[T10] Revisiting the RMB (Removable Medium) bit definition

Ballard, Curtis C (HPE Storage) curtis.ballard at hpe.com
Wed Jul 29 13:43:18 PDT 2020


As do I.  Ralphs version makes better use of a base SCSI architecture concept that should be easily grasped.  Removing the I_T nexus is the way the device server would be removed.

Curtis Ballard
Hewlett Packard Enterprise
HPE Storage R&D
Fort Collins, CO
(970) 898-3013

From: t10-bounces at t10.org [mailto:t10-bounces at t10.org] On Behalf Of Curtis Stevens
Sent: Wednesday, July 29, 2020 2:35 PM
To: Ralph Weber <Ralph.Weber at wdc.com>; t10 at t10.org
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition

I like that one better.


---------------------------------------------
Curtis E. Stevens
Technologist
Seagate Technology

E-Mail: Curtis.Stevens at Seagate.com<mailto:Curtis.Stevens at Seagate.com>
Phone: 949-307-5050

________________________________
From: t10-bounces at t10.org<mailto:t10-bounces at t10.org> <t10-bounces at t10.org<mailto:t10-bounces at t10.org>> on behalf of Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>
Sent: Wednesday, July 29, 2020 1:19 PM
To: t10 at t10.org<mailto:t10 at t10.org> <t10 at t10.org<mailto:t10 at t10.org>>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition


Or... how about something along the lines of:

A removable medium (rmb) bit set to zero indicates that the medium is not removable without removing the I_T nexus that connects the application client to the device server.  A rmb bit set to one indicates that the medium is removable able to be removed without affecting the I_T nexus between the application client and the device server.





From: Curtis Stevens <curtis.stevens at seagate.com<mailto:curtis.stevens at seagate.com>>
Sent: Wednesday, July 29, 2020 3:12 PM
To: Ballard, Curtis C (HPE Storage) <curtis.ballard at hpe.com<mailto:curtis.ballard at hpe.com>>; Kevin D Butt <kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>>; Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>
Cc: t10 at t10.org<mailto:t10 at t10.org>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition



CAUTION: This email originated from outside of Western Digital. Do not click on links or open attachments unless you recognize the sender and know that the content is safe.



I think part of the blatant wording that Ralph is proposing is that those making the requirements on the device side do not fully understand SCSI.  This has been the case from day 1 and has not changed.  It is fairly easy to understand commands and what they do.  Understanding the architecture is a different story.



I think your suggested change could be integrated with Ralphs change...





---------------------------------------------

Curtis E. Stevens

Technologist

Seagate Technology



E-Mail: Curtis.Stevens at Seagate.com<mailto:Curtis.Stevens at Seagate.com>

Phone: 949-307-5050



________________________________

From: Ballard, Curtis C (HPE Storage) <curtis.ballard at hpe.com<mailto:curtis.ballard at hpe.com>>
Sent: Wednesday, July 29, 2020 12:49 PM
To: Curtis Stevens <curtis.stevens at seagate.com<mailto:curtis.stevens at seagate.com>>; Kevin D Butt <kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>>; Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>
Cc: t10 at t10.org<mailto:t10 at t10.org> <t10 at t10.org<mailto:t10 at t10.org>>
Subject: RE: [T10] Revisiting the RMB (Removable Medium) bit definition



Why don't we specify it in architectural terms.  Something along the lines of:

A removable medium (rmb) bit set to zero indicates that the medium is not removable from the device server.

A rmb bit set to one indicates that the medium is removable able to be removed without removing the device server.



Curtis Ballard

Hewlett Packard Enterprise

HPE Storage R&D

Fort Collins, CO

(970) 898-3013



From: t10-bounces at t10.org<mailto:t10-bounces at t10.org> [mailto:t10-bounces at t10.org] On Behalf Of Curtis Stevens
Sent: Wednesday, July 29, 2020 1:04 PM
To: Kevin D Butt <kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>>; Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>
Cc: t10 at t10.org<mailto:t10 at t10.org>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition



I would turn the proposal around and say the RMB bit shall only be set to one if the device is able to report ... after is media is ejected.





