Re: Patching RAC

From: jason arneil <jason.arneil_at_gmail.com>
Date: Wed, 28 May 2008 08:40:59 +0100
Message-ID: <fda4898f0805280040y7b17473flad9ace22a70abe5@mail.gmail.com>


Hello,

Here is my 2p worth:

Are you going to be running the view_recompile sql script? If so, this requires that the database is opened with upgrade. Of course to get the full benefit of patching you need to do this recompilation. So you could argue, if you are needing downtime anyway, if you felt uncomfortable running catpatch.sql with the db open, it will not add much downtime to run this while the db is in the startup upgrade state.

That being said I have applied the catpatch.sql with a RAC db open and it was fine. I think Oracle are under increasing pressure to ensure that patches can be applied with the db open, so this is probably going to be more and more of an occurance - you don't want to pay huge amounts of money for RAC to then find you need downtime to plug security holes 4 times a year.

Another option, if this is open to you, is instead of taking a downtime hit with the april CPU, is to take the downtime hit with an upgrade to 10.2.0.4, as this includes the fixes from the April CPU and you'll some extra bug fixes and enhancements. You might say this requires more testing, but, you are going to test the CPU anyway, right?

jason.

--
http://jarneil.wordpress.com




On Tue, May 27, 2008 at 5:16 PM, Koivu, Lisa <Lisa.Koivu_at_starwoodvo.com>
wrote:


> Hello all,
>
>
>
> I am the proud mother of a new RAC installation J
>
>
>
> My first task with this RAC system was to patch it. I successfully
> completed a rolling patch on the clusterware, and then patched the asm and
> db homes separately. I was thinking through how the database patches had to
> be applied (catpatch.sql) and it makes sense to me that the database should
> be unavailable throughout the execution of this script, as we are updating
> the data dictionary within the database (which there is only one of.)
>
>
>
> Now I'm testing the April 2008 CPU. The docs say that the database can be
> up and available while executing catcpu.sql. It also makes no mention of
> patching the clusterware.
>
>
>
> My questions are:
>
>
>
> 1. Has anyone ever had a CPU patch go south because the database was in
> use?
> 2. Am I the only one who is nervous running something that is modifying
> the data dictionary while the database is open for general use?
> 3. Are there CPU patches for the clusterware? I don't see any.
>
>
>
> And as a tidbit, to those of you who may not already know this: When
> creating a database with dbca, don't use the template databases if you can
> avoid it (general purpose, data warehouse, transaction processing). These
> preconfigured databases save time within the database creation process, but
> they include the java engine within the database. It takes a long time to
> patch the jvm.
>
>
>
> (May eons ago I had a major patch end up in a recovery because of the
> database jvm. What a nightmare that was)
>
>
>
>
>
> *Lisa Koivu*
>
> *Oracle Database Administrator*
>
> *desk: 407-903-4691*
>
> *cell: 954-683-4459*
>
>
>
> This electronic message transmission contains information from the Company
> that may be proprietary, confidential and/or privileged. The information is
> intended only for the use of the individual(s) or entity named above. If you
> are not the intended recipient, be aware that any disclosure, copying or
> distribution or use of the contents of this information is prohibited. If
> you have received this electronic transmission in error, please notify the
> sender immediately by replying to the address listed in the "From:" field.
>
-- http://www.freelists.org/webpage/oracle-l
Received on Wed May 28 2008 - 02:40:59 CDT

Original text of this message