On Tue, 21 Mar 2006 egoggin@xxxxxxx wrote: > Two problems are very noticeable, pgpolicy should be group_by_prio > instead of failover and the get_prio callout should be set to > /sbin/mpath_prio_emc. Tried that. It didn't help with my current situation. My multipath.conf looks like this now: defaults { udev_dir /dev path_grouping_policy group_by_prio prio_callout "/sbin/mpath_prio_emc /dev/%n" path_checker emc_clariion failback immediate } # blacklist { devnode "^sda(\b|[0-9]+)" } devices { device { vendor "DGC " product "RAID 5 " features "0" hardware_handler "1 emc" } } -- Jussi Hämäläinen +358 3 3551 8086 University of Tampere System administrator +358 50 5900641 Computer Centre ccjuha@xxxxxx TietoPinni 4166 http://www.uta.fi/laitokset/tkk -- dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel