How to recreate SBD partition in a SUSE HA cluster.

SUSE Linux Enterprise High Availability Extension 11 Service Pack 4

At times you need to change the default values of the SBD partition or even recreate it on a different device.  This can be done while the cluster and resources are up and running following these steps.

SLES 11:

1. Put the cluster into maintenance mode
# crm configure property maintenance-mode=true

2. Stop the stonith sbd primitive in the cluster. 

3. Check the /etc/sysconfig/sbd configuration file to see which device and options are currently used. 
Example:
SBD_DEVICE="/dev/sdb1"
SBD_OPTS="-W -P"

4. Stop sbd on all nodes and all devices: sbd –d <device> message LOCAL exit
Verify that the sbd processes are no longer running.   These processes will be gone after running the command.
# ps -eaf | grep sbd
root      8695     1  0 13:11 pts/0    00:00:00 sbd: inquisitor                                                   
root      8696  8695  0 13:11 pts/0    00:00:00 sbd: watcher: /dev/sdb1 - slot: 0 - uuid: 47c245cb-2c80-49b5-9d1d-c2ec246c2f0a
root      8699  8695  0 13:11 pts/0    00:00:01 sbd: watcher: Pacemaker                         

5. Create the new sbd partition/s on one of the nodes.  Make sure you use same device that's in /etc/sysconfig/sbd
Example: # sbd -d /dev/sdb1 -1 30 -4 60 create
Note:  If the device is different than before, you will need to update the /etc/sysconfig/sbd on each node with the new device.

6. Restart the sbd processes with the options defined in the /etc/sysconfig/sbd on all nodes in the cluster.
# sbd –d <device> watch -P
Verify that the same processes were restarted as we saw in step 3.

7. Start the stonith primitive in the cluster.

8. Put the cluster back to normal mode.
# crm configure property maintenance-mode=false


Note:

When using the SLES11 method on SLES12, the sbd processes would not stop properly if you needed to stop pacemaker service for any reason.