On jeu, 2006-02-09 at 12:39 -0700, Dale.R.Fowle@xxxxxxxxxxx wrote: > Hi all, I have a question related to failover/failback. I'm running > 0.4.5-0.11. I'm wondering what the path checker does for the multipathd > daemon. > > Here's my senario: > I have two paths to a LUN, configure it with 'multipath -v 2 -p failover'. > Run some I/O down the path of the /dev/mapper/<device>. > I fail one of the paths either by pulling the fibre or disabling the port > it's connected to. > When I run 'multipath -ll' I see the failed path 'failed faulty'. > I restore the path physically. > Run 'multipath -ll' again to see the status and it's 'failed ready'. > > Now, my question is, is multipathd supposed to handle making that path > 'active ready'? Yes, multipathd should reinstate the path upon checker status change from "faulty" to "ready" > I have 'failback immediate' set in the /etc/multipath.conf file. And it > does do the immediate failback,...but based on the 'active ready' status. > To get to that status, I have to actually issue a 'multipath' command to > reinstate the path to active. > Likely a bug you'll have to chase down a little. Starting to challenge the devel branch would be a good idea, too. > Is there something that I'm missing or am I just going to have to make a > cron job run every few minutes to do path status updates? > You should not need to resort to that. Regards, cvaroqui -- dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
