Minneapolis FC-TAPE minutes

WYATT,STEWART (HP-Boise,ex1) stewart_wyatt at am.exch.hp.com
Mon Jun 14 10:13:54 PDT 1999


* From the T10 Reflector (t10 at symbios.com), posted by:
* "WYATT,STEWART (HP-Boise,ex1)" <stewart_wyatt at am.exch.hp.com>
*
Joint T11/T10 TAPE AdHoc Meeting	T11/99-358v0.pdf
June 9, 1999, 5-6PM, June 10, 1999, 8-10AM
Minneapolis, Minnesota
Stewart Wyatt, Hewlett Packard, FC-TAPE secretary.
stewart_wyatt at hp.com
208-396-3594

1. Introductions - Facilitator Dale Lafollette, StorageTek, called the
meeting to order and had the participants introduce themselves.

2. Approval of this agenda: T11/99-298v0 - Group - The agenda was approved
with a few additions.

3. Approval of 5/4/99 Minutes: T11/99-297v0 - Stewart Wyatt - Approved
without change.

4. Review of Old Action Items: Stewart Wyatt

General Action Items

#1. Erich Oetting - Consider specifying the compression algorithm like the
supported density is specified: Dale Lafollette reported that Eric has made
some progress and will report at next months meeting.
#2. Bob Snively - Post concerns about TapeAlert to reflector: The issues are
being discussed on the reflector, a proposal has been posted.
#3. Stewart Wyatt - Update connector proposal and present to SFF. Reply as
noted to Ed Grivna about including in T11.2 documents: Proposal presented to
SFF as presented in these minutes as item 7.

FC-TAPE action items for Dave Peterson, StorageTek, unless noted: Items 1
through 4 have been completed.

#1. Change clause 2.2 References under development, FC-AL-2 should be
specified as X3.xxx (or NCITS.xxx) instead of X3.272. 
#2. Make the changes to table 25 - SCSI Tape device commands. On page 3,
Mode Select, and page 39, Mode Sense, page code hex '02', and page 39 the
DOC column should read FCP-2. Further on page 39 the bits included in Mode
Sense page code hex '18' and '19' should have be placed in the Mode Select
page.
#3. Clause 3.1 is to be eliminated and the other sections in clause 3
renumbered.
#4. Remove from clause 5.2 paragraph 3 the phrase "and address identifier".
#5. Dale LaFollette - Ask T11.3 plenary for public review of FC-TAPE: This
was postponed until this months meeting. Dale will ask for the review in the
plenary later today.

FCP-2 action items for Bob Snively, SUN Microsystems, unless noted: Bob will
issue a new copy of the FCP-2 before next month's meeting that will enable
item close items numbers 1, 2, 5-11.

#1. Bob Snively will review the ABTS changes in the FCP-2 draft with respect
to FC-PH.
#2. Bob Snively and Bob Kembel create a list of the specific BLS and ELS
which are allowed before Login including explicit Login in an annex for
later placement in FC-FS.
#3 Jim Coomes will make a proposal to limit certain of the mode page bits to
private loop applications. Jim will include a "private operations only" mode
bit in this proposal and modify the description of the existing mode bits to
consider the private only bit: Completed, item 8A in these minutes.
#4 Dave Peterson to write a proposal to carry the protocol for discovery to
be included in a to be determined standards document.Completed, item 8B in

these minutes.
#5. Add a comment that in a class 3 non queuing environment, the exchange
status can be discarded after RR_TOV if another command is not received.
#6. Add the following note to clause 10.5, that if REC_TOV associated error
recovery is allowed, RR_TOV must be 3 times REC_TOV and always appropriate
to ADISC address discovery time.
#7. Check and repair references to ABT-LS for recovery. Place a note near
ABTS-LS recovery abort to clarify the changes made to ABT-LS.
#8. Add a note to clause1.10 FCP_RSP field length warning of the existence
of non-conforming devices issuing 12 byte length FCP_RSP.
#9. In clause 1.8, change the reference to a "technical report" to a
"standard".
#10. In clause1.13, install Neil Wanamaker's proposed fix.
#11. In reference to clause 3.3, Mode page support for recovery, Bob is to
make a proposal about how a device indicates support for tape error recovery
support (SRR and REC) and enables this support. This proposal should be
posted to the reflector as soon as possible for the group to review and
comment on.