---------------------------------------------

Curtis E. Stevens

Technologist

Seagate Technology



E-Mail: Curtis.Stevens at Seagate.com<mailto:Curtis.Stevens at Seagate.com>

Phone: 949-307-5050



________________________________

From: t10-bounces at t10.org<mailto:t10-bounces at t10.org> <t10-bounces at t10.org<mailto:t10-bounces at t10.org>> on behalf of Kevin D Butt <kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>>
Sent: Wednesday, July 29, 2020 8:48 AM
To: Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>
Cc: t10 at t10.org<mailto:t10 at t10.org> <t10 at t10.org<mailto:t10 at t10.org>>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition



Ralph,

I applaud your goal, however, I am concerned about this definition. Take, for example, a tape library. It has removable media, yet it would not report one of these two additional sense codes. Even if all user cartridges were removed, the library would still be ready. It could do management type things including using diagnostic cartridges (not seen as medium to the application client) to test drives.

The RMBshould be set to zero, if the device server is not able to terminate a TEST UNIT READY command (6.48) with CHECK CONDITION status, with the sense key set to NOT READY and the additional sense code:
a)set to NOT READY, MEDIUM NOT PRESENT; or
b)having the ADDITIONALSENSECODEfield (see 4.4.2 and 4.4.3) set to 3Ah (e.g., MEDIUM NOT PRESENT - TRAY OPEN).

Thanks,

Kevin D. Butt
SCSI Architect, Tape Firmware, Data Retention Infrastructure
T10 Standards
MS 6TYA, 9000 S. Rita Rd., Tucson, AZ 85744
Tel: 520-799-5280
Fax: 520-799-2723 (T/L:321)
Email address: kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>

=========== Interesting Links ===========
[ IBM Tape Storage ]  https://www.ibm.com/it-infrastructure/storage/tape<https://www.ibm.com/it-infrastructure/storage/tape>
[ SSIC - HBA/OS/Switch/Product interoperation ] https://www-304.ibm.com/systems/support/storage/ssic/interoperability.wss<https://www-304.ibm.com/systems/support/storage/ssic/interoperability.wss>
[ LTO & 3592 ISV Support Matrix ] www.ibm.com/systems/resources/lto_isv_matrix.pdf<http://www.ibm.com/systems/resources/lto_isv_matrix.pdf>
[ LTO SCSI Reference ] http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003556<http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003556>
[ 3592 SCSI Reference ] http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003248<http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003248>
===================================



From:        Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>
To:        "t10 at t10.org<mailto:t10 at t10.org>" <t10 at t10.org<mailto:t10 at t10.org>>
Date:        07/29/2020 06:49
Subject:        [EXTERNAL] Re: [T10] Revisiting the RMB (Removable Medium) bit definition
Sent by:        t10-bounces at t10.org<mailto:t10-bounces at t10.org>

________________________________



Proposal 20-082r0<http://www.t10.org/cgi-bin/ac.pl?t=d&f=20-082r0.pdf>SPC-6: Removable Medium Bit Expectations has been posted to bring this discussion thread to the attention of CAP.



Resisting the ongoing calls to put as many words as possible into every standard, the proposal includes only one of the methods for defining a "real" removable medium device discussed in this thread. For the purposes discussed here, only one method is necessary to sufficiently define the standard.



All the best, .Ralph



From:t10-bounces at t10.org <t10-bounces at t10.org<mailto:t10-bounces at t10.org>> On Behalf Of Michael Webster
Sent: Tuesday, July 21, 2020 7:57 PM
To: t10 at t10.org<mailto:t10 at t10.org>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition



I would like to add one other tidbit:  A device reporting RMB=1 in addition to being capable of a state that reports MEDIUM NOT PRESENT, such a device should also properly support and execute an ejection operation of that media when sent a START/STOP UNIT command with Power_Condition=0, LoEj=1, & Start=0.



