Minutes for RAID working group 9/16/93 Accredited Standards Committee X3, Information Processing Systems Doc. No.: X3T9.2/93-141 R1 Date: Sept 16, 1993 Project: Ref. Doc.: Reply to: G. Penokie To: Membership of X3T9.2 From: George Penokie and Ralph Weber Subject: Minutes of RAID Study Group Meeting (9/16/93) Agenda 1. Opening Remarks 2. Attendance and Membership 3. Approval of Agenda 4. RAB Host Interface Group Activities 5. SCSI Disk Array Model (93-003r9) 6. Action Items 7. Meeting Schedule 8. Adjournment Results of Meeting 1. Opening Remarks George Penokie the RAID Study Group Chair, called the meeting to order at 9:00 am, Thursday September 16, 1993. He thanked Bob Dugan of IBM for hosting the meeting. As is customary, the people attending introduced themselves. A copy of the attendance list was circulated for attendance and corrections. It was stated that the meeting had been authorized by X3T9.2 and would be conducted under the X3 rules. Ad hoc meetings take no final actions, but prepare recommendations for approval by the X3T9.2 task group. The voting rules for the meeting are those of the parent committee, X3T9.2. These rules are: one vote per company; and any participating company member may vote. The minutes of this meeting will be posted to the SCSI BBS and the SCSI Reflector and will be included in the next committee mailing. George stated that this is the tenth meeting of the RAID study group. The purpose of the group is to deal with interface issues related to using RAIDs. The study group will assess the issues and then formulate a strategy for dealing with them. 2. Attendance and Membership Attendance at working group meetings does not count toward minimum attendance requirements for X3T9.2 membership. Working group meetings are open to any person or company to attend and to express their opinion on the subjects being discussed. The following people attended the meeting: RAID Study Group Meeting Attenders Name S Organization Electronic Mail Address ---------------------- - ------------------------- ------------------------- Mr. Norm Harris P Adaptec, Inc. nharris@adaptec.com Mr. David Skinner O Advanced Micro Devices dave.skinner@amd.com Mr. Bob Wheeler O AMD b.wheeler@amd.com Mr. Gerry Johnsen A Ciprico gerry@cipric.mn.org Mr. William Galloway O Compaq Computer Corp Mr. Edward Haske P CMD Technology Mr. Peter Walford O DemoGraFX Mr. Paul Massiglia O Digital Equipment Corp. massiglia@genral.enet.dec .com Mr. Doug Hagerman O Digital Equipment Corp. hagerman@starch.enet.dec. com Mr. Ralph Weber A Digital Equipment Corp. Mr. I. Dal Allen P ENDL 2501752@mcimail.com Mr. Howard Grill O Formation Mr. Bill Hutchison O Hewlett Packard Co. hutch@.boi.hp.com Mr. Jeffrey Williams P Hewlett Packard Co. jlw@hpdmd48.boi.hp.com Mr. Charles Shen O Hitachi Computer Products c_shen@hitachi.com Mr. Yoshihiko Yano S Hitachi Computer Products y_yano@ipd.hitachi.com Mr. George Penokie P IBM Corp. gop@rchvmp3.vnet.ibm.com Mr. Giles Frazier O IBM Corp. gfrazier@asuvm6.vnet.ibm. com Mr. Ray Muggli O IBM Corp. Mr. Chuck Grant A Madison Cable Corp. Mr. Mike Glass V Microsoft mglass@microsoft.com Mr. M. Jibbe S NCR Corp. M.Jibbe@WichitaKS.NCR.COM Mr. Charles Binford S NCR Corp. Charles.Binford@WichitaKS .NCR.COM Mr. Stephen F. Heil P Panasonic Technologies, sfh@research.panasonic.co Inc m Mr. Robert N. Snively P Sun Microsystems, Inc bob.snively@eng.sun.com Mr. Todd Sprenkle S Tandem Computers sprenkle_todd@tandem.com Mr. Greg Kite O The GCA Group Mr. Peter Dougherty P UNISYS Mr. Paul D. Aloisi P Unitrode Integrated Aloisi@uicc.com Circuits Mr. Doug Piper P Woven Electronics WOVEN ELECTRONICS / MCI ID: 549-9900 30 people present Status Key: P - Principal A - Alternate O - Observer L - Liaison S,V - Visitor 3. Approval of Agenda The agenda developed at the meeting was approved. 4. Report on last RAID Working Group Last RAID Working Group was held in Colorado Springs. Several members of the RAID Advisory Board were present. Minutes of the Colorado Springs meeting were published on the SCSI Reflector. 5. RAB Host Interface Group Activities A review of the changes in Rev 9 of the SCSI Disk Array Model document (93-003r9) was made. Many changes where requested and will be incorporated into the next rev of SCSI Disk Array Model. 