[LSF/MM TOPIC] I_T Nexus loss SCSI error handling

[LSF/MM TOPIC] I_T Nexus loss SCSI error handling

[Hannes Reinecke]

Hi all,

here's a topic I'd like to discuss at LSF/MM:

SCSI error handling update to handle I_T Nexus loss.

The current error handler still uses a 'target reset' (or, rather, 
bus reset) strategy, although the respective TMF  has been obsoleted 
since SAM-3. SAM-5 defines an I_T nexus loss event instead, which so 
far has only been implemented in libsas.

There has been some discussions on the mailing list, but so far 
there hasn't been any conclusion.

So I would like to discuss a possible implementation of a I_T Nexus 
loss strategy and how to keep compability with SAM-2 targets.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [LSF/MM TOPIC] I_T Nexus loss SCSI error handling

[Roland Dreier]

On Mon, Jan 21, 2013 at 3:18 AM, Hannes Reinecke <hare@suse.de> wrote:
> The current error handler still uses a 'target reset' (or, rather, bus
> reset) strategy, although the respective TMF  has been obsoleted since
> SAM-3. SAM-5 defines an I_T nexus loss event instead, which so far has only
> been implemented in libsas.
>
> There has been some discussions on the mailing list, but so far there hasn't
> been any conclusion.
>
> So I would like to discuss a possible implementation of a I_T Nexus loss
> strategy and how to keep compability with SAM-2 targets.

Would this handle the case of a SAS fabric with lots of target
devices, where we start to get errors from a single device while IO is
proceeding fine to everything else on the fabric?  It would be really
nice if the Linux stack handled this without escalating to a target
reset and host reset, since that disrupts all the IO that is
proceeding fine (and which won't do much to fix a target drive that
might really be dead).

 - R.