[Pacemaker] 2 web and 2 db servers both with drbd. Questions concerning the set-up

[martin.braun at icw.de]

Hi Oliver,

> > I almost have a set-up which looks like this:
> > 
> > 
> >   Master side        Slave side
> > 
> >   ---------virtual IP----------
> >   |                           |
> > web1----drbd/pacemaker------web2
> >   |                           |
> >   internal                 internal
> >   |                           |
> >   |                           |
> > db1----drbd/pacemaker-------db2

I am not sure if I understand your setup: web1 for apache, web2 for 
tomcat, db1 for mysql db2 for pgsql? 

> I have an idea. I don't want automatic fallback btw. So, why not a
> simple monit, which does a crm standby in case of a failure of the other
> master/slave pair?

I would recommend to go with constraints:

1. location prefer settings like apache on web1 score 200 and apache on 
web2 score 100  and vice versa for tomcat and again the same with the 
dbms.
2.group the resource accordingly.

HTH,
Martin
 


Oliver Hoffmann <oh at dom.de> wrote on 04.05.2010 16:00:52:

> [image removed] 
> 
> Re: [Pacemaker] 2 web and 2 db servers both with drbd. Questions 
> concerning the set-up
> 
> Oliver Hoffmann 
> 
> to:
> 
> pacemaker
> 
> 04.05.2010 16:02
> 
> Please respond to The Pacemaker cluster resource manager 
> 
> Am Tue, 4 May 2010 15:02:54 +0200
> > Hi list,
> > 
> > I almost have a set-up which looks like this:
> > 
> > 
> >   Master side        Slave side
> > 
> >   ---------virtual IP----------
> >   |                           |
> > web1----drbd/pacemaker------web2
> >   |                           |
> >   internal                 internal
> >   |                           |
> >   |                           |
> > db1----drbd/pacemaker-------db2
> > 
> > 
> > The web (apache + tomcat) and db (mysql + postgresql) servers have
> > their data on a drbd device each. The internal connection is for
> > better performance of the db queries and redundancy. 
> > 
> > How do I ensure that web2 AND db2 take the resources if web1 and/or
> > db1 fail?
> > 
> > Thanks for input!
> > 
> > Regards,
> > 
> > Oliver
> > 
> > _______________________________________________
> > Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> > http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> > 
> > Project Home: http://www.clusterlabs.org
> > Getting started:
> > http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> > 
> 
> I have an idea. I don't want automatic fallback btw. So, why not a
> simple monit, which does a crm standby in case of a failure of the other
> master/slave pair?
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf



InterComponentWare AG:  
Vorstand: Peter Kirschbauer (Vors.), Jörg Stadler / Aufsichtsratsvors.: Prof. Dr. Christof Hettich  
Firmensitz: 69190 Walldorf, Altrottstraße 31 / AG Mannheim HRB 351761 / USt.-IdNr.: DE 198388516