Instance Does not Start After 11gR2 Grid Infrastructure Restart or Node Reboot [ID 1149003.1] - Oracle Support

Instance Does not Start After 11gR2 Grid Infrastructure Restart or Node Reboot [ID 1149003.1]
	Modified 08-JAN-2012     Type PROBLEM     Status PUBLISHED

In this Document
  Symptoms
  Cause
  Solution

---

Applies to:

Oracle Server - Enterprise Edition - Version: 11.2.0.1 and later   [Release: 11.2 and later ]
Information in this document applies to any platform.

Symptoms

Some database instances of 11gR2 databases does not come up automatically after Grid Infrastructure(GI) restarts or node reboots.

In example below, database b2n is enabled with MANAGEMENT_POLICY=AUTOMATIC, AUTO_START=restore(by default) and a service b2ns(preferred instance1), and has one instance on each node:

$GRID_HOME/bin/crsctl stat res ora.b2n.db -t
ora.b2n.db
     1        ONLINE  ONLINE       racnode1                    Open            
     2        ONLINE  ONLINE       racnode2                    Open

$GRID_HOME/bin/crsctl stat res ora.b2n.b2ns.svc -t
ora.b2n.b2ns.svc
     1        ONLINE  ONLINE       racnode1

$GRID_HOME/bin/crsctl stat res ora.b2n.db -p
NAME=ora.b2n.db
..
AUTO_START=restore
..
ENABLED=1
..
MANAGEMENT_POLICY=AUTOMATIC
..

$GRID_HOME/bin/srvctl config service -d b2n -s b2ns
Service name: b2ns
..
Preferred instances: b2n1
Available instances:

If the database is stopped with srvctl before restart of GI, once GI comes back up, both instances stay down.
But if the database is stopped with sqlplus, once GI comes back up, instance 1 starts automatically but instance 2 stays down.

Cause

As AUTO_START is "restore", GI will "restore" database state while starting up; so if database was brought down prior to restart of GI, it should stay down once GI comes up.

In agent log:$GRID_HOME/log/$HOST/agent/crsd/oraagent_user/oraagent_user.log

2010-07-07 19:07:42.982: [    AGFW][7] check for resource: ora.b2n.db 2 1 completed with status: PLANNED_OFFLINE
..
2010-07-07 19:07:42.988: [    AGFW][53] ora.b2n.db 2 1 received state from probe request. Old state = UNKNOWN, New state = PLANNED_OFFLINE

But if the instance was brought down with sqlplus, service resource "TARGET" will stay ONLINE, thus when GI comes up, it will try to "restore" service resource state; as service resource depend on instance, the instance will be brought up.

Solution

This is expected behaviour.