On Tuesday 27 May 2008 18:21:42 Troels Arvin wrote:
> Nuno Fernandes wrote:
> >> Can you reveal (perhaps only parts of) your multipath.conf?
> >
> > Inline.
> >
> > defaults {
> > user_friendly_names yes
> > path_grouping_policy group_by_prio
> > no_path_retry queue
> > }
> > devices {
> > device {
> > vendor "IBM"
> > product "1814*"
> > hardware_handler "1 rdac"
> > path_checker rdac
> > failback immediate
> > prio_callout "/sbin/mpath_prio_tpc /dev/%n"
> > }
> > }
> >
> > multipaths {
> > multipath {
> > wwid XXXXXXXXXXXXXXXXXXXXXXXXXXX
> > alias volumeX
> > }
> > }
>
> Why is mpath_prio_tpc being used, and not mpath_prio_rdac?
> - The mpath_prio_rdac would seem more logical, given that several
> other configuration items point to *_rdac versions.
Well... let me explain.. we are using rhel5.0.
In rhel5.0 there isn't support for rdac multipath. It only appeared in 5.1.
Back then, 5.1 was still in testing. We've backported the patch (hwhandler)
for the 5.0 kernel as well as the multipath-tools.
The first configuration had mpath_prio_rdac but the system needed to be in
production. So.. we had tested before mpath_prio_tpc and it worked as
expected. We didn't have the time to properly test mpath_prio_rdac with all
the scenarios... so we've configured as described. When we have time and a
schedule window of downtime (probably when updating to 5.2) we'll try it...
Nevertheless i think the problem that you are having is failback related..
Best regards,
Nunno Fernandes
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
