Internet Info 1994 March

home *** CD-ROM | disk | FTP | other *** search

/ Internet Info 1994 March / Internet Info CD-ROM (Walnut Creek) (March 1994).iso / standards / ansi / X3T9 / area01 / 89-049r0.brf < prev next >

Wrap

Text File | 1992-05-12 | 34.3 KB | 738 lines

Final Agenda -- X3T9.2 Meeting #81 April 24-25, 1989 -- St. Petersburg Beach, FL 1. Opening Remarks 2. Approval of Agenda 3. Attendance and Membership 3.1 Roll Call of Members in Jeopardy 4. Approval of Minutes - February Meeting Austin, TX (X3T9.2/89-028) 5. Document Distribution 6. Liaison Reports 6.1 ISO 6.2 IT8 6.3 SCSI-2 Common Access Method Committee Report 6.4 Fiber Channel 7. Review of Old Action Items 8. Working Group and/or Project Status Reports 8.1 ESDI 8.2 SCSI General Working Group (X3T9.2/89-042) 9. Old Business 9.1 Cable Testing Results 9.2 Tape READ underlength detection (X3T9.2/89-044) [Chan/Bramhall] 10. New Business 10.1 Resolution of X3T9 comments on SCSI-2 10.2 AWRE and ARRE bits in READ-WRITE ERROR RECOVERY Page [Lamers/Lohmeyer] 10.3 Exabyte request for new Density Code (X3T9.2/89-043) [Duran] 10.4 Review of new documents 10.5 Agenda for the Wichita Working Group 10.6 AMP Patent #4,808,125 (X3T9.2/89-46) [Goodman] 10.7 Should REQUEST SENSE be required to use the same queue tag? [Houlder] 10.8 LOGICAL UNIT RESET Message (X3T9.2/89-47) [Lohmeyer] 10.9 Command Queuing, ECA, and Error Conditions [Nitza] 11. SCSI-3 Activities 11.1 Autoconfiguration SSWG 12. Review of Action Items 13. Meeting Schedule 13.1 General Working Group Schedule for 1989 14. Adjournment Minutes -- X3T9.2 Meeting #81 April 24-25, 1989 -- St. Petersburg Beach, FL 1. Opening Remarks John Lohmeyer, the chairman, called the meeting to order at 9:00 a.m., Monday April 24, 1989. He thanked Chuck Brill of AMP for hosting the meeting and he also thanked Classic Conferences for arranging the meeting. As is customary, the attendees introduced themselves. A copy of the X3T9.2 membership list was circulated for attendance and corrections. Copies of the draft agenda and the recent document register were made available to all attendees. Information on X3T9.2 and Pilot Subscription forms were made available for new attendees. 2. Approval of Agenda The draft agenda was approved with the following addition: 10.9 Command Queuing, ECA, and Error Conditions [Nitza] 3. Attendance and Membership The membership requirements were reviewed. X3 rules permit one vote per organization. An individual from a new organization must attend one plenary meeting as an observer before he/she may apply for voting membership via a letter to the chairman of X3T9.2. The individual may vote at the second plenary meeting. Working group meetings do not count toward attendance requirements. Kate McMillan of the X3 Secretariat sent the chairman a copy of the "Control Sheets" for X3T9.2 showing that a lot of members have not paid their 1989 service fees. Final notices will be sent out the first week of May and those who have not paid by June 15 will be dropped from the committee lists. The chairman made the control sheets available for examination during the meeting. To make corrections, contact Katrina Gray at the X3 Secretariat (202-626-5741). The chairman stated that since the last meeting the following changes had occurred in the voting membership: Kim Ryal has replaced Michael Bryan as principal from Miniscribe. John Mangan attended for Miniscribe and replaced Kim Ryal who had left the company. Bill Medlinski has replaced Steve Heissenbuttel as the alternate from Panasonic. The chairman said that Toshiba had requested voting membership in X3T9.2. Since his records did not show anyone from Toshiba attending a recent meeting, he had sent a letter to Kenneth Baun saying that they would need to attend two meetings to be eligible to vote. Dal Allan remembered that Kenneth Baun had attended the Austin meeting, but apparently failed to sign the attendance sheet. Thus Toshiba is eligible to vote at the next meeting they attend. (Toshiba did not attend the St. Petersburg Beach meeting.) Chairman's note: Kenneth Baun has left Toshiba. Another Toshiba representative may attend the San Jose meeting. During the meeting several membership changes were received: Chris Nieves of Computer Consoles (an ICL subsidiary) replaced Tony Salthouse as the principal. Tony Salthouse of ICL became the alternate. Bob Snively of Sun Microsystems replaced Curtis Ridgeway as the principal. Curtis Ridgeway and Paul Rikkonen