Configuring High Availability to reduce NFS timeouts
Because the two Moab servers access the same files, you only need to configure high availability (HA) in the moab.cfg file.
Configuring HA mode to reduce NFS timeouts
- Start pbs_server with HA.
> pbs_server --ha -l host1:<port> -l host2:<port> h1.
<port> is Moab's scheduler port. I ts default value is 15004.
Only specify the Moab hosts and ports if Moab HA is configured on a remote server. Otherwise, run pbs_server --ha. For example, if Moab is running on server1 and you wish to start TORQUE in HA, you only need to make TORQUE aware of server 2.
> pbs_server --ha -l server2:<port> h1.
- Use the SERVER and FBSERVER SCHEDCFG attributes to configure the primary and fallback (respectively) hosts that run Moab.
- To turn on file lock, set FLAGS=filelockha and specify the lock file using the HALOCKFILE attribute.
- Use the HALOCKUPDATETIME parameter to specify how frequently the primary server updates the time stamp on the lock file. Use HALOCKCHECKTIME to specify how frequently the secondary server checks the time stamp on the lock file.
- Confirm that your configuration works. To do so, run mdiag -S -v to view which Moab server is currently scheduling and responding to client requests.
SCHEDCFG[Moab] SERVER=host1:42559
SCHEDCFG[Moab] FBSERVER=host2
host1:42559 is the primary Moab server and host2 is the secondary Moab server.
FBSERVER does not take a port number. The primary server's port is used for both the primary server and the fallback server.
SCHEDCFG[Moab] FLAGS=filelockha
SCHEDCFG[Moab] HALOCKFILE=/opt/moab/.moab_lock
File lock is turned on and uses the /opt/moab/.moab_lock file.
HALOCKCHECKTIME 9
HALOCKUPDATETIME 3
In the preceding example, the secondary server checks the lock file for updates every 9 seconds. The HALOCKUPDATETIME parameter is set to 3 seconds, permitting the primary server three opportunities to update the timestamp for each time the secondary server checks the timestamp on the lock file.
© 2012 Adaptive Computing 