Hi!
I don't know if this is an isse of mdadm in SLES11 SP1, but we had the
situation where a RAID1 ended up with one leg, when a manual "mdadm --re-add
/dev/mdX /dev/disk/..." worked.
Inspecting the RA, I guess it also should have tried that (a bit differently):
ocf_log info "Attempting recovery sequence to re-add devices on
$MDDEV:"
$MDADM $MDDEV --fail detached
$MDADM $MDDEV --remove failed
$MDADM $MDDEV --re-add missing
# TODO: At this stage, there's nothing to actually do
# here. Either this worked or it did not.
How was the problem created? One of the RAID legs had been presented to only
one of three cluster nodes (because nobody expected the software to activate
the RAID elsewhere). Obviously the cluster tried to activate the RAID on any
node, where if failed in 2 of 3 cases. But when the RAID came back to the boog
node, the RAID remained incomplete.
I thought, I'd let you know. I have fixed the presentation in the meantime.
Regards,
Ulrich
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