became alternates. Dexter Anderson, Vit Novak, and Chuck Campbell became observers. Zenith Data Systems joined the committee with Saied Zangenehpour as principal and Thomas Colligan, Greg Leonhardt, and Greg Mart as alternates. Enclosure (1) is the list of attendees at the meeting. Enclosure (2) is the X3T9.2 membership changes since the last minutes and enclosure (3) is the current X3T9.2 membership list. 3.1 Roll Call of Members in Jeopardy The chairman stated that he had sent jeopardy letters to all representatives of the following organizations: Ballard Synergy Corp., Cipher/Optimem, Hirose Electric U.S.A., ICL, Computer Consoles, Intergraph Corp., ITT Cannon, JAE, Panasonic Industrial Co., Peripheral Technology Inc., and Viking Connectors. The letters said that these organizations had failed to attend two of the last three plenary meetings and that their membership would be terminated unless they attended the February plenary meeting. Representatives were present from all of the organizations except: Ballard Synergy Corp. Peripheral Technology Inc., and Viking Connectors. The individuals from these organizations were changed to observer status. 4. Approval of Minutes - February Meeting Austin, TX (X3T9.2/89-028) On Page 22 of the minutes, the spelling of Ricardo Dominguez's name (Texas Instruments) was butchered. The minutes were approved as amended. 5. Document Distribution The last mailing was 908 pages which were mailed in three separate envelopes. The chairman had not received any complains about the mailing other than from a few people who received their mailing out of order. Since there was no cover page on the second and third sections, it was not obvious what was going on. Members with mailing subscription problems should deal directly with Katrina Gray of the X3 Secretariat. She can be reached at (202) 626-5741. John Lohmeyer requested that all documents for the May mailing be sent to him by May 15, 1989. The following new or revised documents were distributed and/or discussed at the meeting: Document Doc Date Author Description of Document ------------- -------- --------------- --------------------------------------- X3T9.2/89-38 4/21/89 V. Novak SCSI-2 Cable Testing Report Rev 1 X3T9.2/89-39 4/22/89 Rikkonen et. SCSI-2 Cable Testing Summary Report Rev 1 al. X3T9.2/89-46 4/5/89 D. Goodman Letter concerning AMP patent #4,808,125 X3T9.2/89-47 4/11/89 J. Lohmeyer Overriding Device Reservations and Pending Commands X3T9.2/89-48 4/24/89 B. Snively Sun Cable Testing Results X3T9.2/89-50 4/24/89 D. Davies Proposed changes and corrections to SCSI-2 R8, Section 9 X3T9.2/89-51 4/11/89 Campbell/ Kalin Measurements of REQ&ACK using 3M connector X3T9.2/89-52 4/12/89 Campbell/ Kalin Measurements of REQ&ACK using AMP connector X3T9.2/89-53 4/21/89 T. Wicklund SCSI-2 R8 comments X3T9.2/89-54 4/21/89 D. Appleyard MESSAGE PARITY ERROR message X3T9.2/89-55 4/16/89 G. Milligan Comments accompanying his X3T9 letter ballot X3T9.2/89-56 4/24/89 D. Smith Letter saying AMP patent #4,808,125 outside scope of SCSI-2 draft stnd. X3T9.2/89-57 4/20/89 Chan/ Matheson AMP SCSI-2 Cable Comparisons X3T9.2/89-58 4/14/89 J. Lohmeyer NCR "No" vote on SCSI-2 letter ballot (X3T9/89-41) X3T9.2/89-59 4/20/89 J. Lawlor July working group meeting announcement X3T9.2/89-60 4/20/89 D. Allan United and Northwest discount air fare numbers Enclosure (4) is the current 1989 document register. 6. Liaison Reports 6.1 ISO Gene Milligan gave the following report: The DIS letter ballot on 9314-3, FDDI PMD closed April 20, 1989. DIS 9318-4, IPI-3 Device-Generic Tape, opened on March 9, 1989 and closes on September 9, 1989. SC-13 agreed to accept SCSI-2 as a draft proposal upon its approval by X3T9. The proposed US delegation to SC-13 was to meet later in the week. There had been a meeting April 21 of several officers of SC-83 and SC-13 to discuss the potential merger of SC-83 and SC-13. They were attempting to reach an agreement on meeting arrangements and working responsibility. If the merger occurs, the merged group was to be called SC-25. The currently scheduled meetings were planned to occur as if there were no merger. The chairman of X3S3 had requested that the USA oppose the merger until their constituency has an opportunity to determine their position. SC-13 plans to meet September 25-29 in Copenhagen, Denmark. SC-83 plans to meet November 1-3 in Venice, Italy. 6.2 IT8 There was no report. 