SSC Action Items: All of these items have been completed. The letter ballot
(item 7) is currently open.

#1. Dave Peterson will remove the references to extent reservation.
#2. Dave Peterson will add a text discussion be on non-immediate command
reporting using queuing.
#3. Dave Peterson will make the changes proposed by Carl Zeitler on page 44
of SSC rev 15, second paragraph, last sentence, change to "A check condition
caused by early termination shall not result in a negative information field
value". Another clarifying sentence should be added: "The information field
value shall be equal to the magnitude of the count field minus the magnitude
of the blocks, filemarks, or setmarks spaced over." Both sentences need to
be moved to the end of the of paragraph 1 on the same page.
#4. Dave Peterson will reject obsoleting the compression mode in the device
configuration table.
#5. Dave Peterson will add a table to describe the required results of
reading the default compression code to include the four states, current,
changed, saved, default and what compression algorithm is returned in each.
#6. Dave Peterson will update the SSC to revision 17 with the changes agreed
to today. 
#7. Eric Oetting will request a letter ballot comment on revision 17 in the
T10 plenary this week. 

5. FC-TAPE: T11/99-069v4 - Dave Peterson: Dave had no issues to discuss.

6. SSC: T10 Working Drafts - Dave Peterson: At letter ballot. Dave had no
issues to discuss. He hopes to have the letter ballot comments ready for
review at next months meeting.

7. Tape Connector: T11/99-234v1 - Stewart Wyatt

Stewart reported on his presentation to the SFF (Small Form Factor
Committee). The SFF preferred that the 80 pin SCA-2 connector be used
instead of the 68 pin SCA connector. The 80 pin connector is currently used
as the alternate 4 connector in the SPI-3 standard. It has the advantage of
being in production and having improved hot plug attributes. In addition
using the larger connector would provide more reserved pins for future
applications. The SFF spokesman thought that the form factor differences

would preclude mixing the applications.

When the new proposal (T11/99-234v1) was posted, it was criticized by Alvin
Cox of Seagate because of the incompatibility of the two pinouts. Stewart
reworked the proposal to resolve the problem by leveraging the SCSI power
pin locations to the Fibre Channel proposal. The revised proposal has been
posted to the reflector (fc-alt4.pdf). Alvin has reviewed the second
proposal and reported that it resolved his objections.

Stewart asked if there were any objection from the tape group to using the
80 pin connector for the FC-TAPE connector proposal. No objection was
raised. Bob Snively made a motion that the FC-TAPE group endorse this
proposal which was seconded by Stewart. Dale took a vote with 11 in favor
and 0 opposed. (A number of participants did not vote.)

8. FCP-2: T10 Working Drafts, T10/99-168r2 - Bob Snively

Bob quickly reviewed some changes and additions to the FCP-2 that he is
working on to meet the action items he has for the next revision. These went
quickly. The only unresolved issue was the negotiation of FCP recovery
procedures. Bob is concerned about "making link behavior LUN sensitive". An
example of a target having different LUNs with different error recovery
procedures is a single target that includes both disk and tape LUNs. This
happens currently with Fibre Channel to SCSI bridges which treat the devices
behind the bridge as LUNs. Bob thinks that a target should use a single
recovery scheme for all LUNs. Allowing LUN based recovery procedure would
allow the disks and targets to use separate recovery procedures. (This
discussion was resumed later in the meeting. See item 8c)

Bob hopes to have a new revision of the FCP-2 for next meeting next month.

8A. Mode Page 19 - Jim Coomes, Seagate.

Jim had posted his proposal to the web page (99-319v0.pdf) and had sent a
notice with explanation to the reflector. 

