[Date Prev][Date Next] [Chronological] [Thread] [Top]
RE: Problems with openldap-2.1.22 slowing down over time.

To: "'Matthew Schumacher'" <matt.s@aptalaska.net>, <openldap-software@OpenLDAP.org>
Subject: RE: Problems with openldap-2.1.22 slowing down over time.
From: "Matthew Hardin" <mhardin@symas.com>
Date: Wed, 12 Nov 2003 16:47:49 -0800
Importance: Normal
In-reply-to: <3FB2B042.1090608@aptalaska.net>
Hi Matt,

>From the behavior you are describing I'd guess you are using back-bdb and an
improperly configured or missing DB_CONFIG file.

Would you please indicate which backend you are using (i.e, back-bdb,
back-ldbm, etc.), and if you are using back-bdb, please tell us what you've
put in the DB_CONFIG file? If you have not set up a DB_CONFIG file, check
the docs on Sleepycat's web site that talk about setting up a DB_CONFIG
file, specifically selecting cache sizes. There are a few other things that
go in that file, and you can check the OpenLDAP Faq-o-Matic for them. Try
this link:

http://www.openldap.org/faq/index.cgi?_highlightWords=tuning&file=190

Hope this helps.

Matthew Hardin
Symas Corporation
Packaged, certified, and supported LDAP software:
http://www.symas.net/download


