[Linux-HA] Filesystem resource agent will not mount device, "ERROR: unpack_rsc_op: Hard error"
martin.oswald at muenchen.de
Tue Oct 26 09:44:20 MDT 2010
I am running a two node Heartbeat Cluster Version 2.1.3 with SLES10 SP2.
My config includes several resources, which are being mounted via the
OCF Filesystem resource script. All those mounts work fine, so I thought
it would be no problem to add another mount to my config...but I was wrong.
My cluster is running, so I added another Filesystem-Resource to a
existing resource group. The device and directory are spelled correctly,
the mountpoint exists. Although the device does not get mounted by
The only error I get in my logfile (I have got debug 0 set in ha.cf) is
-- snip --
Oct 26 17:15:26 db10 mgmtd: : ERROR: unpack_rsc_op: Hard error:
mount_ora37_data2_monitor_0 failed with rc=2.
Oct 26 17:15:26 db10 mgmtd: : ERROR: unpack_rsc_op: Preventing
mount_ora37_data2 from re-starting anywhere in the cluster
-- snip --
I must admit, I am a little bit confused because I have no "monitor"
operation configured. Where does the "mount_ora37_data2_monitor_0" come
I tried to mount the device manual via "mount" and the agent directly
/data2/ora37/ vxfs start) and both options work fine.
It is a Veritas Filesystem but so are the other mounts which work
Can anyone please help or has some ideas to debug? Unfortunately it is a
productive system, so I think I can not increase the debug level without
restarting the cluster?
More information about the Linux-HA