Jim had agreed to look at the mode page 19 bits to see if some were only
applicable to private operation. After completing the review Jim concluded
that the DDIS bit was the only one that was meaningful only in a private
loop. This bit was intended to provide a simple migration path to FC-AL from
parallel SCSI. Jim recommended that DDIS bit be ignored when a valid fabric
initialization occurs. 

The second part of Jim's proposal was to add a bit to disable public mode
operation in FC SCSI targets. Jim noted that this request was driven by his
customers and provided the benefits of enhancing zoning and improving
performance by eliminating the overhead of FLA functions. He suggested
calling it Disable Target Fabric Discovery (DTFD). 

There was some concern expressed about switches with "stealth mode". This is
a "vendor unique" capability where a switch makes a private device visible
to the fabric. The resolution was that since this capability is not
documented in the standards, the solution is not an issue for this
committee. 

Dale Lafollette called for a vote. It was approved 19 in favor and 0
opposed.

8b Discovery procedure: Dave Peterson 99-340v0.

Dave Peterson presented a proposal on the evening on the 9th (99-340v0). 


Bob Kembel, Connectivity Solutions, recommended that the public procedure
also be applied to private loops with FNL ports. A long discussion followed.
Dave Baldwin recommended that the GID_FT be used instead of GA_NXT. This was
accepted. (These are defined the FC-GS2). A discussion followed without
resolution about the applicability of INQUIRY and REPORT LUNS. When time ran
out in the evening meeting, the discussion was tabled for the next days
meeting.

Over night Dave Peterson revised the proposal to 99-340v1, incorporating
Dave Baldwin's suggestion. The discussion was continued the following
morning. The intent of this discussion was questioned. The answer was that
the discovery process is how a SCSI device discovers its surroundings. There
was some question about the need for this effort. Bob Kembel felt that the
some of the documentation was questionable. The next issue was where this
documentation belonged. It was suggested that some aspects may belong in
FC-AL. However this group focused on an informative annex for FCP-2. 

There was some discussion about the appropriateness of documenting lower
level topology issues in an FC-4 document. Another concern was to make the
procedure as topology independent as possible. Yet another concern was
relying on the name server since that is not a required function of fabrics.
Suggestions for the procedure included (1) Is there a name server available?
(2) Is there a loop? (3) Scan for private devices. A suggestions was made to
change the titles to Fabric SCSI device (removing the loop reference) and
Loop SCSI device (removing the private reference).

Another set of suggestions was (1) Perform FLOGI if needed (2) Login with
name server (3) Login with the fabric controller (4) Register with the name
server (4) Register for state change notification. At this point the Bob
Snively, the FCP-2 editor, decided that he had all of the advice he needed
and stopped the discussion, promising to create an annex for future review.

8C LUN sensitive link behavior

This issue had been discussed without resolution in agenda item 8. At this
time the discussion was reopened by Bob Snively who wished to have it
resolved. Bob noted that the CRN function was enabled by a mode page and
that FCP_CONF was enabled by PLOGI. These solutions were accepted. However
the means of enabling SRR and REC, which Bob saw as being linked, was
unresolved.

Charles Binford, LSI, disagreed with linking SRR and REC. He thought that
REC is useful by itself for discovering IO status after a LIP. Dal Allen,
ENDL, suggested discovering lack of support of REC by LS_RJT and SRR by PRLI
bit (with the assumption that support for SRR requires support for REC.) Bob
proposed making this solution applicable for the entire target with all
LUNs. The group agreed that for a device with multiple LUNs, those LUNs not
supporting SRR could send a reject. A device that supports SRR may send a
reject if data is unavailable for some reason also. Bob indicated that if
both the target and initiator indicate support in the PRLI then it shall be
supported. Bridges may need to emulate this behavior. The final resolution

is that there will be a PRLI bit for SRR (which assumes support for REC).
Target support including all LUNs is implied by the PRLI bit for these
functions however a LUN not support SRR or a LUN supporting SRR which is
unable to return the data make reject the SRR. REC support is assumed unless
the LUN rejects it.