Mike Webster

Western Digital



From: <t10-bounces at t10.org<mailto:t10-bounces at t10.org>> on behalf of Mike Webster <Mike.Webster at wdc.com<mailto:Mike.Webster at wdc.com>>
Date: Friday, July 17, 2020 at 10:30 AM
To: "t10 at t10.org<mailto:t10 at t10.org>" <t10 at t10.org<mailto:t10 at t10.org>>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition



I would expect a USB memory stick to report RMB=0 and ejection causes the whole device to be removed from the operating system's driver stack.  Afterward, such an ejected device would require an unplug/re-plug for the device to respond to anything on USB.



Unless that USB memory stick included a slot for media (e.g. an SD slot) then I would expect the USB memory stick to report RMB=1 and ejection would only cause the device to remove the media, the device would remain in the operating system's driver stack, and the device would begin reporting CHECK CONDITION, NOT READY, MEDIUM NOT PRESENT for TEST UNIT READY.



Mike Webster

Western Digital



From: <t10-bounces at t10.org<mailto:t10-bounces at t10.org>> on behalf of Curtis Stevens <curtis.stevens at seagate.com<mailto:curtis.stevens at seagate.com>>
Date: Friday, July 17, 2020 at 10:05 AM
To: Kevin D Butt <kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>>, Paul Suhler <Paul.Suhler at kioxia.com<mailto:Paul.Suhler at kioxia.com>>
Cc: "t10 at t10.org<mailto:t10 at t10.org>" <t10 at t10.org<mailto:t10 at t10.org>>, "t10-bounces at t10.org<mailto:t10-bounces at t10.org>" <t10-bounces at t10.org<mailto:t10-bounces at t10.org>>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition



We had this descussion during the original specification development.  For USB, the device controller is removed along with the media.  There is nothing present to respond with an RMB bit.  If however, you plug in a USB floppy, you can have an RMB bit that enables eject function reporting.





---------------------------------------------

Curtis E. Stevens

Technologist

Seagate Technology



E-Mail: Curtis.Stevens at Seagate.com<mailto:Curtis.Stevens at Seagate.com>

Phone: 949-307-5050



________________________________

From:t10-bounces at t10.org<mailto:t10-bounces at t10.org><t10-bounces at t10.org<mailto:t10-bounces at t10.org>> on behalf of Kevin D Butt <kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>>
Sent: Friday, July 17, 2020 9:40 AM
To: Paul Suhler <Paul.Suhler at kioxia.com<mailto:Paul.Suhler at kioxia.com>>
Cc: t10 at t10.org<mailto:t10 at t10.org><t10 at t10.org<mailto:t10 at t10.org>>; t10-bounces at t10.org<mailto:t10-bounces at t10.org><t10-bounces at t10.org<mailto:t10-bounces at t10.org>>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition



I concur with Curtis and Paul.

Kevin D. Butt
SCSI Architect, Tape Firmware, Data Retention Infrastructure
T10 Standards
MS 6TYA, 9000 S. Rita Rd., Tucson, AZ 85744
Tel: 520-799-5280
Fax: 520-799-2723 (T/L:321)
Email address: kdbutt at us.ibm.com<mailto:kdbutt at us.ibm.com>

=========== Interesting Links ===========
[ IBM Tape Storage ]  https://www.ibm.com/it-infrastructure/storage/tape<https://www.ibm.com/it-infrastructure/storage/tape>
[ SSIC - HBA/OS/Switch/Product interoperation ] https://www-304.ibm.com/systems/support/storage/ssic/interoperability.wss<https://www-304.ibm.com/systems/support/storage/ssic/interoperability.wss>
[ LTO & 3592 ISV Support Matrix ] www.ibm.com/systems/resources/lto_isv_matrix.pdf<http://www.ibm.com/systems/resources/lto_isv_matrix.pdf>
[ LTO SCSI Reference ] http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003556<http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003556>
[ 3592 SCSI Reference ] http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003248<http://www-01.ibm.com/support/docview.wss?uid=ssg1S7003248>
===================================



