[stgt] SCSI Persistent Group Reservations: possible bugs found

[Alexander Nezhinsky]

Hi,

Great to hear about the test suite. This reminds me of something...

I've been looking into SPC support and specifically Persistent Group
Reservations recently and saw that stgt does not support "preempt_and_abort"
reserve_out action.

What are the practical implications? Which packages should break?
I guess vmware can't care less. Does Microsoft cluster require it - i know
they do use PGR, but do they actually need preempt and abort?
RedHat Clustering? Solaris?

Lee, can your test suite group the tests according to the minimal needs
of various clustering clients' requirements?

Tomo, can you give a hint on why preempt_and_abort was commented out.
What stops us from reenabling it and how difficult it is?

Alexander

On 12/15/12, Lee Duncan <lduncan at suse.com> wrote:
> Hi:
>
> I work on SCSI-3 Persistent Group Reservations at times, and I noticed
> there we not test suite for this functionality, so I have created one.
> As part of my work in getting this suite to work, I have been applying
> it to the stgt iSCSI target. My target happens to be running on openSUSE
> 12.2, using the latest target code from your master branch.
>
> I believe I have found a couple of bugs in the PGR functionality, and I
> wanted to know the best way to report them. In addition, I would like to
> submit patches for any bugs I find, and I wanted to make sure you were
> receptive to such patches.
>
> In addition, I'd be glad to share my test suite with you. Right now, it
> uses an open-iscsi initiator (and manual set up), and the Python
> "nosetests" module and the sg_persist SCSI utility command. The tests
> have been developed on SLES 11 SP2. There are almost a hundred test
> cases so far, but they are still in development. For example, I am still
> not testing for Unit Attention, Preempt, and other pieces of PGR
> functionality, but the suite does the 6 types of reservations, and it
> makes sure access is restricted (or not), as appropriate.
>
> So far, I believe I found an error in some case where the Reservation
> Holder unregisters. In this case, the reservation should go away.
> --
> Lee Duncan
> SUSE Labs
> --
> To unsubscribe from this list: send the line "unsubscribe stgt" in
> the body of a message to majordomo at vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
--
To unsubscribe from this list: send the line "unsubscribe stgt" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html