[Pacemaker] Service failback issue with SLES11 and HAE 11

[Dejan Muhamedagic]

Hi,

On Thu, Jun 10, 2010 at 09:40:09AM +0800, ben180 wrote:
> Dear Dejan,
> 
> Thanks for your reply.
> 
> I reviewed the log of tibcodb2 and noticed that after step7.ServerB reboot,
> ServerB CAN heard from ServerA.
> 
> The strange thing is that when ServerA reboot, ServerA CANNOT heart from
> ServerB. But when ServerB reboot, ServerB CAN heard from ServerA.
> Why ServerA cannot recognize ServerB, but ServerB can recognize ServerA ?
> Is that means something such as a firewall or the network switch port
> restrict some kind of packet to send from ServerB ?

Don't know. Never heard of such a case. Some network equipment
doesn't deal well with multicasting.

> I'm curious about the way how suse cluster nodes communicate  with each
> other.

Multicast UDP.

> Can you point the way to shed me some light ?

No, sorry.

Thanks,

Dejan

> Thanks.
> 
> 
> 2010/6/9 Dejan Muhamedagic <dejanmm at fastmail.fm>
> 
> > Hi,
> >
> > On Wed, Jun 09, 2010 at 03:27:08PM +0800, ben180 wrote:
> > > Dear Dejan,
> > >
> > > The test sequence is:
> > >
> > > 1.Service is running on ServerA(tibcodb)
> > > 2.The Network Cable on ServerA is pulled out
> > > 3.ServerB(tibcodb2) fenced ServerA，ServerA reboot
> > > 4.ServerB take over the service
> > > 5.ServerA restart and the network is up
> > > 6.ServerA fence ServerB and take over the service
> > > 7.ServerB reboot
> > >
> > >
> > > My question is : after step5:ServerA restart and the network is up, it
> > > should not fence ServerB again but it did.
> > > What happened between step5 and step6 ?
> >
> > According to the logs, the tibcodb never heard from the other
> > node after reboot. That's strange. You sure you don't run
> > firewall? Or something else preventing packets reaching the node.
> >
> > Thanks,
> >
> > Dejan
> >
> > _______________________________________________
> > 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
> > Bugs:
> > http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
> >

> _______________________________________________
> 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
> Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker