[Date Prev][Date Next] [Chronological] [Thread] [Top]

RE: [openldap 2.1.12 + BerkeleyDb 4.1.24] db_checkpoint pb



Hi Helene,

 

Place a checkpoint statement in your slapd.conf and remove the db_checkpoint command from your cron job. It’s better to let the application using the db (slapd) do its own checkpointing.

 

Continue to run db_archive from your cron job to identify unneeded log files. Either delete them or archive them to tape as desired.

 

You should also upgrade to the latest stable version of OpenLDAP. 2.1.12 is pretty old and a lot has been fixed since then. Remember, too that there’s one patch for Berkeley DB that needs to be applied.

 

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 Helene.Lastennet@alcatel.fr
Sent:
Wednesday, November 19, 2003 3:27 AM
To: openldap-software@OpenLDAP.org
Subject: [openldap 2.1.12 + BerkeleyDb 4.1.24] db_checkpoint pb
Importance: High

 

Hy,

I have some pb with the checkpointing procedure.
In my crontab I added a shell script (cf. file in attachment) enabling to make every hour:
1. db_archive (in order to remove the inactive database log files)
2. db_checkpoint (to write into the backing database file the changes to the database that are part of committed transactions)
3. db_archive (in order to remove again the inactive database log files)

But, often my script can be blocked (db_checkpoint seems failed) and I can't restart correctly openldap
To restart correctly my openldap server I need to make a db_recover
But our customer is absolutely not glad to stop the service regularly ...

So do you have an idea about this pb ?
Best regards,
Helene

Working Environment :
Linux redhat 7.3
Berkley 4.1.24
openldap 2.1.12
database-backend: bdb