[DRBD-user] Digest integrity check FAILED

[Raoul Bhatia [IPAX]]

On 06/24/2011 07:36 PM, Thiago Vinhas wrote:
> Hi Lars, thanks for your response.
> 
> What I'm trying to achieve is a load balanced MySQL cluster, where my
> application could write to both MYSQL servers, under DRBD.
> I'm actually using OCFS2 as the filesystem. But after several hours my
> nodes are always getting disconnected due to the failure I pointed out
> in the last message.
>
> The reason I'm running DRBD in dual primary mode, is that I expect to
> have MySQL writing on both nodes. Am I missing something?

this isn't going to work. mysql must be run in a master-master setup
itself, using either replication or nbd.

putting mysql on a shared storage will surely mess things up, as
concurrent access to the same database files doesn't work (at least to
my knowledge).

in plain english: using this setup will corrupt all your data.

> What benefit would I have by running a Primary/Secondary configuration,
> instead of a Primary/Primary?

no data corruption and "clean" failover - at least as clean as it can
be when one node dies and not all data has been written to disk.

honestly, i suggest not using drbd for this scenario but using
pacemaker and a master-slave or master-master with the latest and
greatest cluster-agents.

details can be found, for example, at [1]

cheers,
raoul
[1]
http://fghaas.wordpress.com/2010/03/30/mysql-scale-out-with-replication-and-pacemaker/
-- 
____________________________________________________________________
DI (FH) Raoul Bhatia M.Sc.          email.          r.bhatia at ipax.at
Technischer Leiter

IPAX - Aloy Bhatia Hava OG          web.          http://www.ipax.at
Barawitzkagasse 10/2/2/11           email.            office at ipax.at
1190 Wien                           tel.               +43 1 3670030
FN 277995t HG Wien                  fax.            +43 1 3670030 15
____________________________________________________________________