Ctdb Error

Martin Schwenke martin at meltin.net
Fri Dec 19 02:08:36 MST 2014


Hi Zinny,

On Thu, 18 Dec 2014 11:33:48 -0500, ezinne ibeagbazi
<eibeagbazi at gmail.com> wrote:

> I am using ctdb-2.5.1-1.el6.x86_64 and samba-3.6.23-12.el6.x86_64

That sounds like a combination that should work OK.  However, if
you're able to build CTDB 2.5.4 then it has a bunch of fixes in it.

If you can get that GDB backtrace then we might be able to help figure
out why smbd was crashing.

> This morning I started the machines and one node became healthy and trying
> to work around the second node to become healthy, the two nodes are now
> showing unhealthy again.

Same thing?  smbd crash?

> But what I don't understand is after a reboot of the two nodes, one of it
> lost all its configuration files including smb.conf and /gluster/lock/ctdb
> files yet the ctdb still starts well but still unhealthy.
> 
> Any deals on this?

That's very strange.  You're storing the configuration in /etc?  I
think some early documentation suggested storing configuration in the
cluster filesystem but that isn't generally recommended.  If you're
storing configuration in /etc then I can't imagine why it would
disappear...

peace & happiness,
martin


More information about the samba-technical mailing list