HP Serviceguard Toolkit for NFS on Linux User Guide Version A.12.00.00

6 Troubleshooting
This section explains some of the problem scenarios that you might encounter while working with
the HP Serviceguard Toolkit for NFS on Linux in a Serviceguard cluster.
Recommended actionPossible causeProblem scenario
All Lock migration related attributes
must be configured in the package
Lock migration feature is not enabled
for the NFS package. This does not
The package log contains the
following message:
configuration file before applying the
configuration file.
move the file locks to the adoptive
node whenever the current node
reboots or crashes.
The parameter LOCK_MIGRATION is
not set. NFS client locks will not be
migrated.
Start the process manually using the
command provided by the NFS
application on OS.
Particular process,
<process_name>, which is required
for NFS to function properly has gone
down.
When important daemons go down,
the package fails over to adoptive
node. The package log contains the
error:
<process_name> has died or is not
running. NFS monitoring process
stopped.
NOTE: The daemons include
mountd, rpc.statd,
rpc.rquotad, nfs , lockd,
portmapper, rpcbind,
rpc.idmapd.
Create a file system on a shared vg to
be used as an exported file system as
The specified exported file system is
not on a shared vg.
If the package log contains the
message:
WARNING! <export
directory> is not mounted on a
volume group”
it will be available to the adoptive node
in case of failover.
Restart the package using the
cmhaltpkg and cmrunpkg commands
The error message for NFSv3 package
indicates that stopping synchronization
of rmtab background process has
The package log contains the
following error:
Failed to stop sync rmtab background
process.
so that file system is an exported and
entries are added freshly in to rmtab.
failed, and some other process have
locked this file for write process.
Start the particular rpc process,
<rpc_process_name> manually
The particular rpc,
<rpc_process_name> process
might have gone down.
The package fails over to adoptive
node and the package log file
contains a message:
Could not ping
<rpc_process_name> NFS
monitoring process stopped.
before starting the package again on
the node.
Use the cmmodpkg command to enable
the package to run on the particular
Package might not have been enabled
to run on this cluster node. Check the
package log file and syslog file for
more details.
Starting the package using the
cmrunpkg command fails with the
error:
package script for <pkgname> failed
with no restart. You must not start the
<pkgname>.
node. After correcting the problem
mentioned in the log file, start the
package using cmrunpkg.
Create the rpc_pipefs directory
manually and try to start the
The Directory /var/lib/nfs/
rpc_pipefs directory must be
For NFSv4 packages, rpc.idmapd
daemon fails to start and ps -ef |
rpc.idmapd process using the
present for the rpc.idmapd daemongrep rpc.idmapd does not show
22 Troubleshooting