I currently work on a 4-node cluster with 12.1.0.2 Grid Infrastructure installed, hosting several 11g and 12c databases. As it is a cold failover cluster, databases are not RAC databases and are active on only one node at a time.
Thoses databases share the same action script. The “start” section of this action script is quite complex and also manages the relocation of the database target on Enterprise Manager 12.1.0.5. But this specific task is time-consuming (approximately 30 seconds for 1 target) and is not always successful. Which means I often have to relocate some targets in Enterprise Manager by hand.
Recently, 2 nodes of our cluster did not survive a major storage outage. After their reboot, the databases restarted (quite slowly) according to the action script. But target relocation did not work as expected for all targets and I had to manually fix the mess.