Hello, I've on several occasions seen that multipathd fails to reinstate paths that has recovered. Yesterday I had a controller meltdown on an array, and several hosts ended up stuck in a all-paths-down state, which a simple restart of multipathd fixed. It's as if multipathd at some point stopped to check dead paths for revival. The hosts were running 0.4.7 and 0.4.5 (RHEL4/5 and Debian 4.0). Since getting things back on line was the priority I didn't get around to debugging it, unfortunately. I've seen this before many times, though, so I'm wondering if anybody else has seen this and know what causes it? Is there any known workarounds or ways to avoid it? Any suggestions on how to debug it if I see it happen again? Regards, -- Tore Anderson -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
