What is the correct order to shutdown and startup RAC system

Question

anyone please take a look if my order of shutdown and restart is correct.
This is two node RAC,
1) srvctl stop database -d db_unique_name
2)srvctl stop nodeapps -n node1
3)srvctl stop nodeapps -n node2
4)srvctl stop asm -n rac1
5)srvctl stop asm -n rac2
6)srvctl stop listener -n rac1
7)srvctl stop listener -n rac2
8) crsctl stop cluster all
when I restart RAC, use
1)crsctl start cluster all
2)srvctl start listener -n rac1
3)srvctl start listener -n rac2
4)srvctl start asm -n rac1
5)srvctl start asm -n rac2
6)srvctl start nodeapps -n node1
7)srvctl start nodeapps -n node2
8)srvctl start database -d db_unique_name

Answer 1

Hi,
1.) I would specify -o immediate. Otherwise your stop of the database can take quite a while.
2.) Actually nothing else (except the database) needs a clean shutdown. I would simply stop the cluster than with crsctl stop cluster all.
4.) Also note if you have Voting and OCR in ASM this step won't be successfull (since it can't stop ASM as long as the cluster is running).
8.) Note that crsctl stop cluster will leave the OHASD running (and also some other processes, check with crsctl stat res -t -init).
If you need to really stop everything you have to execute crsctl stop crs on each node (they are not clusterwide).
Startup:
CRSCTL start CLUSTER ALL should actually bring a lot of things up... but this depends also on what has been stopped (since this is recorded in the OCR).
But at least ASM and listener should startup automatically (even if they were stopped before).
So normally it is enough to start the database. But I would not start the database, I would startup the services (which the database services). Because this will bring up the database.
Otherwise you might end up with a running database, but with no active services.
Regards
Sebastian