Thursday, June 23

10G patching

Just finished patching one of our 10g development systems to the latest patchset from Oracle (10.1.0.4 Patchset 2). Smooth as always with 10g. catpatch.sql may take quite some time on slow systems, same with utlrlp.sql.

1. Download patchset 4163362 (and the README!) from Metalink and decompress it.
2. Stop the listener, cssd, em, all instances etc.
2b. If you are upgrading an important production box (do test on a staging server before going live!), do a backup. I'm to lazy to do this for a development box.
3. Start the runInstaller and click through all the defaults. (I always do this in a VNC session to the server so I can detach if I get bored or need to reboot my WS).
4. For each instance login (using sqlplus) as sysdba and run:
SQL> STARTUP UPGRADE
SQL> SPOOL patch.log
SQL> @?/rdbms/admin/catpatch.sql
SQL> SPOOL OFF
[ review patch.log for any errors, fix
them and if needed run catpatch.sql again]

SQL> SHUTDOWN
SQL> STARTUP
SQL> SPOOL utlrlp.log
SQL> @?/rdbms/admin/utlrp.sql
SQL> SPOOL OFF
[ review utlrlp.log for errors, ignore recyclebin
errors, shouldn't be any other errors really (unless
you i.e. disabled public execute access to utl_file as
Oracle so kindly recommends) ]
5. Start the listener, em etc.
Done!

Don't forget to upgrade RMAN catalogs and stuff like that if you use that and if you get problems with EM just reset the timezone as described in the Readme.