 a8 ba f7 0d 1c ..........&..... 00048(0030): d0 41 00 00 00 00 00 00 00 00 43 3a 5c 53 50 4d .A........C:\SPM 00064(0040): 32 56 32 5c 53 50 4d 44 43 46 2e 45 58 45 00 00 2V2\SPMDCF.EXE.. @ 1994-04-04-16.33.05.43 P:181 T:1 L:4 SPMISTOP: DRAIN of C:\SPM2V2\SPM_DLP.LOG initiated ->0037:2a30: 00 00 ................ @ 1994-04-04-16.33.05.47 P:181 T:1 L:4 SPMISTOP: SPMISTOP Ended return code 0 ->0037:2a26: 00 00 ................ @ 1994-04-04-16.33.06.15 P:178 T:1 L:4 SPMILOG: SPMILOG Stop requested @ 1994-04-04-16.33.06.18 P:178 T:1 L:4 SPMILOG: DRAIN command acknowledged @ 1994-04-04-16.33.07.12 P:179 T:4 L:4 SPMDCF: Sample point pool new maximum encountered: 181740 bytes @ 1994-04-04-16.33.07.81 P:179 T:1 L:4 SPMDCF: DCF stop requested @ 1994-04-04-16.33.07.84 P:179 T:2 L:4 SPMDCF: Shutdown command received via pipe @ 1994-04-04-16.33.07.87 P:178 T:10 L:4 SPM0450: SPMILOG: Have initiated stop of DCF M874729A1, and broken the pipe to it! @ 1994-04-04-16.33.08.22 P:178 T:7 L:4 SPMILOG: .LOG shut down as requested @ 1994-04-04-16.33.08.43 P:178 T:1 L:4 SPMILOG: SPMILOG Ended return code 0 ->003f:ac4c: 00 00 ................ @ 1994-04-04-16.33.08.56 P:178 T:1 L:4 SPMILOG: {ExitList Msg} NORMAL TERMINATION ->003f:acec: 00 00 00 00 ................ @ 1994-04-04-16.36.00.09 P:184 T:1 L:4 Using C:\SPM2V2\SPM_DLP.RDF @ 1994-04-04-16.36.00.65 P:184 T:1 L:4 Data collection started for C:\SPM2V2\SPM_DLP.LOG @ 1994-04-04-16.36.02.37 P:184 T:1 L:4 Report started for C:\SPM2V2\SPM_DLP.DMP 2\SPM_DLP.LOG Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** The problem is in your .SPM file. The .SPM file must match the SPMRegister call in the application. The .SPM file tells SPM about the user metrics; The application's SPMRegister call tells OS/2 about the user metrics. In this case, to match the SPMRegister call's group name DLPCPU, change the .SPM line to: groupnme="DLPCPU" Another parameter in the .SPM file, the grouptag, is commented in the example documentation with the comment "name of group as registered with SPM", but I think in fact it is the groupnme parameter that matters. To be sure, make them both the same: groupnme="DLPCPU" ... grouptag=DLPCPU ********************************************************************** ═══ 190. SPM0302 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Customer has same problem (error message SPM0302) as Martin Murhammer, The machine being monitored is not down, either. The monitoring machine is not on the same ring as the machine monitored (OS LAN Server 3.0), the 2 rings are connected through a bridge. What would you recommend ? Thanks. C. de la CHANONIE, France ********************************************************************** ═══ 191. Support for OS/2 2.11 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi .... does SPM/2 2.0 (with its CSD) support OS/2 2.11? Are there any SPM/2 fixes required to support OS/2 2.11? Thanks.... Hugo Garcia. Personal Systems Solutions. IBM Canada Ltd. ********************************************************************** Yes, SPM 2.0 supports OS/2 2.11. There is still some testing underway to determine if any fixes will be required. ********************************************************************** ═══ 192. LAN Server for Workgroups ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: I am pleased to announce the availability of SPM 2.0 in a new IBM product offering. "LAN Server for Workgroups" is a new IBM product packaging that includes: OS/2 2.11 LAN Server 3.0 Lotus Notes SPM 2.0 The product is available on compact disk. Please see the official announcement for more details. ********************************************************************** Is Lotus Notes certified for use with OS/2 2.11? Dave ********************************************************************** >Is Lotus Notes certified for use with OS/2 2.11? I can't really speak for Lotus Notes on this forum, but I would certainly hope it is certified for use with OS/2 2.11 since both are part of the LAN Server For Workgroups product. ********************************************************************** in NOTES FORUM That is why I asked the question. See second referenced append for an indication that the packaging of OS/2 2.11 with LAN Server for Workgroups and the current level of Notes may be a little premature. Dave ********************************************************************** ═══ 193. SPM/2 APIs & NetBIOS ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I'm using the SPM/2 APIs in a program which uses NetBIOS for its own communication needs. The program has a thread which calls SPMAPIInit() followed by SPMAPIOpen() and then repeatedly calls SPMAPIGetData(). At the time SPMAPIInit() returns with a good return code, it is apparently not completely done initializing; if the call to SPMAPIOpen() is done too soon, it fails; but more importantly, it also seems to deallocate all of my program's NetBIOS resources. So, my question is: is there any way that my program can determine when SPMAPIInit() is actually done