I was using the term a bit loosely, perhaps. I was referring to the ability
to bring down an instance and apply a patch/patchset to the ORACLE_HOME on
that node while leaving the rest of the cluster running. At some point, the
entire cluster has to come down for a short while to apply database updates,
but the downtime for this is shorter than if I had a single ORACLE_HOME and
had to have the cluster down for both the database updates and software
updates. Once the database has been updated, those nodes that have the
patched software can be brought back up while the ORACLE_HOMEs on the
remaining nodes are updated.
Another strike against a single shared ORACLE_HOME is that it creates an
unncessary single point of failure.
John Smiley
Technical Management Consultant
TUSC, Inc.
On 6/22/05, Mladen Gogala <mgogala_at_allegientsystems.com> wrote:
>
> John Smiley wrote:
>
> > Because if you have a single, shared ORACLE_HOME, you can't do rolling
> > upgrades with RAC.
> >
>
> John, rolling upgrades are a legend. I work with RAC, I used to work
> with OPS and I've never
> seen that done. Allegedly, it has only been done in the land of El
> Dorado. It's easier to find WMD
> in Iraq then someone who has really done a rolling upgrade. The best you
> can do is to switch over
> to standby, upgrade the primary database, then upgrade the standby and
> rebuild the standby.
>
> --
> Mladen Gogala
> Oracle DBA
> Ext. 121
>
>
>
--
http://www.freelists.org/webpage/oracle-l
Received on Wed Jun 22 2005 - 16:00:27 CDT
