Relinking Grid Infrastructure

This is the official document when you need to relink Oracle after an o/s kernel upgrade.

How To Relink The Oracle Grid Infrastructure Standalone (Restart) Installation Or Oracle Grid Infrastructure RAC/Cluster Installation (11.2 or 12c). (Doc ID 1536057.1)

Note the roothas.sh or the rootcrs.sh -unlock changes the ownership of the grid home from root to oracle to allow for updates on that directory. You can see this by running the "ls" command before and after running that script.

For RAC follow :

Instruction Steps
Task Steps		Reference/Screenshots
1.	Verify that cluster services is running on all nodes	$GRID_HOME/bin/crsctl stat res -t
2.	Shutdown OEM agent – if available	$AGENT_HOME/bin/emctl stop agent
3.	Shutdown or Failover Goldengate if required
4.	As root, disable automatic restart for HAS	sudo bash . /usr/local/bin/oraenv +ASM1 (or active ASM instance) $ORACLE_HOME/bin/crsctl disable has
5.	Shutdown HA services to prepare for relinking	$ORACLE_HOME/crs/install/rootcrs.sh -prepatch For pre-12c installs use: perl $ORACLE_HOME/crs/install/rootcrs.pl -unlock
6.	Confirm all Oracle processes are down, apply O/S patches and reboot	ps -ef | grep -E 'oracle|has'
7.	After reboot, login as oracle and verify HAS did not start	$GRID_HOME/bin/crsctl stat res -t
8.	As oracle (or owner of $GRID_HOME), relink executables in GRID_HOME	. /usr/local/bin/oraenv +ASM1 (or active ASM instance) $ORACLE_HOME/bin/relink all
9.	Check log for relink errors	egrep 'error|fail' $ORACLE_HOME/install/relink.log
10.	Double check relinked binaries. Files should have date stamp of current date/time	ls -ltr  $ORACLE_HOME/bin
11.	As oracle, relink executables in DB_HOME	. /usr/local/bin/oraenv nextfare1 (or other db name on server) $ORACLE_HOME/bin/relink all
12.	Check log for relink errors	egrep 'error|fail' $ORACLE_HOME/install/relink.log
13.	Double check relinked binaries. Files should have date stamp of current date/time	ls -l  $ORACLE_HOME/bin
14.	As root, re-enable HAS	sudo bash . /usr/local/bin/oraenv +ASM1 (or active ASM instance) $ORACLE_HOME/bin/crsctl enable has
15.	As root, run rootadd_rdbms.sh	$ORACLE_HOME/rdbms/install/rootadd_rdbms.sh
16.	If GRID_HOME is owned by a user other than oracle, ie grid, verify <DB_HOME>/bin/oracle has correct permission/uid/gid as below.  If not, run setasmgidwrap to reset the oracle binary owner/group. Replace <DB_HOME> with correct path for database home. Verify permissions/uid/gid as below: ls -l <DB_HOME>/bin/oracle permissions: -rwsr-s—x uid: oracle gid: asmadmin	su – grid $ORACLE_HOME/bin/setasmgidwrap o=<DB_HOME>/bin/oracle
17.	Restart HA services	$ORACLE_HOME/crs/install/rootcrs.sh -postpatch For pre-12c installs use: perl $ORACLE_HOME/crs/install/rootcrs.pl -patch
18.	Verify that HA services have restarted	$ORACLE_HOME/bin/crsctl stat res -t
19.	As oracle, start OEM agent – if available	$AGENT_HOME/bin/emctl start agent
20.	Start Goldengate if required
21.	Repeat Steps 2-19 for each additional node replacing ASM instance and DB instance as needed.
22.