Veritas Storage Foundation 5.1 SP1 Cluster File System Administrator"s Guide (5900-1738, April 2011)

■ Stops lock and status services on all nodes to prevent granting locks.
■ Copies all the files from /locks/sm/lastonline/sm/ to
/locks/sm/nextonline/sm/ directory.
where locks is the file system created for storing lock information.
where lastonline is the node on which the VIP resource was previous online.
where nextonline is the node on which the VIP resource will go online next.
■ Calls /opt/VRTS/bin/fsclustadm frlock_resume to resume giving out locks.
Note: At the end of the preonline trigger all lock services have stopped on all
nodes and no NFS locks can be requested until they are restarted.
postonline trigger
The postonline script for each VIP does the following during an IP failover or
node failure:
■ Starts lock services, triggers reclaim, and grace mode on all nodes.
■ The restarting of status monitor scans all lock status files in the state directory
and contacts all nodes to reclaim their locks. The state files get deleted after
they are processed and reclaim messages sent appropriately.
■ The lock server goes into grace mode and only allows clients to recover their
locks during the grace period. It does not give out any new locks during the
grace period.
postoffline trigger
The postoffline script does the following on IP failover:
■ Calls /opt/VRTS/bin/fsclustadm frlpause_disable to reduce the internal
usage counter.
■ Each call to /opt/VRTS/bin/fsclustadm frlpause_enable needs to be matched
with a call to /opt/VRTS/bin/fsclustadm frlpause_disable as the kernel
keeps an internal counter to track the number of IP addresses active on a
system. If there are no active IPs on a system, it will be in disabled mode.
Note: This trigger is only called for an administratively initiated failover. An actual
system failure and reboot discards the local state being manipulated in this stage.
This is the one trigger called on the node that was previously hosting the VIP,
while the others are called on the server taking over.
143Clustered NFS
Understanding how Clustered NFS works