A Data Protector 10.xx Cell Manager running on linux within a ServiceGuard (SG) cluster
With versions of Data Protector running in a prior to 10.00 a time out of 3 seconds was usually adequate.
With Data Protector 10 versions we find 3 seconds is not long enough.
If you have unwanted failovers please check the value of
$OB2SBIN/omnisv.sh start_mon -timeout
This can be seen within a `pf -ef | grep omni` output.
Consider increasing the value.
If it is 3 seconds we suggest extending it to 30.
This is done by editing the csfailover.ksh script changing the line…
"$OB2SBIN/omnisv.sh start_mon -timeout 3"
"$OB2SBIN/omnisv.sh start_mon -timeout 30"
The changes will only be active when the Data Protector package is next re-started.
Keyword list:- Unexplained intermittent Failover Failovers no reason metro.