[Pacemaker] Replacing sbd devices in running cluster

Gregory House raintown.us at gmail.com
Thu Jan 29 13:00:07 EST 2015

Hi Pacemaker Experts

I have inherited a pacemaker cluster running on the SLES11SP1 stack.
Sadly there is no support contract any more.

# rpm -qa | egrep 'pacemaker'

I need to migrate both SBD devices.  I.e. /dev/sdX and /dev/sdY. to
/dev/sdA and /dev/sdB

It is required to do this without service downtime.

The sbd resource is running OK:

# crm_mon -1 | grep sbd
 stonith_sbd_is (stonith:external/sbd): Started node1

and is configured as follows

# crm configure show | grep sbd
primitive stonith_sbd_is stonith:external/sbd \
        params sbd_device="/dev/sdX;/dev/sdY"

and sbd processes are running from this config

# cat /etc/sysconfig/sbd

pstree -halp | grep sbd
  |   |-sbd,8798
  |   `-sbd,8799

(similarly on node2)

The resources are all running, and both nodes are OK

# crm_mon
Last updated: Thu Jan 29 18:52:53 2015
Stack: openais
Current DC: s987l1020 - partition with quorum
Version: 1.1.5-5bd2b9154d7d9f86d7f56fe0a74072a5a6590c60
2 Nodes configured, 2 expected votes
61 Resources configured.

Online: [ node1 node2]

stonith_sbd_is  (stonith:external/sbd): Started node1

Is this change possible without stopping the cluster? If so, how
should I best implement?

With downtime I guess all I would need to do is change the
/etc/sysconfig/sbd file, change the settings within the cluster via
crm, and restart everything. But I cannot see the cluster surviving
this without a downtime?


More information about the Pacemaker mailing list