Hi
I think control files management is so low level that Oracle didn't spend
time taking care of the scenario of online changes to control_files - loss
of 1 at least. It was most likely done as a compromise to code complexity
long-long time ago. It would be a good improvement to availability to allow
online control_file changes I think. However I've no idea if it's something
doable at all.
ER anyone?
On Tue, Nov 15, 2016 at 9:59 PM, Robert Freeman <rfreeman@xxxxxxxxxxxxxxxxx>
wrote:
I’ve been sitting here wondering about a subject that might seem a bit
mundane, but I thought I’d throw the question out there to see what others
think…
What is the logic in terminating a database/database instance should you
lose one of your control files, when other copies of that control file
continue to survive?
Note, I am talking specifically about abnormally terminating the database
instance rather than preventing it from starting up after a shutdown has
been executed.
I realize that one might argue, but what if you lose the surviving control
file? This does not seem like a reasoned response to the problem though.
I mean, if you lose a redo log member, your database does not crash.
You can lose a whole redo log group and the database won’t crash. All of
which are good things… J
This just seems to be an odd thing to me. Thoughts…
Robert
Robert G. Freeman
Deliverer of Data
Businessolver
Cell: 801-703-3405
“Personally, I think that’s a hell of a bird.”
-- Dr. Who (Peter Capaldi)
[image: calvindba]