Found in the docs:
".... If the instances that support these services, including available
instances that the service uses for failover, are not running but are
enabled, then SRVCTL starts them. "
Some reason I've never noticed that because I guess I'm usually working on
nodes that the instances are up, or standby dbs that are only 1
On Thu, Aug 6, 2020 at 2:24 PM Chris Taylor <
Setting up a new Standby DB on a RAC cluster
I configured the services with " -role "PRIMARY,SNAPSHOT_STANDBY" so that
the services will only run when the role is primary or snapshot_standby.
Now when I start the service, it brings up the database on the other 3
nodes and tries to start the db on node1 (but it's already running)
[oracle@xmp2-cluster101vm01 ~]$ srvctl start service -db db_name -service
PRCD-1084 : Failed to start service svc_name
PRCR-1079 : Failed to start resource ora.db_name.svc_name.svc
CRS-2800: Cannot start resource 'ora.db_name.db' as it is already in the
INTERMEDIATE state on server 'node_name'
CRS-2632: There are no more servers to try to place resource
'ora.db_name.svc_name.svc' on that would satisfy its placement policy
CRS-2674: Start of 'ora.db_name.db' on 'node2' failed
CRS-2674: Start of 'ora.db_name.db' on 'node3' failed
CRS-2674: Start of 'ora.db_name.db' on 'node4' failed
Even though it says it failed on the other 3 nodes, it did start up the
Is there was some way to turn that behavior off?
Did configuring the services with a "-role" flag have this side effect?