Re: Oracle and Linux kernal versions
- From: Carel-Jan Engel <careljan@xxxxxxxxxx>
- To: Jon.Crisler@xxxxxxx
- Date: Tue, 09 Oct 2007 18:07:44 +0200
In theory, and practically, it will _probably_ work. I don't know
whether I would relink the Oracle software.
Data Guard documentation (Concepts & Administration Manual, for release
9.2, Page 2-6 says that the OS releases may differ between Primary and
Standby, but Oracle SW release must be the same.
Others might be able to advise about the relink better than me.
I would go along the following steps:
1. Create a test database at the Primary site (always good to have
one, to be able to test the actual stack). It can be minimally
sized.
2. Instantiate this test db to the Standby, and bring the standby
to managed recovery mode.
3. Make a copy of the Oracle home at the SB,
e.g. /opt/oracle/product/9.2 to /opt/oracle/product/9.2a
4. Stop redo transport to both test database and production
database at the Standby
5. upgrade the kernel at the Standby
6. Relink Oracle in the copied Oracle home 9.2a
7. Start the test instance from the old oracle home and bring it to
managed recovery mode
8. create some load at the test on the Primary, and see how the
Standby behaves with the old code
9. If everything OK, you can run the production db at the standby
from the old oracle_home
10. If not OK, test the test-Standby from using the relinked code
11. If OK now, run the production database at the Standby from the
new oracle home.
This approach is not 100% guaranteed, as your test database might not
completely represent whatever is happening in your production database.
But for now, I think you cannot get any closer. But feel free to teach
me else ;-)
Best regards,
Carel-Jan Engel
===
If you think education is expensive, try ignorance. (Derek Bok)
===
On Tue, 2007-10-09 at 11:13 -0400, Crisler, Jon wrote:
> We have a situation where we need to fix a SAN disk issue on the
> target side of a Oracle 9.2.0.7 with Dataguard. The fix is to upgrade
> the kernel version from 2.6.42 to 2.6.55. However, we cannot get
> downtime right now to fix the source side.
>
>
>
> My feeling is that it is ok to update the target Linux server kernel
> for now, and they update the source side as time permits. The only
> thing I am advising is to relink Oracle. Any thoughts on this
> strategy?
>
>
>
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
Other related posts:
