home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.unix.questions:9403 comp.unix.sysv386:12397
- Path: sparky!uunet!gatech!usenet.ins.cwru.edu!ncoast!fmsystm!bmhalh!actcon!kearns
- From: kearns@actcon.canton.oh.us (Steve Kearns)
- Newsgroups: comp.unix.questions,comp.unix.sysv386
- Subject: PS/2 SCO XENIX woes
- Keywords: PS/2,problems
- Message-ID: <1992Jul23.215618.962@actcon.canton.oh.us>
- Date: 23 Jul 92 21:56:18 GMT
- Organization: A.C.T. Consulting, Inc.
- Lines: 101
-
- Fellow netters,
-
- My company installed an SCO XENIX system about 8 months ago on an old
- IBM PS/2 model 80 (micro-channel architecture). Since that time we have
- had nothing but problems; along several fronts. Let me detail them...
-
- HARD DISK:
-
- We've had two major hard disk problems, but first a brief history of
- the hard disk. The system has its original 70mb ESDI drive. When we
- initially installed XENIX, badtrk reported 2 bad tracks and locked them
- out of use. The system ran fine for about 3 months. Then, in February
- of this year, the system started barfing out ESDI errors at the console
- and died two hours later (the customer was ignoring the messages).
-
- One of my fellow employees arrived and ran badtrk; this time it found
- 28 bad tracks. We allowed it to lock them out and restored their system
- from the previous night's backup.
-
- Things were fine again for a couple more months, then in April the system
- again died. This time, the customer blamed us for their problems and
- had another vendor "repair" the machine; this repair involved the
- same process, accept this time they reformatted the drive and completely
- re-installed XENIX. They told our customer that we had originally
- installed XENIX wrong and that there was no hardware problem (ha!).
- (The customer's system crashed recently, and this time their "other
- vendor" told them it was because we configured the UPS wrong; of course
- this is a UPS without an auto-shutdown ability :-) ).
-
- Strangely, after they reformatted the drive, badtrk reported no errors
- at all. About a month ago, the system again started spitting out
- ESDI errors and crashed. The errors are of the following form:
-
- esdi: ERROR: on fixed disk dev=1/40 block=5
- cmd=00000002, isr=0000000C, sir=00000702
- drive=0, sector=10, cylinder/head=1/0
-
- esdi: ERROR: on fixed disk dev=1/40 block=47
- cmd=00000002, isr=0000000C, sir=00000702
- drive=0, sector=30, cylinder/head=1/2
-
- esdi: ERROR: on fixed disk dev=1/40 block=38781
- cmd=00000002, isr=0000000C, sir=00000702
- drive=0, sector=26, cylinder/head=38/55
-
- We've had about 60 such errors over the past eight months. Each time
- this happens, I run badtrk (intensive, non-destructive) and each time
- it finds no bad tracks (the bad track table is empty). Does anyone
- know how to interpret this error message (it's on the console and in
- /usr/adm/messages)? NOTE: there is no consistency in the cylinder/head
- numbers; all errors logged are unique.
-
- My second hard disk problem deals with a Procomm Technologies SCSI
- drive and controller for the PS/2. We purchased Procomm's XENIX
- drivers and installed a 200mb drive as /usr2. The drive works fine; for
- about 3 days, then all processes which attempt to access the drive will
- get "hung." The processes are hung to the point that kill -9 will not
- even work. I can't even shut the system down, as the shutdown eventually
- hangs and I have to power off. I've also found that performing two or
- three df's in a row will cause the system to hang as well.
-
- I've done about 10 rounds with Procomm Technologies who have no answers
- or advice (actually, they don't even return my calls anymore). Has any
- one out there every used this configuration before?
-
-
- FOXBASE:
-
- My next category of problems is SCO Foxbase (version 2.1.1). I am having
- very unusual and inconsistent problems with Foxbase. Strangely, I have
- never been able to replicate these problems on XENIX running on an ISA
- machine nor on our Altos 2000 machine. Problems:
-
- - a program which works for days, weeks, months, will suddenly
- develop a "memory variable not found" error message. Syntax
- wise the programs are fine; I've been over them a thousand times.
- It seems that Foxbase is losing some of its internal variables
- (yes, I've changed all my assignments which used '=' to STORE's).
-
- - I sometimes get "file in use" while trying to open a database.
- Other process will have the database open, but no exclusive
- opens are ever being done. I'm very careful not use any database
- operations which lock the entire database.
-
- - Every now and then, especially when printing, a user will get
- kicked all the way out of Foxbase with an "Invalid Signal"
- message (no, they didn't hit the break key).
-
- Like I said, I've only had these problems on the PS/2 MCA machine; never
- in DOS or on an ISA box.
-
-
- Any help or insight offered will be greatly appreciated.
- Thanks in advance.
-
- steve
- --
- Steve Kearns A.C.T. Consulting, Inc. 216-455-1444
-
- Internet: kearns@actcon.canton.oh.us
- UUCP : uunet!aablue!redpoll!mrsmouse!actcon!kearns
-