initializing and thus know that it is safe to go ahead with the SPMAPIOpen()? This problem should be solvable by making the thread doing the SPM calls into a separate program, but I'm hoping to avoid that. Pat Nickel (pjnickel@rchland) department 480 IBM Rochester, Minnesota (507)253-3420 t/l 553-3420 ********************************************************************** (This was handled offline, but here's the answer.) A good return code from SPMAPIInit doesn't mean that everything ything is ready to go, it just means that the your request was received and initialization is taking place. What you have to do is a while loop that does an SPMAPIQuery of the SPM_QU_ILOG_NAME until you get a good return code. his, The answer isn't pretty, but our SPM/2 graph has to use this technique if it makes you feel any better: While (returncode == error) { returncode = SPMAPIQuery(SPM_QU_ILOG_NAME, szlogfile, blah, blah) if (returncode == error) DosSleep() } ********************************************************************** ═══ 194. Problems with Lotus Notes Servers ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am helping a customer who has the following problem: Lotus Notes Servers give the message "Netbios Names are in Conflict" then the Server and requestor lockup. They have isolated the problem: any Lotus Notes Server that has the distributed SPM/2 code installed and running has this problem. Without SPM/2 distributed feature there is no problem. Any ideas? Thanks, Lee ********************************************************************** ═══ 195. Questions concerning User Metric Timers ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 1) What level of granularity is possible using the SPMStartTimer and SPMStopTimer functions ? The .INF file speaks about the use of the 838 nanosecond timer ticks of the PS/2. Does this mean that timings can be made to the microsecond ? The SPM Dump report seems to only report timings to the millisecond level (e.g., 4.295 seconds). If greater granularity exists, how can it be obtained ? 2) Can a timer be started (SPMStartTimer) in one process (.EXE) and stopped (SPMStopTimer) in a separate process (.EXE) ? 3) What is the proper way to display a timing from the SPM timers in a user application - given that the timings are kept in 8-byte qwords - rather than looking at them in an SPM Dump report ? 4) If you display the contents of the address holding the timer prior to issuing the SPMStopTimer, what do the values in the .ulLo and .ulHi fields represent ? How does this compare to the values in those fields when the SPMQueryTime function is used instead ? Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** 1) The granularity/frequency of the timer is 1193167 Mhz (838 ns). There is overhead in acquiring the time, but microsecond precision is possible with proper determination of the overhead amounts, etc. Tim Holck will have to answer the report precision question. 2) It doesn't matter what processes the timers are manipulated in, as long as they both have addressability to the timer. 3) In my test program, I simply print out the hex values for the Hi & Lo ulongs. Don't know of any way of converting them to decimal with C. However, if the timer has been started and then stopped, so that a delta is in it that is less than 4 gigaticks, then you could simply format the lower ulong and ignore the high one. 4) SPMStartTimer subtracts current time (in ticks) from the value in the timer. SPMStopTimer adds current time to the value, to generate the delta. So, after a start and before a stop, assuming the value was zero before the start, what you see is the negative of the start time. SPMQueryTime simply returns the current time (in ticks). SPMStartTimer and SPMStopTimer calles SPMQueryTime and then does the math. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 05:42 on 94/04/13 by DENNIS at RALVM0 ] The resolution of the SPM timers is very good. The SPM Dump Report rounds off to milliseconds. To see the timers at their best resolution you will have to use the SPM API (not just the User Metric API) to collect the performance data and format the timers as you wish. You can use the SPM API to get realtime data or get the data from an SPM log file; either way you should be able to get the timers at their best resolution. Converting quadword (8 byte) timers to a decimal number can be done. SPM does it, and I have the source code that does it. Send me a note if you want me to dig it up. ********************************************************************** [Reference to append at 14:30 on 94/04/13 by SUMMERS at AUSVM1 ] In my initial append, I asked the following: 2) Can a timer be started (SPMStartTimer) in one process (.EXE) and stopped (SPMStopTimer) in a separate process (.EXE) ? The response was: 2) It doesn't matter what processes the timers are manipulated in, as long as they both have addressability to the timer. Does anyone have an example of starting and stopping timers in separate programs ? I have not been successful as yet. My SPMRegister gives a return code of 0 in both programs; however, when the SPMStopTimer executes in the 2nd program, it appears as though the SPMStartTimer has not yet executed in the first program even though I'm sure it has. In addition, I try to update the same counter in both programs, but the results seem to indicate that neither program is aware of the other program. I guess the real question is how do I ensure both programs have "addressability" to the same timers and counters ? Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** 1) The SPMRegister should only be done in 1 process. (It must be deregistered in the same process that registered it.) 2) The DataBlock should be in shared memory, with both processes having access to it. Some of our LAN Server code does this and this is the way they do it. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 05:42 on 94/04/13 by DENNIS at RALVM0 ] > There is a problem of getting my user metrics to show up in an > SPM Report Dump file unless I ensure that the user application which > SPMRegisters my user metrics starts and remains in memory BEFORE I > start DCF via an SPM monitoring session. This restriction may not > allow us to use SPM in our product since there is no guarantee that > all the processes which may use the SPM hooks will all be already > loaded (and stay loaded) before the SPM monitoring session starts. > > I seem to remember the words "persistent data" being used to explain > why an application - which is not in memory when SPM starts - may or > may not be seen by SPM. There is a deficiency in how SPM 2.0 collects User Metrics. SPM will not collect a user metric resource group unless the user application is already loaded and has already registered its user metrics. The problem is that user metrics may be registered and deregistered at any time and are therefore "transient", as opposed to most OS/2 resources which are registered when OS/2 boots up and are "persistent". The workarounds are: 1) Always start the user application before SPM data collection (i.e. always start SPM data collection after the application has started), OR 2) Collect file or thread data along with the user metrics. Files and threads are "transient" resources, so SPM will be be prompted to collect all transient resources including user metrics. ********************************************************************** ═══ 196. SPM0302 error message ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Has anyone read my append on SPM0302 and could help ? Thanks ? C. de la CHANONIE ********************************************************************** ═══ 197. LAN Server app data still available ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * With the new "talklink" a WATCHIT demo is still available. Type os2info and find 'WATCHIT' for a demo of a: LAN SERVER application oriented data collection level program to determine capacity and performance. Send comments to dev2401 for a personalized answer. Labenski CSN Inc Home of WATCHIT and CONDUIT (203) 233-2951 ********************************************************************** From: Vince Coen To: Dev2401 At Hone83ж Hi, -=> Quoting Dev2401 At Hone83 to All <=- DAH> With the new "talklink" a WATCHIT demo is still available. DAH> Type os2info and find 'WATCHIT' for a demo of a: DAH> LAN SERVER application oriented data collection level DAH> program to determine capacity and performance. DAH> Send comments to dev2401 for a personalized answer. Is this product (demo and full) available in the UK? Regards, Vince. ___ Blue Wave/QWK v2.20f OS2 Beta * EMEA OS/2 Developer Assistance Program BBS * gHostNet(0057002C01000000 2daf4f4c) reply to (0008037800580000 B4CB7F05) ********************************************************************** The demo is a an introduction. It will give you a chance to look at some one elses data. We have no European distributer. You may buy direct. If you use a credit card $$$ conversion is easiest. Indelible Blue has the product if you have an account with them. Labenski CSN Inc Home of WATCHIT and CONDUIT (203) 233-2951 ********************************************************************** [Reference to append at 21:05 on 94/04/14 by DEV2401 at HONE83 ] From: Vince Coen To: Dev2401 At Hone83ж Hi, -=> Quoting Dev2401 At Hone83 to All <=- DAH> With the new "talklink" a WATCHIT demo is still available. DAH> Type os2info and find 'WATCHIT' for a demo of a: DAH> LAN SERVER application oriented data collection level DAH> program to determine capacity and performance. DAH> Send comments to dev2401 for a personalized answer. Is this product (demo and full) available in the UK? Regards, Vince. ___ Blue Wave/QWK v2.20f OS2 Beta * EMEA OS/2 Developer Assistance Program BBS * gHostNet(0057002C01000000 2daf4f4c) reply to (0008037800580000 B4CB7F05) ********************************************************************** ═══ 198. How to get started for local workstation ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * | User error Ed ********************************************************************** ═══ 199. Monitor a server ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None From: Andre Faure To: all How does one install SPM 2.x to monitor a Lan Server 3.0 from a workstation: a) from a workstation? b) from the console? Must the workstation have SPM already installed? I'm not too keen on experimenting with a software which performs changes to the Config.Sys of a server. I've upgraded from SPM 1.0 and the sole documentation which came with the software consisted in a three page flyer. SPM 1.0 was removed from the server when upgrading to OS/2 2.1. Regards, Andre Valmet Paper Machinery Inc. Jyvaskyla, Finland ___ KWQ/2 1.2d "640K ought to be enough for anybody." - Bill Gates, 1981 * EMEA OS/2 Developer Assistance Program BBS * gHostNet(0057002C01000000 2db205c0) ********************************************************************** Yor connemt or "640k should be enougth for anyone caught me eye!! My 1st machine was a 4k 360 in 1966. Haven't we grown? OR ........... LOST control ?? Labenski CSN Inc Home of WATCHIT and CONDUIT (203) 233-2951 ********************************************************************** [Reference to append at 09:28 on 94/04/18 by EMEADAP at WINVMB ] > How does one install SPM 2.x to monitor a Lan Server 3.0 from a > workstation: a) from a workstation? b) from the console? Must the > workstation have SPM already installed? At the LAN Server, install the Distributed Feature from diskettes: SPMINST /DISTFTR At the workstation, install the full product from diskettes: SPMINST /API for example. There are changes made to CONFIG.SYS. C'est la vie. The workstation must have LAN Requester + Peer Services to monitor a remote system such as the LAN Server. You must also increase the NetBios configuration parameters for the workstation and maybe the server. See the online SPM reference once you have installed for more details. ********************************************************************** ═══ 200. Look for a Sample of SMP/2 Output ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hi! Say I was looking to provide an example of the types of reports and information that spm/2 could capture for a server, might anyone know of a good summary file. Might anyone have something online which they could share with me? Thanks! Lou Rivas 8-391-4771 ********************************************************************** ═══ 201. What has to be running to monitor other stations ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Other then having the distributed feature, or the full product installed on other OS/2 lan machines. What has to be running to monitor other OS/2 workstations. Do you have to have OS/2 requester up? Thanks Ed ********************************************************************** ═══ 202. Remote Monitoring and Peer Services ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am in the process of installting SPM/2 and require remote monitoring to be established to an OS/2 LAN Requester V3.0 from a LAN Server V3.0. I have read previuosly in this forum of two appends that have since been removed and archived that would be of great help to me, these are: 'Remote Monitoring doc confusing 16:42:09 93/03/30' and 'SPM/2 and Peer Services 15:16:07 93/04/16'. I noted at the top of this form that they are available from the OS2 BBS, but I have no access to this bullent board. Is there any way I can obtian these append from other sources? Thanks.................................................Jack ********************************************************************** ═══ 203. convert spm2 data into DBF format ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Has anyone written a program which reads SPM2 data from a LOG file and puts them into DBF format? (Or something similar.) It would be helpfull for me. Thanks. ********************************************************************** On a PS/2 Model 95 with 32MB of RAM, SPM/2 Graph is only showing RAM graph of 15.6MB. What fix(es) do I need to have SPM/2 see all of the memory in the system? Theseus sees all 32MB of RAM. Thanks, James Lenaburg ********************************************************************** 1) Are you graphing "working set" when the graph only show 15.6M? 2) Where in THESEUS2 are you looking when you say: "Theseus sees all 32MB of RAM"? (There are several places where system memory is shown.) James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 204. What watchit does! ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * WATCHIT gathers LAN SERVER data that can be collected (off line) to identify resource usage not system usage for the application but not the total machine. LAN server performance problems are oriented to the app (LAN SERVER) and not the system. A demo is on the info and doc section of this bbs or call us for detail. (203) 233-2951 Labenski CSN Inc Home of WATCHIT and CONDUIT (203) 233-2951 ********************************************************************** The demo for version one of WATCHIT is still available in the information and Documentation of this board. WATCHIT provides a tool to collect response, session and alias usage of your server. Labenski CSN Inc Home of WATCHIT and CONDUIT (203) 233-2951 ********************************************************************** ═══ 205. SPM/2 on a LAN ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Reply-To: KARIUS at WATSON.IBM.COM We have SPM/2 installed on a LAN server and grant Read and Execute (RX) rights to the program directory. The problem is that everytime someone starts SPM they have to change the drive and directory to select/create a Monitor File since it defaults to the network drive that SPM/2 is accessed on. It also does not let users write to a NODE.INI file since SPM wants the file on the server also. Is there a way to specify a working directory other than the program directory for SPM/2. I've tried doing this the simple way by changing the working directory to a local directory in the program object but then SPM doesn't even start. LIBPATH, PATH, and DPATH are all set in CONFIG.SYS. I'd like to start SPM/2 from the network and have the default directory for the monitor files and the application's datafiles (NODE.INI, file where colors info is kept etc.) be a local directory. Is it possible to do this? Regards, Matthew Karius - Watson Office Applications From: Matthew G. Karius 914-945-3503 (t/l 862-3503) KARIUS at WATSON ********************************************************************** ═══ 206. SPM/2 not working ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Ref: None From: Andy Grieve To: All I've been having problems getting SPM/2 V2.0.1 workin on an OS/2 V2.1 system. When I select a graph for the local station, it starts up, but nothing appears on the graphs. I've just tried it on a new, format the disk, install of OS/2 V2.1, followed by installing SPM/2, noy joy, installed CSD WR06075, no joy, installed the OS/2 V2.1 Service Pack (UK version), still no joy. Anyone got any ideas? The machine's I'm testing on are a PS/2 95-AJF and a PS/2 65-121, both have IBM Token Rings cards, but the 65 has NO LAN software or drivers loaded. Any one else running V2.0.1? Thanks, Andy Grieve. * EMEA OS/2 Developer Assistance Program BBS * gHostNet(0057002C01000000 2db57618) ********************************************************************** Which machine is not graphing? Monitoring LOCAL with no LAN software is OK. I'm running SPM/2 2.0.1 on OS/2 2.1 or 2.11 and can do it local or with LAN running. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** [Reference to append at 14:01 on 94/04/20 by EMEADAP at WINVMB ] Here's the situation for SPM 2.0 LOCAL COLLECTION only, based on my experimentation. Requester Logged On Workstation Result Started Name in SPM ---------- ----------- ------------- ------------------------- No * No LOCAL Works like a charm. No No Machine Name SPM0298 Yes No LOCAL SPM0529 Yes No Machine Name SPM0298 Yes No LOCAL Should work. Yes * Yes Machine Name Works like a charm. * If you stop LAN Requester and restart it, you must stop and restart SPMNBL. -- From the SPM 2.00.1 Readme.inf. ********************************************************************** ═══ 207. Missing Records ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am monitoring a workstation cross-domain from a Server (NO peer services). Both PC's are OS/2 2.01 with CSD 6055. Monitoring appears to run fine (measuring every second), but when I look at a tabular report, I see spots of missing data (ie I'll get a row for every second, then it will skip up to 10 seconds and then continue). I know that when I'm missing data, it is usually when the monitored PC is busy. Is there an IBMLAN.INI setting I should be adjusting? Thanks David Kerr - Advisory S.E. - Insurance Services and Solutions IBM Canada - Toronto, Ontario - (905) 316-6664 ********************************************************************** It is possible for SPM/2 to get locked out for several seconds at a time, especially if the paging activity is high. I see it on my graph as long periods (up to 30 seconds) of not updating and then a jerk with the lines drawn from where it left off to the new values at the new time. I don't know of anything that can be done about this. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** If I set PRIORITY=ABSOLUTE and/or play with the TIMESLICE parms within the CONFIG.SYS, I SHOULD be able to force the system to give SPM some time to write some records. The question is : how much time does it need to do that? I know that fiddling with those settings isn't a good idea necessarily in a production environment, but I'm still testing, and REALLY want to see what's going on within the system... Comments anyone? David Kerr - Advisory S.E. - Insurance Services and Solutions IBM Canada - Toronto, Ontario - (905) 316-6664 ********************************************************************** Further update : I added PRIORITY=ABSOLUTE and PRIORITY_DISK_IO=NO to the CONFIG.SYS. SPM/2 overhead DROPPED to 5 to 