6.3 SCSI-2 Common Access Method Committee Report Dal Allan reported. The CAM committee has had two meetings and three working groups since the last X3T9.2 meeting. The AT bus register set effort has made the most progress. Gene Milligan's and Tom Treadway's documents will be merged into one document by Dal Allan. The operating system dependent functions needed for CAM were discussed in Denver and progress is occurring. The transport layer is also making progress. The next meeting is May 10, 1989 in Wichita, KS, the day following the SCSI working group meeting. 6.4 Fiber Channel Dal Allan reported. The Fiber Channel working group met with only Dal representing the SCSI interests. Dal was concerned that the needs are being driven by the high-end interests and that without more participation from people with interests in using Fiber Channel for SCSI applications, the architecture adopted may become very expensive for both the copper and fiber versions. Three fiber channel layers were planned: FC0 - physical, FC1 - encoding/decoding/framing, and FC2 - interface to applications (SCSI, IPI, HSC et al). A mechanism was developed to detect whether a destination is attached and can accept information from the source. The next fiber channel meeting was planned for May 22-23 in Minneapolis. 7. Review of Old Action Items 1. Paul Boulay will develop an appendix on single-ended terminators. Carry over. 2. Larry Lamers will prepare revision 8 of SCSI-2. Complete. 3. John Lohmeyer will request X3T9 letter ballots on forwarding SCSI-2 and ESDI. Complete -- ESDI was forwarded by an unanimous meeting vote and the SCSI-2 letter ballot closed on 4/17/89. The X3T9 voting results are covered under item 10.1. 8. Working Group and/or Project Status Reports 8.1 ESDI The second ESDI Public Review period began April 21, 1989 and it ends June 20, 1989. Copies of this document (X3.170-198x) can be purchased from Global Engineering Documents at 800-854-7179. Any public comments should be addressed to Marilyn Kornfeld, X3 Secretariat, 311 First St., N.W., Suite 500, Washington, DC 20001. 8.2 SCSI General Working Group (X3T9.2/89-042) Twenty-four people attended a working group meeting hosted by Jim McGrath of Quantum on March 6, 1989 in Milpitas, CA. The principal work accomplished was to review the interim revision 8 document that Larry Lamers had prepared based on the February plenary meeting. While the last plenary meeting agreed to forwarded revision 8, it did not exist on paper. What was actually forwarded, was revision 7 plus marked-up changes. The exact wording of most of these changes was not reviewed at the February plenary meeting. Thus the working group was acting as a quality control group. The working group also dealt with clarifying an issue about a SCSI-2 device that responds to SCSI-1 protocol. Larry Lamers asked whether a SCSI-2 device is required to honor all of the SCSI-2 protocol after it has responded to the single-initiator option of SCSI-1. This option was removed from SCSI-2, but SCSI-2 devices are permitted to respond to it in support of SCSI-1 hosts. The working group felt that a target which responds to SCSI-1 protocol should be bound by the SCSI-1 rules. Thus the target would not be required to support all of the SCSI-2 messages if it was talking to a SCSI-1 host. A clarification was added to the implementors note saying that the target may respond as an SCSI-1 device. A correction was made in Section 9 where some the the values for the density code corresponding to QIC 320 were in error. Bill Duran of Exabyte presented a document requesting a new tape density code. While the working group generally favored adding the code, procedurally it did not have the authority to make such a change after the document was forwarded. (See agenda item 10.3.) The working group spent a few minutes reviewing the list of ideas for SCSI-3. Several new items were added to the list including: * Autosense [Nitza] * Apple request for Data DAT device type * Documentation Layering [Gary Stephens] * Single-cable 16-bit wide SCSI * Alternate physical layers (e.g., fiber optics) 9. Old Business 9.1 Cable Testing Results Bob Snively gave a brief overview of Sun's cable testing efforts. Vit Novak presented a comparison of high-density round versus low-density flat cables. Three physical test configurations were used (see 89- 038R1). The tests were done using both asynchronous and synchronous data transfers. The test cable was 28 AWG twisted-pair round cable. With only two SCSI devices (one initiator