[Pacemaker] SBD kills both nodes in a two node cluster.

Ulf mopp at gmx.net
Wed May 18 09:15:51 EDT 2011


> Don't clone it.
Removed the clone

>And it'd probably make sense to introduce a sleep() in the start action
>there, that would give the partition that has to instantiate sbd first a
>slight malus ...

I added a start-delay of 15, which seems to give the node with the running sbd resource an advantage over the other node, so the passive node gets killed if the primary is still active.

primitive stonith_sbd stonith:external/sbd \
        params sbd_device="/dev/disk/by-id/scsi-360012345678900000000123400005678-part1" \
        operations $id="stonith_sbd-operations" \
        op start interval="0" timeout="15" start-delay="15"


Thank you Lars.

Ulf
-- 
NEU: FreePhone - kostenlos mobil telefonieren und surfen!			
Jetzt informieren: http://www.gmx.net/de/go/freephone




More information about the Pacemaker mailing list