From:        Paul Suhler <Paul.Suhler at kioxia.com<mailto:Paul.Suhler at kioxia.com>>
To:        "Ballard, Curtis C (HPE Storage)" <curtis.ballard at hpe.com<mailto:curtis.ballard at hpe.com>>, Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>, "t10 at t10.org<mailto:t10 at t10.org>" <t10 at t10.org<mailto:t10 at t10.org>>
Date:        07/17/2020 09:19
Subject:        [EXTERNAL] Re: [T10] Revisiting the RMB (Removable Medium) bit definition
Sent by:        t10-bounces at t10.org<mailto:t10-bounces at t10.org>

________________________________



I agree with Curtis about the need to be able to report NOT READY, MEDIUM NOT PRESENT. I've worked on both SSC and SBC devices like that.

Paul
Kioxia

From:t10-bounces at t10.org <t10-bounces at t10.org<mailto:t10-bounces at t10.org>> On Behalf Of Ballard, Curtis C (HPE Storage)
Sent: Friday, July 17, 2020 9:54 AM
To: Ralph Weber <Ralph.Weber at wdc.com<mailto:Ralph.Weber at wdc.com>>; t10 at t10.org<mailto:t10 at t10.org>
Subject: Re: [T10] Revisiting the RMB (Removable Medium) bit definition

Ralph,

That bit is used with tape drives and was used with SCSI magneto optical drives, DVD drives, etc.

Even system I've used with that bit set to one, and I've worked with several devices from different manufacturers that set that bit to one, only use it in reference to the media with no device server disruption other than the expected Unit Attention condition transitions as the media is removed and re-loaded.

SCSI devices should only set that bit to one if they are able to report an 02h/3Ah/00h - NOT READY, MEDIUM NOT PRESENT

Curtis Ballard
Hewlett Packard Enterprise

From:t10-bounces at t10.org<mailto:t10-bounces at t10.org><t10-bounces at t10.org<mailto:t10-bounces at t10.org>> On Behalf Of Ralph Weber
Sent: Thursday, July 16, 2020 5:27 PM
To: t10 at t10.org<mailto:t10 at t10.org>
Subject: [T10] Revisiting the RMB (Removable Medium) bit definition

Regarding the Standard INQUIRY data RMB bit, SPC-6 r02 says...
"A removable medium (RMB) bit set to zero indicates that the medium is not removable. A RMB bit set to one indicates that the medium is removable."

An interesting debate has developed regarding whether a USB "memory stick" is an RMB=0 or an RMB=1 device.

One wag has suggested that the RMB bit definition be conditionalized on the presence/absence of a coincident Logical Unit Reset condition or I_T Nexus Loss condition, something along the lines of...
"A removable medium (RMB) bit set to zero indicates that the medium is not removable without resulting in a Logical Unit Reset condition (see SAM-6) or an I_T Nexus Loss condition (see SAM-6). A RMB bit set to one indicates that the medium is removable with no concurrent Logical Unit Reset condition or I_T Nexus Loss condition."

How says the T10 body politic?

Thanks,

.Ralph_______________________________________________
T10 mailing list
T10 at t10.org<mailto:T10 at t10.org>
https://www.t10.org/mailman/listinfo/t10<https://www.t10.org/mailman/listinfo/t10>

 _______________________________________________
T10 mailing list
T10 at t10.org<mailto:T10 at t10.org>
https://www.t10.org/mailman/listinfo/t10<https://www.t10.org/mailman/listinfo/t10>


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.t10.org/pipermail/t10/attachments/20200729/7b4e4e10/attachment.html>


More information about the T10 mailing list