and one target), all configurations passed a read-write-verify test of 1011 bits transferred. With three devices all configurations passed, but synchronous data transfer with 70 ohm cable required a Vterm of 4.40 volts in order to pass. With eight devices the high-density 70 ohm cable failed. High- density cable required a Vterm of 4.60 volts for the synchronous data transfer to pass; a Vterm of 4.28 volts was required for asynchronous data transfer to pass. Attaching a device to the high-density cable changed the impedance from 100 ohms to 75 ohms using a Time Domain Reflectometer to measure the impedance. Vit offered the following rationale for this drop: The abundance ground wires twisted around the signal wires cause the lowering of impedance in round cable. With flat cable the impedance only drops approximately 5 ohms. The thick shielded round cable showed about the same characteristics as the flat cable. The main difference was the cable jacket thickness. Alex Papas presented the second part of the Sun report (see 89-039R1) and spoke primarily about the "Configuration C" portion of the report. The oscilloscope pictures in the report show REQ on top and ACK on the bottom at each of the four test points. The waveforms shift as the signals propagate down the cable and actually look better at the far end. The end device failed frequently with a reported error message of selection time out. However, the actual failure mode had not been determined. It did not matter which device was placed at the end of the cable -- they all would fail in this position. Devices in other positions also failed with selection time out failures, but at a much lower rate. While selection time out failures were the most frequent failure mode (99%), some synchronous data errors also occurred (1%). Sun intends to investigate this phenomena further after they receive additional test equipment. Bob Snively summarized Sun's testing to date and their plans for further testing, including a test of higher impedance cables (120 ohms). The B cables may have a problem due to their larger number of conductors which may cause problems with the cable not fitting into the connector shell. Termination power was critical to correct performance; 4.5 volts was highly desirable at both ends of the cable. This required 28 AWG wire minimum. The current limiting devices must not have a large voltage drop. The receiver devices must have good characteristics. Some CMOS devices are adversely affected by undershoot. Bob had three recommendations: 1) An implementors note should be added to warn of the impedance drop problem with round cable and single-ended devices. 2) The receiver transition threshold should be specified at 1.4 volts dc + 0.1 volts dc. 3) Change the voltage requirement for Vterm to be 4.5 to 5.25 volts dc. Bob stated that the primary problem was cable impedance, not termination impedance. Paul Boulay's proposed termination was mocked up; the results were that it did not materially affect the near-end results. Although it should help with the Vterm problem. We should recommend a 1.5 amp fuse instead of the current 1.0 amp fuse for the A cable to allow for temperature and lifetime de-rating. More exotic fuses have higher voltage drops. The net result of the discussion was that the issues can be solved but they have to be addressed by the system integrator. The problems cannot be resolved by just making changes to the standard's requirements of the individual devices. Kurt Chan gave a report on the HP cable testing. They used TTL breadboards to simulate seven devices. The breadboards were designed to represent worst case characteristics of stub length and impedance. Skew effects were not taken into account, only signal quality was evaluated. Driving the terminator power from the middle of the cable, the test failed at 4.73 volts Vterm for the single-ended SCSI-1 configuration. Paul Boulay's termination scheme (2.85 volt 110 ohm) did not fail. John Morse presented the results of the DEC testing (see 89-051, and 89- 052). They used 130 ohm characteristic impedance flat cable. Bob Snively moved and Jim Schuessler seconded that the recommended implementors note on page 3 of X3T9.2/89-048 be accepted for inclusion into SCSI-2 in section 4.2. The motion passed unanimously. The question of the proposed addition of a transition threshold voltage and changing the Vterm voltage was assigned to the May working group. 9.2 Tape READ underlength detection (X3T9.2/89-044) [Chan/Bramhall] Document 89-21 from IBM proposed adding a SULI bit to only suppress the reporting of underlength conditions when reading tape blocks. IBM withdrew this proposal in favor of document 89-35 from Cipher which effectively changed the definition of the SILI bit so that it now only suppressed reporting of blocks that were smaller than the block length specified in the mode data. This proposal was accepted and the document was forwarded. After checking further, HP had discovered that the change does not work for some of their applications. These applications would now be forced to either issue a MODE SELECT command between each READ command or stop using the SILI bit and deal with the ILI statuses on each READ command. HP voted no on the X3T9 letter ballot urging that 89-21 be reconsidered as a better solution than 89-35. IBM also voted no at X3T9 citing essentially the same reasons. Dan Davies presented a proposed solution (see 89-050) that would resolve the problem without requiring both a SILI bit and a SULI bit. Both IBM and HP asked for time to check with their organizations to make sure we do not repeat history. Final resolution of this issue was postponed until the June plenary meeting. 10. New Business 10.1 Resolution of X3T9 comments on SCSI-2 The chairman did not have the final count of the X3T9 letter ballot on forwarding SCSI-2 to X3, however he did indicate that X3T9 had received the following ballots in addition to a number of "yes" votes with no comments: HP (My Le) No with one comment (on Tape underlength detection) IBM (Bob Dugan) No with one comment (on Tape underlength detection) NCR (John Lohmeyer) No with one comment (on ARRE and AWRE) EG&G (Arnold Roccati) Yes with 24 comments IMPRIMIS (Gene Milligan) Yes with 477 comments (Chairman's note: The final tally on the X3T9 letter ballot was 16 yes, 3 no, 0 abstain, and 4 did not return their letter ballots.) The chairman expressed appreciation to the people who read and commented on the document. He recognized that it took a lot of time, but he expected that it would result in a better standard. The chairman suggested that the plenary focus on resolving the no votes first. Then, time permitting, the comments from EG&G and IMPRIMIS should be addressed. Any remaining comments should be assigned to the May working group. He requested that the editor attempt to create a revision 9 document in time for the May mailing and that X3T9.2 plan to forward that document to X3T9 at the June meeting. The resolution of the HP and IBM negatives is under agenda item 9.2 and the resolution of the NCR negative is under agenda item 10.2 Most of the Roccati comments pointed out instances were the document failed to follow the ANSI style. While the comments were completely correct, the editors may postpone their incorporation until later in the process. This is because the document will likely be converted to the ISO style prior to publication. The comments concerning how section numbers should be formatted will also be delayed because the section numbering program does not have the necessary flexibility. The Roccati comments on the high-density connector dimensions will result in notes being added to the appropriate figures. The notes will point out that the socket dimensions are of the entire hole in the plastic and do not show the actual contact. Several hours were spent reviewing the Milligan comments. Most comments were purely editorial and were directed to the document editor. Other comments required group discussion to reach proper resolution. The first 151 comments were covered during the plenary meeting and the remaining comments were remanded to the May working group meeting for review. 10.2 AWRE and ARRE bits in READ-WRITE ERROR RECOVERY Page [Lamers/Lohmeyer] John Lohmeyer stated that these bits are documented in a way that permits the target to reallocate a logical block when the target is unable to recover the data. This would result in potential data integrity problems. He said that his proposed wording for the AWRE bit may contain some errors in that it is not necessary to recover the data from the medium if the data is still available in the target's buffer. His major concern was that the target not be given permission to irrevocably discard the data when making automatic reallocations. John Lohmeyer moved and George Penokie seconded that the AWRE and ARRE bit descriptions be corrected so that the target only performs these operations if data integrity can be insured. The motion passed unanimously. 