Oracle DB RAC Server Reboot Process

 ORACLE RAC DATABASE SERVER REBOOT PROCESS

-----------------------------------------

Current RAC 4 node Setup

• FOR EASYDB (4-NODE RAC DATABASE on ASM) with four Node Standby Database

➢ DOWN PROCESS:

1) Check log sequence number of all four nodes and make sure it is in sync (DC & DR).

SQL> SELECT MAX(SEQUENCE#),THREAD# FROM V$LOGHIST GROUP BY THREAD#; (both DC & DR).

2) DEFER dataguard (DC side).

SQL> ALTER SYSTEM SET log_archive_dest_state_2='DEFER' SID='*';

3) Stop MRP process (DR side).

SQL> alter database recover managed standby database cancel;

4) Down all Instances.

Login with ‘oracle’ session on master node i.e., EASYNODE1:

EASYNODE1:/home/oracle#srvctl stop database –d EASYDB (both DC & DR).

5) Dismount diskgroups such as EASYDBDATA_DG1, EASYDBDATA_DG2,

EASYDBFRA_DGNEW, EASYDBREDO_DGNEW, and EASYDBTEMP_DGNEW.

Login with ‘orarac’ session and connect to ‘sqlplus / as sysasm’:

SQL> alter diskgroup EASYDBDATA_DG1 dismount;

SQL> alter diskgroup EASYDBDATA_DG2 dismount;

SQL> alter diskgroup EASYDBFRA_DGNEW dismount;

SQL> alter diskgroup EASYDBREDO_DGNEW dismount;

SQL> alter diskgroup EASYDBTEMP_DGNEW dismount;

6) Stop ASM service.

Login with ‘orarac’ session on master node i.e., EASYNODE1:

EASYNODE1:/home/orarac#srvctl stop asm

7) Stop cluster service and handover to UNIX team for reboot the nodes.

Login with ‘orarac’ session on master node i.e., EASYNODE1:

EASYNODE1:/home/orarac#srvctl stop asm

EASYNODE1:/home/orarac#crsctl stop cluster all

or 

stop cluster/crs on all node by root user after shutting down database .

cd /GRIDHOME/bin

./crsctl stop crs - run on each node or one by one as per requirement

➢ UP PROCESS:

1) After get back the servers from UNIX team, First check binary mount points

such as ‘/oracle’ & ‘/asm_grid’ from all 4 nodes.

EASYNODE1:/home/dxc_subhankar#df –gt

2) Start cluster services if it is not started automatically

Login with ‘orarac’ session:

EASYNODE1:/home/orarac#crsctl start cluster all

3) Start ASM services.

Login with ‘orarac’ session on master node i.e., EASYNODE1:

EASYNODE1:/home/orarac#srvctl start asm

4) Mount diskgroups such as EASYDBDATA_DG1, EASYDBDATA_DG2, EASYDBFRA_DGNEW,EASYDBREDO_DGNEW, and EASYDBTEMP_DGNEW.

Login with ‘orarac’ session and connect to ‘sqlplus / as sysasm’:

SQL> alter diskgroup EASYDBDATA_DG1 mount;

SQL> alter diskgroup EASYDBDATA_DG2 mount;

SQL> alter diskgroup EASYDBFRA_DGNEW mount;

SQL> alter diskgroup EASYDBREDO_DGNEW mount;

SQL> alter diskgroup EASYDBTEMP_DGNEW mount;

or 

start the clusterware if it is not started from root user

cd $ORACLE_GRID/bin

./crsctl start crs -- run it on all node or one by one based on requirement

5) UP all Instances.

Login with ‘oracle’ session on master node i.e., EASYNODE1:

EASYNODE1:/home/oracle#srvctl start database –d EASYDB (both DC & DR).

6) ENABLE dataguard (DC side).

SQL> ALTER SYSTEM SET log_archive_dest_state_2='ENABLE' SID='*';

7) Start MRP process (DR side).

SQL> alter database recover managed standby database disconnect from session;

8) Do some alter switch (DC side) and check archives are receiving (DR side).

SQL> alter system switch logfile;

9) Do tnsping to check network connectivity, if everything looks fine handover to application team.