6.4. Triggering the SAPHanaFilesystem failure action

Procedure

1. Go to the hidden directory which the SAPHanaFilesystem resource uses to test read and write filesystem access:

   [root]# cd /hana/shared/<SID>/.heartbeat_SAPHanaFilesystem/<node>

2. Set the existing test file to become immutable, which prevents write access by the resource monitors:

   [root]# chattr +i test

3. Verify the behavior during the simulated failure.

   1. You can check the /var/log/messages file for the related log message if the resource action is set to ignore:

      [root]# grep ON_FAIL_ACTION /var/log/messages
      ... node1 SAPHanaFilesystem(rsc_SAPHanaFil_RH1_HDB02)[715184]: INFO: -2- RA monitor() ON_FAIL_ACTION=ignore => ignore FS error, do not create poison pill file

   2. You can verify the fence behavior if the resource action is set to fence:

      [root]# pcs status --full
      ...
      
      Failed Resource Actions:
        * rsc_SAPHanaFil_RH1_HDB02_stop_0 on node1 'error' (1): ...
      
      Pending Fencing Actions:
        * reboot of node1 pending: client=pacemaker-controld.1694, origin=node2

4. You must remove this blocker again after the test. If the node is fenced, do this after the node is running again:

   [root]# chattr -i test