RE: ZERO Database Downtime???
- From: "Mercadante, Thomas F (LABOR)" <Thomas.Mercadante@xxxxxxxxxxxxxxxxx>
- To: <MFontana@xxxxxxxxx>, <oracle-l@xxxxxxxxxxxxx>
- Date: Wed, 29 Mar 2006 10:02:10 -0500
Michael,
Simply put, the "Rolling Upgrade" statement is a lie. Let me explain:
Lets say you have a RAC configuration. You need to perform a minor
upgrade (10.2.0.1 to 10.2.0.3). You shut down one node. Upgrade the
software. Now what do you do? Internal (Data Dictionary) patches need
to be applied (catalog, catproc etc). You can only do this with the
database up and (for safety reasons) with no users connected. How the
heck are you going to do this with the other node up and running?
I was told that this "is being addressed in Oracle 11x".
I would love to hear something different from the better educated folks
on this list.
Good Luck!
Tom
________________________________
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Michael Fontana
Sent: Wednesday, March 29, 2006 9:50 AM
To: oracle-l@xxxxxxxxxxxxx
Subject: ZERO Database Downtime???
Has anyone yet achieved this with Oracle 10g/GRID.
I just went to an Oracle 10g release2 presentation, where it was
unequivocally stated that with proper implementation of GRID, downtime
for database and server patching can be completely avoided using
"rolling upgrade" functionality.
I still have users who insist they must have complete application
availability even during schema changes.
Has anyone actually been successful at doing this?
Is it even worthwhile (time, effort, resources) to attempt it?
Is it safe?
Michael Fontana
NTT/Verio
Senior Database Administrator
mfontana@xxxxxxxxx
214.853.7405(Office)
214.912.3705(Wireless)
Other related posts:
