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

Re: Corrupt LDAP DB ...

Greetings ...

Thank you to everybody, whom has put there 2c worth in ... Distro discussions aside, I was hoping I put a little more in here ...

>>> Started and stopped correctly, do you mean running slapd with
>>> options to
>>> start and then kill -9 slapd ... Is this incorrect or is there a
>>> better way to shutdown slapd?
>> Using kill -9 on slapd is the same as causing an abnormal shutdown.
>> If you
>> are using kill -9, then you are likely the cause of your data
>> corruption.
> More generally: You should *never* use -9 (SIGKILL) on *any* process
> unless all other measures have failed. SIGKILL causes the process to
> be terminated without it being given the chance to exit cleanly.
> Particularly for anything with persistent data store, such as a
> directory service (LDAP) or database server, killing the server daemon
> with SIGKILL will likely result in corrupted data.
Okay, but then what is the correct signal to give slapd to shutdown nicely? This could be half the problem ...

Just the reason why I think this might be my problem, I have two LDAP dn DB's going ( dought that is the correct term, but I'm hoping that everybody understands )

The default that ships with the packages and then the LDAP DB that I use with an include, because it's easier to diff the two conf files to see what might have changed between release, than to do by hand and eye.

Now, every time I have had DB corruption, I have had it in the Default DB, which I don't write too, so I don't understand, how that could get corrupted ... Maybe one of the OP Guru's could let me in on the secrets ;-)

	I hope this helps shed a little light on this problem.