We've got a vRO cluster running with an MSSQL database backend that is also clustered. When the SQL database is failed over to the secondary for maintenance we find that both vROs gets disconnected from the database and don't reconnect and require us to restart the application servers to recover. Is this expected behaviour? Is there a way to extend the timeout for the database connection or anything along those lines?
Upgrade to vRO 6.0.3. VMware added a watchdog timer that will keep the vco-server process running during MSSQL failover or reboot. I've only tested reboots and it works as advertised.