[BlueOnyx:01446] Re: db corrupt?

[Robert Fitzpatrick]

On Fri, 2009-06-19 at 02:44 +0200, Michael Stauber wrote:
> Hi WebTent,
> > [root at tpa-hosting ~]# /etc/init.d/dbrecover start
> > Checking consistency of user and group information... *** glibc detected
> > *** double free or corruption (top): 0x080d56d0 *** /etc/init.d/dbrecover:
> > line 49:  3704 Aborted                 $DBRECOVER -v -h $DBHOME >>$LOGFILE
> > 2>&1 failed
> >
> > I restarted the server and still getting this error, no users can login. Do
> > I need to start loading backups on another machine or can something be done
> > about this?
> 
> Try this:
> 
> 	/etc/init.d/ dbrecover stop
> 	/etc/init.d/ dbrecover start
> 
> If that doesn't fix it, make a copy of the directory /var/db/ like this:
> 
> cp -R /var/db /var/db.bak
> 
> And then run this command to delete the transaction logfile:
> 
> rm -f /var/db/log.*
> 
> Then try ...
> 
> 	/etc/init.d/ dbrecover stop
> 	/etc/init.d/ dbrecover start
> 
> ... again and it should work.
> 

Thanks Michael. I ended up restoring /var/db from previous nights backup
and that fixed it, will be doing some testing later.

Seems this one server is problematic for a few months now. More than
once resetting passwords or other general GUI functions have stopped
working and only start again after a reboot, then today with this
problem. We'll see how it goes tonight while adding some users and
changing passwords. Is dbrecover the best way to determine if any db
corruption lies within? And do you think I should put back my backup
of /var/db done today prior to the restoration and delete the logs as
you suggested? GUI has been locked with no changes since then.

Getting a good backup with your CMU 2.81 fix now, thanks for everything.

-- 
Robert