6. SCSI Disk Array Model (93-003r9) [Penokie] George Penokie presented revision 9 of the SCSI Disk Array Model. The R9 document includes changes from the August RAID SSWG meeting and changes from the joint meeting with the RAID Advisory Board. The following was discussed during the presentation of revision 9: In the glossary, the LBA and LUI (Logical Block Address and Logical Unit Identifier) have usage identifiers prefix, e.g. P-LUI is Physical Logical Unit Identifier. A more complete definition of extent has been added. The notion of "pool" has been used in replacement of "set." Redundancy sets will be allowed to overlap. This permits a RAID set construction that the RAID Advisory Board believes is useful. Questions were raised about representing data compression in the model. George was open to advice about how to represent data compression, however, he felt incapable of defining such a capability without some kind of specific proposal from someone. George described the process of creating a redundancy set from P-extents. This is the first step in creating a usable disk array LBA addressing space. The subsystem has a report-of-free-space function that is used to determine what P-extents are available for creating the redundancy set. Next, R-extents are assigned to volume sets. The volume sets are the data access mechanism for user applications. Note: today most redundancy sets are exactly overlaid by their defined volume sets, but this in not required by the Model. A lengthy debate arose around whether the check data is part of the addressing space in the volume set. Viewed from one perspective, the check data is a "shadow" component of the volume set address space. The check data is "automatically" addressed based on redundancy set status and the asserted volume set address. The opposite perspective, says that the check data is not in the volume set because no address asserted in the volume set address space can specifically address a specific address in the check data space. The latter perspective was the consensus choice of the study group. Additional confusion was found on later Model drawings. The current drawings overlay a redundancy set drawing and a volume set drawing. George concluded that he needs three drawings; the redundancy set drawing, the volume set drawing, and the overlay of the two. Another problem is that redundancy sets and P-extents are model abstractions- , not definitions of data paths. The abstractions are intended for use in the construction and definition of disk array sets, not as a definition of how data is passed from one model layer to the next lower model layer. Similarly, spare pool addressing is an abstraction for constructing and defining spare physical address space within a DACL. George described the need for maintenance functions setting and returning a geographical address for the SCSI Disk Array (SDA) cabinet. George and Bob Snively debated the addressing mechanics for maintenance functions. Does a LUN value need to be reserved for maintenance functions? George is ready to make such a LUN value reservation, but is unwilling to do so until it is really needed. George described the redundancy set configuration parameters. These parameters permit description of every known RAID configuration. Doug Hagerman suggested that the RAID Advisory Board members be polled regarding the completeness of the proposed redundancy set configuration parameters. The usefulness and mechanics of vendor-unique configuration description tools was discussed. 7. Action Items 1) George Penokie will prepare revision 1 of the SCSI Disk Array Questions and Answers which includes larger LUNs and incorporates the features of Bob's proposal. 2) George Penokie will prepare a document to incorporate a new identify message into SPI to handle 8 byte LUN addresses. 3) George Penokie will prepare another reversion the the SCSI Disk Array Model. George noted that, due to the 9.2 document numbering process, the next revision of the SCSI Disk Array Model will be X3T9.2/93-140r0, not X3T9.2/93-003R10. 8. Meeting Schedule The next meeting of the RAID Study Group is planned for Oct 13, 1993 at the Ramada Hotel and Conference Center in West Palm Beach, FL.. The meeting is expected to start at 9:00am-12:00am. This meeting will be a joint RAID Advisory Board Host Interface Group and X3T9.2 RAID Study Group meeting. George stated that on even numbered months the RAID Study Group meetings will be held on Wednesday morning of the X3T9.3 meeting week. In the odd numbered months the RAID Study Group will meet on Thursday morning with X3T9.2. On the even numbered months RAB Host Interface Group meetings will be held all day Tuesday of the X3T9.3 meeting week. If possible these meetings will be held at the same hotel as the X3T9.3 meetings. On the odd numbered months the RAB host Interface Group meetings will be held all day Tuesday of the X3T9.2 meeting week. If possible these meetings will be held at the same hotel as the X3T9.2 meetings. 9. Adjournment The meeting was adjourned at 12:00 pm. on Thursday Sept 16, 1993.