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

Re: Slapd database goes corrupt repeatedly after recovery



--On Thursday, March 25, 2010 4:06 PM +0100 Kaspar Fischer <kaspar.fischer@baselgovernance.org> wrote:

On 25.03.2010, at 14:27, Dieter Kluenter wrote:

Kaspar Fischer <kaspar.fischer@....org> writes:

Dear list,

We are using an OpenLDAP/slapd server to manage the user accounts of
our Samba server and have recently run into the problem that users
cannot connect to Samba drives anymore after some time. Samba
complains that it cannot connect to the LDAP server (see below for
error message in Samba log) and the slapd log shows
[...]

 Mar 25 11:38:15 office-server slapd[3433]: bdb(dc=foo,dc=org): file
id2entry.bdb has LSN 1/382892, past end of log at 1/283666
[...]
Strangely, restarting slapd helps and users can use Samba again for a
limited and arbitrary period of time until the problem pops up
again. I tried fixing the database using

 db4.7_recover -v -h /var/lib/ldap

but again, the problem pops up again later.
[...]

For some unknown reason the last transaction logfile has been removed,
therefor the database connot be recovered. Unfortunately you have to
rebuild the database

-Dieter

Dieter,

First of all, thanks for your prompt and helpful reply!

I have just rebuild the database and will check now whether that did it.

For the sake of completeness (for others with similar problems), here's
how I rebuilt the db:

  /etc/init.d/slapd stop
  slapcat -l slapd-export.ldif -f /etc/ldap/slapd.conf -b "dc=foo,dc=org"
  rm -fr /var/lib/ldap/ # make sure this IS the path to your db as
specified in slapd.conf   mkdir /var/lib/ldap/
  chgrp -R openldap /var/lib/ldap/; chown -R openldap /var/lib/ldap/

These next two steps are wrong.  slapadd will create the database.

  /etc/init.d/slapd start # creates an empty db
  /etc/init.d/slapd stop


--Quanah

--

Quanah Gibson-Mount
Principal Software Engineer
Zimbra, Inc
--------------------
Zimbra ::  the leader in open source messaging and collaboration