[Pacemaker] Problem with one drbd dual primary ressource (fwd)

georg at riseup.net georg at riseup.net
Mon Aug 12 09:41:58 UTC 2013


Sorry, didn't send this to the list...

----- Forwarded message from "georg at riseup.net" <georg at riseup.net> -----

Date: Sun, 11 Aug 2013 17:13:25 +0200
From: "georg at riseup.net" <georg at riseup.net>
To: Digimer <lists at alteeve.ca>
Subject: Re: [Pacemaker] Problem with one drbd dual primary ressource
User-Agent: Mutt/1.5.21 (2010-09-15)

Hi,

On 13-08-10 20:48:08, Digimer wrote:
> There is a DRBD-users mailing list, but I suspect there is a lot of
> overlap here.

Yeah, I tought so too...

> Wrong assumption (though a common one). The idea of a test cluster
> is to get things working. Fencing is a _critical_ part of this. When
> I started clustering, I made the exact same assumption. I wasted
> weeks trying to get even basic stuff working.
> 
> Only when I went back and added fencing did everything start to
> work. Please take the time to set it up now!

So I did. But, nothing changed. For this scenario there shouldn't be
fencing involved at all, right? I just issued 
crm ressource migrate nfs ha1, which led to the same failure as the last
days.

However, in the logs now I got:
[  100.995696] block drbd4: Split-Brain detected, 0 primaries, automatically solved. Sync from this node

Any ideas?

Cheers,
Georg

----- End forwarded message -----




More information about the Pacemaker mailing list