RE: Rman Locking

We run into this problem quite a lot.

The issue seems to be that during a RMAN backup it takes an enqueue lock = (seems to be of type CHPT) against the RMAN catalogue. This is held whilst it searches for obsolete/redundant rows to delete. As you will be aware there is quite a lot of referential integrity = within the catalogue and some of the full table scans can be quite = intensive on a very large catalogue.
The process then exaggerates as another backup is kicked off from a = different target database.

The end result is the need to kill of the locked processes and restart. Solutions can be to reduce the number of concurrent rman backups from = seperate servers and to increase the number of catalogues in use. We have a RMAN recovery database with a number of schemas which we back = various versions of Oracle software into and we can also split into = schemas for application types (i.e. rman8i, rmman9i, rmancust etc)

HTH
John

-----Original Message-----

From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org]On Behalf Of Spears, Brian Sent: Tuesday, November 09, 2004 4:35 PM To: Oracle-L (E-mail)
Subject: Rman Locking

Anybody ever run into a resync catalog while backing up archive logs = causing
problems. It is happening once a week randomly. What happens is = deadlock
on chp table and of course all other Rman stuff is held up. It seems to = be
this one small rarely used database. The way we solve it is to go out = and
kill the process on unix and everything is freed up. I would like to say that the network is burping or something and disconnecting the process ..but its on the same database all the time.

Brian

--

http://www.freelists.org/webpage/oracle-l
--

http://www.freelists.org/webpage/oracle-l Received on Wed Nov 10 2004 - 02:53:25 CST