RAC: How to restart a node in RAC Cluster for OS maintenance

Ticket:

Hi Oracle DBA,

UNIX team needs to apply some OS Patches on linux-1. We believe this node  has an active Oracle instance. Please shutdown the oracle instance and stop CRS. Notify Unix team when done.

Solution:

Phase-1

How to shutdown Oracle CRS on a node?

10g	11g R2
DBA Performs the following: DBA notifies the teams involved about database instance going down. ps -ef |grep pmon. You see an instance called racDB1 srvctl stop instance -d racDB -i racDB1 srvctl stop ASM -n linux-1 srvctl stop nodeapps -n linux-1 DBA Notifies UNIX Admin to perfom the following unless DBA has the sudo permissions to crsctl: crsctl disable crs crsctl stop crs	DBA Performs the following: DBA notifies the teams involved about database instance going down. ps -ef |grep pmon. You see an instance called racDB1 srvctl stop instance -d racDB -i racDB1 srvctl stop has DBA Notifies UNIX Admin to perfom the following unless DBA has the sudo permissions to crsctl: crsctl disable crs crsctl stop crs

Phase-2

Unix Admin can start their work. Once done, they UNIX team notifies the DBA to start the Oracle instance.

Phase-3

How to start the CRS and start the

10g

11g

DBA notifies UNIX Admin to perform the following unless DBA has the sudo permissions to crsctl: crsctl enable crs crsctl start crs Once UNIX Admin confirms the above two, DBA performs the following: Verify CRS is up crsctl check crs srvctl start nodeapps -n linux-1 srvctl start ASM -n linux-1 srvctl start instance -d racDB -i racDB1 verify all services are up crs_stat or crs_stat -t If any of the services are still down, DBA can individually start the services or start all services at once. crsctl start resources. When verification is done all services are up, notify Team database instance is up and ready.