> -----Original Message-----
> From: owner-openldap-software@OpenLDAP.org [mailto:owner-openldap-
> software@OpenLDAP.org] On Behalf Of Matthew Schumacher
> Sent: Wednesday, November 12, 2003 2:12 PM
> To: openldap-software@OpenLDAP.org
> Subject: Problems with openldap-2.1.22 slowing down over time.
> 
> List,
> 
> I have been fighting with slow down issues with openldap.  I wrote a
> query that I run ever hour to monitor performance.  Over time the query
> gets to be very slow (47 secs at the beginning of the attached log) only
> using 2% of the cpu, then I tried to slapindex the database and it takes
> minutes per record to index so I ctl-c and it leaves the database
> completely corrupt.
> 
> If I export the database into an ldif file then delete the openldap-data
> dir and reload it, then the query goes back to 1sec for a while and
> eventually slows down again.
> 
> Anyone know what this could be, I'm not sure how to identify the problem.
> 
> This database is only updated though a php interface I wrote, but it is
> queried by sendmail, nss_ldap, pam_ldap, php, and perl.
> 
> Thanks,
> schu
> 
> Logs:
> 
> Sun Nov  9 14:07:01 AKST 2003: 1.01user 0.05system 0:47.65elapsed 2%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 15:07:01 AKST 2003: 1.00user 0.05system 0:46.57elapsed 2%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 16:07:01 AKST 2003: 0.97user 0.06system 0:49.70elapsed 2%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 17:07:01 AKST 2003: 1.06user 0.04system 0:44.72elapsed 2%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 18:07:01 AKST 2003: 1.04user 0.01system 0:45.85elapsed 2%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 19:07:01 AKST 2003: 1.02user 0.06system 0:45.59elapsed 2%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 20:04:12 AKST 2003: 1.02user 0.05system 0:01.10elapsed 96%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 20:07:01 AKST 2003: 1.09user 0.07system 0:01.55elapsed 74%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 21:07:01 AKST 2003: 1.14user 0.04system 0:01.68elapsed 70%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 22:07:01 AKST 2003: 1.03user 0.07system 0:01.51elapsed 72%CPU
> (0avgtext+0avgdata 0maxresident)k
> Sun Nov  9 23:07:01 AKST 2003: 0.88user 0.07system 0:01.50elapsed 63%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 00:07:01 AKST 2003: 1.11user 0.05system 0:01.55elapsed 74%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 01:07:01 AKST 2003: 0.94user 0.06system 0:01.50elapsed 66%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 02:07:01 AKST 2003: 1.11user 0.02system 0:01.56elapsed 72%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 03:07:01 AKST 2003: 1.06user 0.05system 0:01.53elapsed 72%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 04:07:01 AKST 2003: 1.05user 0.05system 0:01.54elapsed 71%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 05:07:01 AKST 2003: 1.03user 0.08system 0:01.56elapsed 70%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 06:07:01 AKST 2003: 1.03user 0.08system 0:01.66elapsed 66%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 07:07:01 AKST 2003: 1.02user 0.05system 0:01.67elapsed 63%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 08:07:01 AKST 2003: 0.98user 0.09system 0:01.58elapsed 67%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 09:07:01 AKST 2003: 1.07user 0.09system 0:02.55elapsed 45%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 10:07:01 AKST 2003: 1.06user 0.03system 0:01.56elapsed 69%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 11:07:01 AKST 2003: 0.93user 0.08system 0:02.56elapsed 39%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 12:07:02 AKST 2003: 0.90user 0.04system 0:02.62elapsed 35%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 13:07:01 AKST 2003: 1.03user 0.08system 0:03.49elapsed 31%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 14:07:01 AKST 2003: 1.03user 0.03system 0:03.58elapsed 29%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 15:07:01 AKST 2003: 1.05user 0.02system 0:02.52elapsed 42%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 16:07:01 AKST 2003: 1.05user 0.07system 0:02.53elapsed 44%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 17:07:01 AKST 2003: 0.97user 0.07system 0:03.57elapsed 29%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 18:07:01 AKST 2003: 1.02user 0.04system 0:02.52elapsed 42%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 19:07:01 AKST 2003: 0.99user 0.10system 0:03.55elapsed 30%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 20:07:01 AKST 2003: 0.94user 0.06system 0:02.57elapsed 38%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 21:07:02 AKST 2003: 0.92user 0.08system 0:03.54elapsed 28%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 22:07:01 AKST 2003: 0.96user 0.03system 0:03.60elapsed 27%CPU
> (0avgtext+0avgdata 0maxresident)k
> Mon Nov 10 23:07:01 AKST 2003: 1.07user 0.06system 0:04.55elapsed 24%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 00:07:01 AKST 2003: 1.04user 0.04system 0:03.58elapsed 30%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 01:07:01 AKST 2003: 1.00user 0.03system 0:03.69elapsed 27%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 02:07:01 AKST 2003: 1.02user 0.04system 0:03.70elapsed 28%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 03:07:01 AKST 2003: 0.91user 0.11system 0:04.59elapsed 22%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 04:07:01 AKST 2003: 1.07user 0.03system 0:03.53elapsed 31%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 05:07:01 AKST 2003: 1.04user 0.02system 0:04.59elapsed 23%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 06:07:01 AKST 2003: 0.99user 0.08system 0:03.56elapsed 30%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 07:07:01 AKST 2003: 1.02user 0.08system 0:04.57elapsed 24%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 08:07:01 AKST 2003: 0.96user 0.04system 0:05.56elapsed 17%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 09:07:01 AKST 2003: 1.01user 0.04system 0:03.57elapsed 29%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 10:07:02 AKST 2003: 1.05user 0.07system 0:05.62elapsed 19%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 11:07:01 AKST 2003: 1.01user 0.08system 0:07.73elapsed 14%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 12:07:01 AKST 2003: 0.97user 0.06system 0:06.52elapsed 15%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 13:07:01 AKST 2003: 1.04user 0.04system 0:09.49elapsed 11%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 14:07:01 AKST 2003: 1.04user 0.07system 0:09.88elapsed 11%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 15:07:01 AKST 2003: 1.02user 0.04system 0:06.57elapsed 16%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 16:07:01 AKST 2003: 0.93user 0.08system 0:06.57elapsed 15%CPU
> (0avgtext+0avgdata 0maxresident)k
> Tue Nov 11 17:07:01 AKST 2003: 1.11user 0.07system 0:07.57elapsed 15%CPU
> (0avgtext+0avgdata 0maxresident)k
References:
- Problems with openldap-2.1.22 slowing down over time.
  - From: Matthew Schumacher <matt.s@aptalaska.net>
Prev by Date: Re: Problems with openldap-2.1.22 slowing down over time.
Next by Date: Re: Intro and question
Index(es):
- Chronological
- Thread