10.3 Exabyte request for new Density Code (X3T9.2/89-043) [Duran] Bill Spence moved and Jeff Stai seconded that X3T9.2/89-043 be included in the SCSI-2 document. The motion passed with 24 in favor and 1 opposed. Dan Davies suggested that QIC-600 density code be replaced by the 1.3 GByte tape density specifications. There was general agreement that the QIC-600 density code would not be needed and that the 1.3 GByte should be included. Dan was directed to bring the appropriate values for the 1.3 GByte density code to the May working group meeting. 10.4 Review of new documents 89-050 - Davies proposed changes and corrections to Section 9 (see 9.2). 89-051 - Part of DEC's cable testing (see 9.1). 89-052 - Part of DEC's cable testing (see 9.1). 89-053 - Wicklund's comments on SCSI-2 Rev 8 - remanded to the working group. 89-054 - MESSAGE PARITY ERROR message - remanded to the working group. 89-055 - Milligan's comments on SCSI-2 rev 8 - partially addressed and the remainder remanded to the working group. 89-056 - AMP letter saying their patent does not have to be infringed to conform to the SCSI-2 draft standard. 89-057 - HP testing on the AMP SCSI-2 cable. 89-058 - NCR Comment on SCSI-2 rev 8 (see 10.2). 89-059 - July working group announcement 89-060 - How to avoid high airfare to working group meetings. 10.5 Agenda for the Wichita Working Group The following agenda was assembled for the May working group meeting: SCSI-2 Items 1. 89-055 Resolution of the X3T9 letter ballot comments on SCSI-2 R8 2. 89-048 Should we define a transition threshold voltage? [Snively] 3. 89-048 Change Vterm to 4.5 - 5.25 volts? [Snively] 4. 89-21,44,50 Resolution of the SILI/SULI issue. 5. Better definition of queuing error handling [Eneboe, Nitza] 6. 89-54 MESSAGE PARITY ERROR handling (archaeological dig for the resolution of 87-213). [Appleyard, Spence] 7. QIC-600 density code replacement. [Davies] 8. Redefine contingent allegiance (semantics) [Lohmeyer] SCSI-3 Items 50. 87-186 SEND DIAGNOSTIC Pages [Bill Spence] 51. 87-203 & 87-217 LOAD SKIP MASK command proposal [Greg Floryance, Dave McIntyre] 52. 87-206 More that 8 devices on wide SCSI [David Harms] 53. 88-002 Search Command modifications [Jeff Stai] 54. 88-007 Expanded RelAdr Bit Definition [Paul Boulay] 55. 88-69R1, 88-92, & 88-100 Autoconfiguration SSWG [Jerry Marazas, Paul Nitza, Jim McGrath, Paul Boulay] 56. 88-127 Error Handling Action Codes [George Penokie] 57. Autosense [Nitza] 58. Apple request for Data DAT device type 59. Documentation Layering [Gary Stephens] 60. Single-cable 16-bit wide SCSI 61. Alternate physical layers (e.g., fiber optics) 62. 89-058 LOGICAL UNIT RESET message 10.6 AMP Patent #4,808,125 (X3T9.2/89-46) [Goodman] Dave Goodman of ITT Cannon had indicated to John Lohmeyer that while AMP does not have patents on the "intermating dimensions" of its Amplimite 0.050 Series connectors, AMP has recently been issued U.S. Patent #4,808,125 which covers the construction technique of the back-end portion of the connectors. The alternate sources of this connector may infringe AMP's patent. The existence of these alternate sources was a principal reason many people voted for this high-density connector. Honda and Fujitsu did not have enough time prior to the meeting to investigate their legal position on this issue. David Smith, a patent attorney for AMP, presented an overview of what is claimed by AMP's patent. He also explained some of the fundamentals in reading a patent. He was asked whether AMP had any other pertinent patent applications and he responded that patent applications are trade secrets and therefore he is not permitted to answer the question. Dal Allan moved and Gene Milligan seconded that X3T9.2 request all X3T9.2 connector vendors respond as to their business intention with respect to the ANSI patent policy on any patent or patent application that pertains to any of the SCSI-2 connectors at the June 1989 plenary meeting. The motion passed with 32 in favor and 2 opposed. 10.7 Should REQUEST SENSE be required to use the same queue tag? [Houlder] The intention of the committee is that REQUEST SENSE commands do not have a queue tag. Furthermore an untagged command should be treated as head of queue. Queuing still needs some work in the area of error recovery. Snively, Nitza, McGrath, Eneboe were assigned an action item to analyze and propose a solution for the next working group meeting. 