Re: [Scst-devel] [REQUEST TO TEST]: ALUA changes toward better SCSI compliance

[Vladislav B. <vs...@vl...>]

Bart Van Assche wrote on 02/26/2016 04:22 PM:
> On 02/26/2016 03:45 PM, Sebastian Herbszt wrote:
>> I just enabled explicit ALUA support in SCST and noticed STPG gets send
>> down the standby path. I think it comes from the ALUA handler and not multipathd.
>>
>> multipathd: 26162356632353435: sdc - tur checker reports path is in standby state
>> multipathd: 8:32: reinstated
> 
> Hello Sebastian,
> 
> This multipathd behavior has been introduced in 2009 (see also below). 
> I'm afraid that anyone who is using an enterprise OS release around or 
> before that date will see the previous multipathd behavior, namely that 
> standby paths for which SCST fails TUR are considered failed by the 
> initiator and hence that the initiator won't send an STPG over these paths.
> 
> Bart.
> 
>  From the multipath-tools changelog:
> 
> commit 5da642ff71254ecda685e82c681e78cab4c75b87
> Author: Hannes Reinecke <ha...@su...>
> Date:   Thu Feb 19 16:19:45 2009 +0100
> 
>      Return 'ghost' state when port is in standby
> 
>      Currently, issuing a TUR command on a path in standby
>      mode, it returns a Unit Attention status. This is
>      reported as a Failed path by the SLES11 tur checker.
>      Instead, this has to be reported as Ghost path based
>      on the additional sense code and sense code qualifier values.
> 
>      References: 475816
> 
>      Signed-off-by: Hannes Reinecke <ha...@su...>

OK, good, first time something real. However, I'm still can not say that I'm fully
convinced, because for formal full proof we need to see that without this patch
multipath from that time fails in some bad, unrecoverable way.

Also, enterprise OSes have 5 years release cycle, which means that they all must have
already upgraded to the fixed multipath-tools. For those users who still stays with 7
years old systems, it would be a good incentive to upgrade. Any objections?

Actually, it might be a good idea to cleanup SCST sources from support of kernels below
2.6.33 (first in 2010)

Thanks,
Vlad