Re: PSU - Application Order

From: Andy Colvin <acolvin_at_enkitec.com>
Date: Sat, 30 Jun 2012 08:14:31 -0500
Message-Id: <A58E8038-DDD7-44C5-8AEB-85AF570734D4_at_enkitec.com>

The rule for the GI home to be same or newer refers to a home with processes running. Since the database home isn't started back up until after GI is patched, the rule isn't broken. The thought process (my guess) is that if you have multiple homes, your downtime overall won't be as large. Say that you had 3 database homes, each running different databases, and you wanted to patch all of them. The patch process would be:

shut down and patch dbhome_1
shut down and patch dbhome_2
shut down and patch dbhome_3

shut down and patch GI home
start cluster

This way, the databases in dbhome_3 aren't down as long as they would be if you patched the GI home first. A different option would be to patch the GI home, followed by each db home, starting them as the patches complete. Personally, I like to use the -oh switch and patch the GI home first. Once that's done, I'll move on to the database homes.

Andy Colvin

Principal Consultant
Enkitec
andy.colvin_at_enkitec.com
http://blog.oracle-ninja.com

On Jun 26, 2012, at 8:04 AM, Purav Chovatia wrote:

> Hi,
> I observed that when applying PSU via "opatch auto", the patch is first
> applied to DB home and then to GI home. Because GI should always be at a
> higher version, shouldn't the patching order have been other way round?
>
> Also, on of our sites, I observed that applying 112022 via "opatch
> auto" succeeded on DB home but failed on GI home. Again, because GI should
> always be at a higher version, isn't that incorrect?
>
> Thanks.
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>

--
http://www.freelists.org/webpage/oracle-l

Received on Sat Jun 30 2012 - 08:14:31 CDT