10.8 LOGICAL UNIT RESET Message (X3T9.2/89-47) [Lohmeyer] This proposal from NCR would add permit an initiator to take over for a failed initiator. The failed initiator may have outstanding reservations and queued commands. The good initiator would issue a LOGICAL UNIT RESET message to each logical unit of each target that it wishes take over from the failed initiator. The LOGICAL UNIT RESET message would cause the target to release any device reservations and abort any queued commands for the failed initiator (specified in the message). John Lohmeyer said that he had not attached this proposal to his X3T9 letter ballot because it could be viewed as a new proposal and hence a subject for SCSI-3. However, he felt that this capability is an important feature for multi-initiator systems and it should be standardized at the earliest possible time. While NCR may add this feature in a vendor specific manner, it is likely that other vendors will also need it. Several people voiced concerns over whether the proposal was adequate for other applications. In particular, there was a strong concern that aborting the outstanding I/O processes would make recovery difficult or impossible in some applications. Most people felt that the issue required more study and that it would delay the SCSI-2 approval. Therefore the committee remanded this proposal to the SCSI-3 project. 10.9 Command Queuing, ECA, and Error Conditions [Nitza] There was a straw poll of 15 to 4 in favor of retaining the statement that requires a target to return BUSY status for any tagged I/O processes that do not have the DiscPriv bit set to one. The discussion of this agenda item exposed several problems in handling error conditions in tagged queuing. The issue was assigned to the individuals mentioned in 10.7 and to the May working group for resolution. 11. SCSI-3 Activities There was only minor work on SCSI-3 since the last plenary meeting. Please refer to agenda item 8.2 and to X3T9.2/89-42 for additional information. 11.1 Autoconfiguration SSWG There was no activity on this subject. 12. Review of Action Items 1. Snively, Nitza, McGrath, Eneboe will analyze and propose a solution for error handling of tagged commands for the May working group meeting. 2. Dan Davies will propose a replacement for the QIC-600 density code to support 1.3 GByte tapes. 3. Larry Lamers will reword the paragraphs describing ASC and ASCQ hierarchy on page 7-46. 4. Larry Lamers will prepare SCSI-2 R9 for the next X3T9.2 mailing. 5. The X3T9.2 connector vendors will prepare statements concerning their business intentions on SCSI-2 connector patents and patent applications. 13. Meeting Schedule The next meeting of X3T9.2 will be June 19-20, 1989 at the Red Lion Inn (408-279-0600) in San Jose, California hosted by AMD. Please mention X3T9 - AMD when making reservations to receive the special room rate of $99.00 single or double occupancy. The cut-off date for reservations is May 22, 1989. The plenary meeting schedule for the remainder of 1989 and 1990 is: Date Location Host -------------------- ------------------------ --------------------- August 21-22, 1989 Denver DEC October 16-17, 1989 Research Triangle Park, NC IBM December 4-5, 1989 San Diego, CA NCR February 19-20, 1990 Austin, TX Motorola April 23-24, 1990 St. Petersburg Beach, FL AMP June 18-19, 1990 Wichita, KS NCR August 20-21, 1990 Seattle, WA Boeing October 15-16, 1990 Valley Forge, PA UNISYS December 3-4, 1990 Southern CA ?? (Chairman's note: The 1990 dates and locations were not discussed during the meeting, but were established by X3T9 later in the week.) 13.1 General Working Group Schedule for 1989 John Lohmeyer announced that he would host the May working group meeting in Wichita, Kansas at the Hilton Inn East (316-686-7131). The cut-off date for room reservations was Friday April 28. Please mention NCR/SCSI to get the $60 room rate (corporate rate is $64). The SCSI-2/SCSI-3 Working Group meeting schedule for 1989 is: Date Location Host Note -------------------- ----------------- ------------- ------------------- May 8-9, 1989 Wichita, KS John Lohmeyer {see 89-041} July 10-12, 1989 Chicago, IL Joe Lawlor {see 89-059} September 6-8, 1989 Oklahoma City, OK Gene Milligan {tentative} Oct 30 - Nov 1, 1989 ? ? {tentative date} The May meeting has been shortened to two days. A SCSI CAM meeting will be held on May 10 in the same hotel. 14. Adjournment The meeting was adjourned at 5:00 p.m. on Tuesday, April 25, 1989.