6% (from 15% and more), I captured nearly all the records I wanted, and my application performance IMPROVED! I still have some research to do, but this was a bit of a surprise.... David Kerr - Advisory S.E. - Insurance Services and Solutions IBM Canada - Toronto, Ontario - (905) 316-6664 ********************************************************************** ═══ 208. HPFS386 cache effectivness ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * On my LAN Server, I have collected SPM/2 data specifying LAN Server 3.0 advanced in the LOG. I would like to get the same cache effectiveness info that I could get with the base OS/2 HPFS driver in the tabular report. Since it seems (based on things that I read in previous appends) that the HPFS386 data is only available in the DUMP report, can someone definitely confirm this? All I'm interested in an overall cache effectiveness (hit %) by time. If the only way to get this is to compute it myself, then does anyone have a spreadsheet already written for the data that is available to compute HPFS386 cache effectiveness? Gary Eheman - IBM PC Company Server Solutions ********************************************************************** Here's the formula to calculate LAN Server 3.0 Advanced (HPFS386) cache hit ratio. I'll send you a document with some other formulas as well. Unfortunately, in SPM 2.0 there's no report on HPFS386 cache hit % in either the Summary or Tabular report, and no way to conveniently import the data from the Dump report into a spreadsheet. CacheReadHit% = bcCaRdReqC / (bcCaRdReqC+bcCaRdReqD) CacheWriteHit%= bcCaWrReqL / (bcCaWrReqD+bcCaWrReqL+ctCaForceL) HPFS386 CacheHit% = (bcCaWrReqL+bcCaRdReqC) / (bcCaWrReqD+bcCaWrReqL + ctCaForceL+bcCaRdReqC+bcCaRdReqD) ********************************************************************** ═══ 209. SPM/2 Overhead ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I have been doing some remote monitoring of a workstation, measuring process-level data, and collecting every 1 second. When I look at the summary reports, SPM seems to be consuming about 15% of the CPU on the monitored PC. Is that normal? I would have hoped that it would be lower since I was running remotely.... David Kerr - Advisory S.E. - Insurance Services and Solutions IBM Canada - Toronto, Ontario - (905) 316-6664 ********************************************************************** 15% for SPM/2 when collecting process level data at a 1 second rate seems about normal to me. In fact, I'm surprised its not a little higher. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 210. SPM/2 V2 Incorrect Totals ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am looking for some guidance on reconciling information in an SPM Report. I have found nothing in the last 6 months of this forum, etc. My server reports 35 seconds CPU busy, 10 seconds of Interrupt Svc time. I have a number of processes reporting abut 15 seconds of CPU time, leaving 10 seconds unaccounted for. Where are they?!? My server is a PS/2 9595 xNx model, running OS/2 V2.1 with LS 3.1. I know that the HPFS386 file accesses will not be recorded, but what about the CPU required to handle the accesses (there's a lot of 'em)? I also have read that some of LAN Server and Requester code is counted in the Interrupt service time. Are there any guidelines to let me know what is reasonable? I have run this with and without the CSD from the TOOLS disk. Thanks. ********************************************************************** ═══ 211. How can I get a copy of the latest installation disks? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Reply-To: KARIUS at WATSON.IBM.COM I have a licensed copy of the SPM/2 2.0 disks and the SPM/2 CSD disks. I am planning to reinstall SPM/2 and would like to some how get a set of SPM 2.01 disks so that I don't have reinstall both the product and CSD. Also, the latest documentation I have is dated October 29, 1992. Is there a newer version of the documentation available in hardcopy (LIST3820 files perhaps). Regards, Matthew Karius - Watson Office Applications From: Matthew G. Karius 914-945-3503 (t/l 862-3503) KARIUS at WATSON ********************************************************************** ═══ 212. 3514 DISK ARRAY ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Is it possible to monitor a server with the 3514 disk array? (This question was posed to me. I have not tried this myself.) Thanks! ********************************************************************** ═══ 213. Printing out Elapsed times ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * After using SPMStartTimer and SPMStopTimer to time a path, what are the proper SPM function calls to be used in conjunction with "printf's" to display the elapsed time in the application itself (with as much granularity as is meaningful) rather than looking at an SPM Report Dump file ? > For that matter, how do you print out the current time (QWORD) > obtained from using the SPMQueryTime function ? Should the overhead of starting and stopping the SPM timer also be measured and subtracted from the path timings obtained ? Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** I would do a few start/stop pairs to see what the overhead is. Since the lower ULONG will hold about 4000 seconds (4 Gigaticks), you can use it by itself and convert it to whatever you want. SPMQueryFreq will give you the number of ticks in a second, so you can convert the deltas to ms or us as you please. James Summers (OS/2 Performance & THESEUS2) ********************************************************************** ═══ 214. (no subject) ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * message Ref: Append at 14:40:34 on 94/04/13 GMT (by F077072 at DCTVM1) My customer looks like they have a similar situation with one major twist...one that I really need help with! In loooking at SPMSNAP.ERR, I see seeral occurances of the 4 messages (SPM0302/511/513/514) which appear to have no outward symptoms (that is, the machine was not down or anything) However, the past 2 mornings, their managed server has hung at exactly 8AM each morning (screen goes blank, LMU/2 reports heatbeat overdue, noone can logon to server, keyboard/mouse dead). They noted that the above messages appear in the log at 8AM each morning. To add to the confusion, they just began autmatically starting DCF data collection on Monday. It starts at 7:55AM. WHile it worked fine Monday & Tuesday, the machine hung yesterday and today at 8AM (just after the autostart of data collection). The customer is concerned that it is DCF causing the failures. My question: SHould we interpret the messages above as DCF causing a problem or DCF reacting to the problem? NOTE: Each day we have rebooted the hung server, waited an hour or so and then reinitiated DCF data collection and everything has been fine for the rest of the day. Does anyone have any ideas, or can anyone confirm that these messages are not indicative of SPM causing the hang? Rick St. Onge One Commercial Plaza Sr. Networking Marketing Specialist Hartford, CT 06103 IBM Connecticut RICKST @ BOSTON or USKIB2794 @ IBMMAIL ********************************************************************** ═══ 215. spmsnap.dll locked on application of WR06075 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I'm hitting this during the install of the Service Pack. I have installed this before with success, using SPMNBL /HALTALL and SPMNBL /STOPLISTEN as directed to quiesce the system before the install. This problem is on a Pentium system. Any ideas? The process, on indication that this file is locked, tells me to reboot the system off the A drive...but offers no clue what to do after that is done. Dave LeGrys ********************************************************************** ═══ 216. disk utilization 100.02% on LS 3.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Our customer monitored one LS 3.0. The disk graphs showed a very high utilization about 4-6 times a day on all four physical disks for about 2 minutes. The report (summary) for this intervall: % Util Access Read/Write (Requests Sectors ........) Rate Disk 01 100.02 0 0 0 0 0 Disk 02 100.02 0 0 0 0 0 Disk 03 100.02 0 0 0 0 0 Disk 04 100.02 0 0 0 0 0 All the other time (which we are not interested in), the report is ok, and all the % util, reads and writes are meaningful numbers. The same results are in the dump file: if the report is generated over this 2 minutes, we get a list of files, but no reads/writes. Enviroment: LS 3.0 7005, SPM/3 6075, OS/2 2.1 without SP HPFS386 Can anybody help me with this ? Wolfgang Mild, PC-Services Austria / ATSPCS18 at VIEVMA ********************************************************************** >The report (summary) for this intervall: > > % Util Access Read/Write (Requests Sectors ........) > Rate >Disk 01 100.02 0 0 0 0 0 >Disk 02 100.02 0 0 0 0 0 >Disk 03 100.02 0 0 0 0 0 >Disk 04 100.02 0 0 0 0 0 I don't know why this is happening. Could you provide some more info? Either the data is right and OS/2 is acting weird, or the data is wrong. Can you tell if there is any activity on this server during this period, like the disk light on or work getting accomplished? Did you really collect file I/O data and run a Summary Report? Do all files show zero activity? What is the date on your C:/OS2/OS2DASD.DMD file? Can you send your log file to SARACINI at ROMEPPC for further analysis? We have a tool called spmlogpr that will format a log file. ********************************************************************** ═══ 217. How to use the SPMAPIOpen function of the SPM╪s API? ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * In page 7-34 (Chapter 7. SPM/2 Application Programming Interface) of the SPM/2 User╪s Guide and Reference (December 4, 1993)I can read: "SPMAPIOpen: Use the SPMAPIOpen function to open the connection to an active or INACTIVE monitor session (.LOG) file so your application can begin the input processing of data from that file." What means INACTIVE? I supposed that INACTIVE means that it is NOT necessary to issue an SPMAPIInit function before issuing the SPMAPIOpen function.But if I don╪t do so the SPMAPIOpen function fails. Where am I wrong? All I want to do is to get historical data stored in a monitor session (.LOG) file. ********************************************************************** ═══ 218. Support for SPM 2.0 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Effective immediately, support for the SPM 2.0 product has shifted from IBM Austin to IBM Networking Systems in Raleigh NC and Rome Italy. Development support will be handled by Andrea Saracini in Rome; SARACINI at ROMEPPC. ********************************************************************** ═══ 219. watchit license ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Even with multiple licenses it is best you install the tool on each server so the session, alias and logon activity data collected be on the server that is associated with the data. Note the sample data name is a composite of MMDDMMM which is M=Month D=DAY M=minutes from midnite. This naming convention will not allow easy seperation of data to the problem server. Labenski CSN Inc Home of WATCHIT and CONDUIT (203) 233-2951 ********************************************************************** ═══ 220. System Hangs and requires Power On Resetwith LAN Netview ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I am in the process of trying to get a LAN Netview Managed system running the performance agent at level IP07000 to run with SPM/2 V2.00.1 level WR06075 wich I require to do data collection for LMU/2. When I start SPM/2 and LMU/2 I have no problem collecting performance data and sending alerts to LMU/2, but when I start the LNV code the system runs for approx. five minutes and then I get an indication that my LMU/2 agent has failed, and when I get to the machine it is locked up and I must power of in order to get the system back. It appears to be a problem between SPM/2 and LAN Netview. Any help in this area would be greatly appreciated. Has anyone else run into this problem? Thanks and regards......................................Jack ********************************************************************** SPM/2 and LAN NetView were not designed to work together, since it was thought that either one or the other would be used but not both. But I see that you need to use LMU, which uses the SPM Data Collection Facility (DCF), with LNV. The only incompatability between SPM and LNV is the performance DCFs; OS/2 supports only one DCF. I haven't yet determined if just loading SPM or LNV causes their respective DCFs to start, but actual perf data collection does for sure. Anyway, here's the workaround I think. When you start LNV, don't start the PerformanceAgent. In /LNV/ORS/SVSUF, make the performanceagent and performancelog processes non-startable by putting a 1 in column one. ********************************************************************** ═══ 221. Documentation regarding experiences of using SPM/2 ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Hello, Does anyone in the SPM/2 2.0 world if there is any documentation regarding experiences of using SPM/2 2.0??? We have installed it in order to tune the machines at our installation The installation goes fine but we think that the problem regards with the right comprehension of the information and the reports you obtain. We would like to know if there is any kind of documentation regarding these subjects: - Range of values (CPU, DISK, MEMORY) that could be dangerous in case they are reached. - Any experience and/or advise of the use of SPM/2, results obtained, examples......... Thanks in advance Raquel Balaguer Ortiz 83811246 at spavm1 ********************************************************************** There is no information available like you describe. Would anyone like to volunteer? Here are a few sources of information on using SPM. Send me your address and I'll mail you the first one. SPM/2: Fine-Tuning Your Application, OS/2 Developer magazine, Sept/Oct 1993. SPM VS. LNV MONITOR, a white paper by Laura Adams CICS OS/2 V2.0 Sizing and Performance GG24-4190 Client/Server Data Collection / Cap. Planning for OS/2 & AIX ZZ81-0342 ********************************************************************** How could I get a copy of the following reference from your append ? SPM VS. LNV MONITOR, a white paper by Laura Adams Dennis L. Price (919) 301-6257 (T/L 352) ********************************************************************** [Reference to append at 14:31 on 94/05/06 by HOLCK at AUSVM1 ] Hello, thank you very much for your reply to my append. I am sending you my mail address in a note. Raquel Balaguer Ortiz 83811246 at spavm1 ********************************************************************** ═══ 222. How does an IBMer buy SPM/2!! ═══ * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * I called the 1-800-IBM-3040 to order it, and the person that answered the phone told me they no longer carry SPM/2. And didn't know how I could buy a copy! Help. I need to buy a copy ASAP. Ming Dong