MPX restore was on 60 seconds
and unmount delay was on 30 seconds, so I se set those per your instructions. I’ll
forward your note to the ORACLE people. Thank you.
From: Kevin Corley
[mailto:Kevin.Corley AT apollogrp DOT edu]
Sent: Wednesday, November 04, 2009 12:20 PM
To: Reynolds, Susan K.
Subject: RE: Oracle restore question
From
the admin console, and in the media server properties, under General Server is
the delay on multiplexed restore setting. Don’t set it too high, but if you set
it to like 240 seconds, the tape will stay in the drive and wait 4 minutes for
another oracle restore job to go active before unloading.
This is
assuming the DBA is using multiple rman channels in the restore, which can be
done with the allocate command:
allocate
auxiliary channel t1 type 'SBT_TAPE'
parms="ENV=(NB_ORA_CLIENT=dbservername)";
allocate
auxiliary channel t2 type 'SBT_TAPE'
parms="ENV=(NB_ORA_CLIENT=dbservername)";
allocate
auxiliary channel t3 type 'SBT_TAPE'
parms="ENV=(NB_ORA_CLIENT=dbservername)";
allocate
auxiliary channel t4 type 'SBT_TAPE' parms="ENV=(NB_ORA_CLIENT=dbservername)";
Also
check the media unmount delay in the media server properties, under media. If
that is too low, it will unmount the tape too soon. Default is 300 seconds.
From: Reynolds, Susan K.
[mailto:SReynolds2 AT semprautilities DOT com]
Sent: Wednesday, November 04, 2009 1:12 PM
To: Kevin Corley
Subject: RE: Oracle restore question
Hi
Kevin…how is that done? Susan
From: Kevin Corley
[mailto:Kevin.Corley AT apollogrp DOT edu]
Sent: Wednesday, November 04, 2009 12:11 PM
To: Reynolds, Susan K.; veritas-bu AT mailman.eng.auburn DOT edu
Subject: RE: Oracle restore question
Look
at increasing the “delay on multiplexed restores” on the media server and also
increasing the number of channels being allocated in the rman restore script.
From: veritas-bu-bounces AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Reynolds,
Susan K.
Sent: Wednesday, November 04, 2009 12:55 PM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Oracle restore question
We have a group of DBAs running a
restore and I was wondering if there was something in the way they kicked off
their restore that caused the following:
“They initiate restore, tape is
loaded, image is restored, tape is returned to slot in the library…then restore
moves to the next image…goes back…loads the same tape, restores the images,
tape is returned to slot in the library.”
Is there something on the Oracle
side that needs to be specified that says, “load tape, restore ALL the relevant
images on the tape (without unloading and reloading it again in a drive), then
go to the next tape when the restore is completely done with the images on the
first tape.”
This loading and unloading is
adding 3-4 minutes to the restore of each image.
Are there any oracle experts out
there who understand why this is happening (so we’ll know for the future and it
will not be such a time-consuming process)?
|
This message is private and confidential. If you have received
it in error, please notify the sender and remove it from your system.
|
|
This message is private and confidential. If you have received
it in error, please notify the sender and remove it from your system.
|
