Re: ORA-04030: out of process memory errors
From: John Hurley <hurleyjohnb_at_yahoo.com>
Date: Mon, 21 Oct 2013 16:35:47 -0700 (PDT)
Message-ID: <1382398547.76507.YahooMailNeo_at_web181201.mail.ne1.yahoo.com>
Sandra: remember that ORA 4030's are problems with programs chewing up too much PGA memory ( not SGA ). Most of the time oracle bugs.
You should have trace files and/or entries in alert log that you can open an SR with Oracle against.
Last year we found a bad problem for us with a program that ran a bunch of updates on a text index. Oracle had a fix already identified but decided apparently that it was not common enough to include in regular maintenance. ( Heres my blog post on that one http://grumpyolddba.blogspot.com/2013/04/nasty-ora-4030-in-112035-using-text.html ).
Odds are very good that oracle support can identify a specific fix for you probably already on hand.
11.2.0.2 ? Really ? Odds are even better than the fix is probably built into ( perhaps 11.2.0.3 ) or almost certainly 11.2.0.4 ... but you need to know what oracle bug ...
If you have something configured like OS Watcher and get it running every 60 seconds then you should be able to look at and correlate minute by minute os memory stats and process statistics against when the 4030's happen. That will lead you directly to the suspects ...
From: Sandra Becker <sbecker6925_at_gmail.com> To: oracle-l <oracle-l_at_freelists.org> Sent: Monday, October 21, 2013 1:08 PM
Subject: ORA-04030: out of process memory errors
Date: Mon, 21 Oct 2013 16:35:47 -0700 (PDT)
Message-ID: <1382398547.76507.YahooMailNeo_at_web181201.mail.ne1.yahoo.com>
Sandra: remember that ORA 4030's are problems with programs chewing up too much PGA memory ( not SGA ). Most of the time oracle bugs.
You should have trace files and/or entries in alert log that you can open an SR with Oracle against.
Last year we found a bad problem for us with a program that ran a bunch of updates on a text index. Oracle had a fix already identified but decided apparently that it was not common enough to include in regular maintenance. ( Heres my blog post on that one http://grumpyolddba.blogspot.com/2013/04/nasty-ora-4030-in-112035-using-text.html ).
Odds are very good that oracle support can identify a specific fix for you probably already on hand.
11.2.0.2 ? Really ? Odds are even better than the fix is probably built into ( perhaps 11.2.0.3 ) or almost certainly 11.2.0.4 ... but you need to know what oracle bug ...
If you have something configured like OS Watcher and get it running every 60 seconds then you should be able to look at and correlate minute by minute os memory stats and process statistics against when the 4030's happen. That will lead you directly to the suspects ...
From: Sandra Becker <sbecker6925_at_gmail.com> To: oracle-l <oracle-l_at_freelists.org> Sent: Monday, October 21, 2013 1:08 PM
Subject: ORA-04030: out of process memory errors
Solaris 5.10, 128G RAM, 8G swap
Oracle EE 11.2.0.2
We've been getting the ORA-04030 errors several times a week since they did
an OS patch in August. The SE team says there isn't anything wrong with
the server or configuration on their side, although they do see some swap
messages in their logs. Other than the OOM errors, we don't see any issues
in the database.
What should we be looking at/for to determine the cause? The databases and server they are running on have been in place for at least of couple of years. Does anyone know of memory issues with this version of Solaris?
Thanks in advance for any suggestions or ideas.
-- Sandy Transzap, Inc. -- http://www.freelists.org/webpage/oracle-l -- http://www.freelists.org/webpage/oracle-lReceived on Tue Oct 22 2013 - 01:35:47 CEST