[dm-devel] Question regarding failback option in multipath.conf

[Eli Klein]

On Fri, Oct 2, 2009 at 8:30 AM, John Brier <jbrier at redhat.com> wrote:

> On 10/01/2009 09:01 PM, Eli Klein wrote:
>
>> Sorry, I forgot to mention that I did.. it didn't help.
>> In addition, from the output of multipath -v3, multipathd is picking up
>> all
>> of the correct settings, even with the two separate default sections.
>>
>>
> It seems like you set it all up right but failback isn't working like it
> *seems* it should. (I have never needed it) If you're interested in a hack..
>
> Since you don't seem to care how long it takes to failover, you could do
> the *opposite* of what this kbase recommends:
>
> How can I improve the failover time of a faulty path when using
> device-mapper-multipath over iSCSI?
>
> http://kbase.redhat.com/faq/docs/DOC-2877
>
>
> Aside from that I would suggest you open a support case with Red Hat if the
> box is covered under a support contract.
>
>
> John Brier
>
> --
> dm-devel mailing list
> dm-devel at redhat.com
> https://www.redhat.com/mailman/listinfo/dm-devel
>

I've seen that doc before (it was helpful!).  Unfortunately, I *do* care
about how long it takes for the initial failover.  I have this configured to
take 15 seconds at the moment.  One thing that's interesting is that even
when I had "node.session.timeo.replacement_timeout" set to 120 seconds, the
failback was still immediate. The initial failover would take at minimum 120
seconds.

All of my problems seem to be cause by what appears to be a failback
parameter that doesn't actually have any effect.  I'll have to open up a
support case to get to the bottom of the failback issues.

Thank you very much for your help!

-Eli
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/dm-devel/attachments/20091002/13afad8a/attachment.htm>