9. T11 New Business: Group: None.

10: T10 New Business: Group: None.

10A. Future Meeting Requirements: Dale Lafollette collected information to
request time for the next two months.

11. Review New Action Items: Stewart Wyatt

General action items

#1. Erich Oetting - Consider specifying the compression algorithm like the
supported density algorithm is specified.
#2. Bob Snively - Resolve TaleAlert discussions. 
#3. Stewart Wyatt - Complete FC-TAPE connector proposal for SFF review next
month. Reply as noted to Ed Grivna about including in T11.2 documents.

FC-TAPE action items.

#1. Dale LaFollette - Ask T11.3 plenary for public review of FC-TAPE. 

FCP-2 action items for Bob Snively unless noted:

#1. Bob Snively will review the ABTS changes in the FCP-2 draft with respect
to FC-PH. 
#2. Bob Snively and Bob Kembel create a list of the specific BLS and ELS
which are allowed before Login including explicit Login in an annex for
later placement in FC-FS.
#3 Bob Snively will incorporate Jim Coomes proposal on changes to mode page
19 into the next draft of the FCP-2.
#4 Bob Snively will incorporate Dave Peterson' proposal for discovery
protocol into an informative annex in the FCP-2.
#5. Add a comment that in a class 3 non queuing environment, the exchange
status can be discarded after RR_TOV if another command is not received.
#6. Add the following note to clause 10.5, that if REC_TOV associated error
recovery is allowed, RR_TOV must be 3 times REC_TOV and always appropriate
to ADISC address discovery time.
#7. Check and repair references to ABT-LS for recovery. Place a note near
ABTS-LS recovery abort to clarify the changes made to ABT-LS.
#8. Add a note to clause1.10 FCP_RSP field length warning of the existence
of non-conforming devices issuing 12 byte length FCP_RSP.
#9. In clause 1.8, change the reference to a "technical report" to a
"standard".
#10. In clause1.13, install Neil Wanamaker's proposed fix.
#11. In reference to clause 3.3, Mode page support for recovery, Bob is to
make a proposal about how a device indicates support for tape error recovery
support (SRR and REC) and enables this support. This proposal should be
posted to the reflector as soon as possible for the group to review and
comment on.

SSC Action Items:

1. Stewart Wyatt and/or Ian Crighton will make a presentation for inclusion
of Media Auxiliary Memory (T10/99-148r1) in the SSC at next month's meeting.
HP will create a letter ballot comment proposing the changes.

12. Adjournment: Group

Attendance

Dale Lafollette	StorageTek		Dave Peterson	StorageTec
Stewart Wyatt	HP			Bob Snively	SUN
Paul Suhler	Seagate		Bob Kembel	Connectivity Solutions
Carl Zeitler	Compaq		Brad Culp	HP
Mark DeWilde	Pathlight Technology	Joe Breher	Exabyte
Ron Martin	Veritas			Matt Gaffney	StorageTek
Matt Wakeley	HP			Neil Wanamaker Crossroads

Alan Gagliardotto Sony Data Systems	Harry Levinson	Computerm Cord
Damian Bannon	SSL			Horst Truestedt	ENDL
Michael Hoard	Sequent		Dave Baldwin	Emulex
Toshi Terasauic	Sony			Curt Ridgeway 	LSI
Charles Binford	LSI			Arlan Stone	UNISYS
Dal Alan	ENDL			Jim Coomes	Seagate
Danielle Maze	DataCore Software	Jeff Williams	Western Digital
Bill Martin	Gadzoox Networks	Ken Fredericks	McData
Kumar Malavalli	Brocade		Earl E. Rydell	Rockwell-Collins
Schelto van Doorn	Infineon		John Lohmeyer	LSI
Macaulay Osaisai	Harris		Edward A. Gardiner Ophidian Designs
Gene Milligan	Seagate		Gary Warden	SRB Consulting, Inc.

*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at symbios.com





More information about the T10 mailing list