RE: Max Logging

From: Powell, Mark D <mark.powell_at_eds.com>
Date: Fri, 12 Dec 2008 09:29:32 -0500
Message-ID: <D1DC33E67722D54A93F05F702C99E2A9033AE0BC@usahm208.amer.corp.eds.com>

At the OS level two suggestions:  

Double check the kernel parameters settings specified in the vendor specific install guide.
Verify you have adequate page/swap space allocated

In Oracle:

Get the hang analyze dump and look to see if you can identify what Oracle is waiting on. If not or if you need help with figuring out what to do open an SR with support and upload the hang analyze and system state dumps.

  • Mark D Powell -- Phone (313) 592-5148

-----Original Message-----
From: Howard Latham [mailto:howard.latham_at_gmail.com] Sent: Friday, December 12, 2008 2:45 AM
To: Powell, Mark D
Cc: ORACLE-L
Subject: Re: Max Logging

Its the whole server hanging most nights. I have shut down the database two nights running and the server hasnt hung. Its beginning to look like someyhing oracle does or asks for that is killing the box.

On 11/12/2008, Powell, Mark D <mark.powell_at_eds.com> wrote:
>
> If by lockup you mean hang of the Oracle database and not the entire
> server then besides the system state dump Mark mentioned I would take
> a hang analyze dump:
>
> oradebug setmypid
> oradebug unlimit
> oradebug hanganalyze 3
> select 'Wait 1 minute then repeat hang cmd' from dual; oradebug
> tracefile_name
>
> OR
>
> ALTER SESSION SET EVENTS 'IMMEDIATE TRACE NAME HANGANALYZE LEVEL 3';
>
> -- Mark D Powell --
> Phone (313) 592-5148
>
>
> -----Original Message-----
> From: oracle-l-bounce_at_freelists.org
> [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Bobak, Mark
> Sent: Thursday, December 11, 2008 11:49 AM
> To: howard.latham_at_gmail.com; ORACLE-L
> Subject: RE: Max Logging
>
> Hi Howard,
>
> You could set SQL tracing at the instance level, but that's probably
> almost certainly NOT what you want to do.
>
> Most likely, when the lockup happens, you'll want to take a system
> state dump, wait a few minutes, and take another one.
>
> Oracle support will be able to use the two dumps to determine if the
> system is REALLY locking up, or if it's just running really slowly.
> They'll look at the two dumps and see if they look the same. If the
> system state changed between the two dumps, then it's not a lockup,
> just a severe performance problem. The trouble shooting approaches
> could be significantly different.
>
> Other things to consider are, what's CPU usage look like on the box?
> Is it pegged at 100%? Is it practically at 0%? This will reveal
> something about the nature of the lockup, i.e., is the system
> spinning? Or sitting blocked on a lock somewhere?
>
> You're probably going to need to open an SR and work w/ Oracle Support

> on this. I my estimation, diagnosing a lockup without Oracle support
> is going to be tough.
>
>
> Hope that helps,
>
> -Mark
>
> --
> Mark J. Bobak
> Senior Database Administrator, System & Product Technologies ProQuest
> 789 E. Eisenhower, Parkway, P.O. Box 1346 Ann Arbor MI 48106-1346
> +1.734.997.4059 or +1.800.521.0600 x 4059
> mark.bobak_at_proquest.com
> www.proquest.com
> www.csa.com
>
> ProQuest...Start here.
>
>
> -----Original Message-----
> From: oracle-l-bounce_at_freelists.org
> [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Howard Latham
> Sent: Thursday, December 11, 2008 9:10 AM
> To: ORACLE-L
> Subject: Max Logging
>
> Database Version 11G
> Anyone know of a parameter to set maximum activity logging to a trace
> level at instance
> level- We are trying to diagnose a lockup on a Redhat box.
>
> --
> Howard A. Latham
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

--
Howard A. Latham
--
http://www.freelists.org/webpage/oracle-l
Received on Fri Dec 12 2008 - 08:29:32 CST

Original text of this message