[Linux-cluster] ZEN with GFS

[Bernard Chew]

> -----Original Message-----
> From: linux-cluster-bounces at redhat.com [mailto:linux-cluster-
> bounces at redhat.com] On Behalf Of jr
> Sent: Tuesday, March 11, 2008 5:49 PM
> To: linux clustering
> Subject: RE: [Linux-cluster] ZEN with GFS
> 
> 
> > Hi,
> >
> > Just to share,
> >
> > In my production environment, I've 5 physical hosts connecting to a
> > central shared storage (MD3000i) using iscsi. 2 virtual disks were
> > created in the MD3000i; a 200GB virtual disk for 10 xen guest images
> > (Apache, MySQL, Tomcat, etc), and a 1TB virtual disk for data
storage.
> > Both virtual disks are formatted with GFS. The first virtual disk is
> > mounted in the dom0 cluster (containing all the physical hosts) and
2
> > xen guests are started in each host from the shared storage. The 1TB
> > virtual disk is then mounted after the domU cluster (containing all
> the
> > virtual guests) is up.
> >
> > The multipath works nicely for the 1TB data virtual disk, and the
> > virtual guests have no problems accessing the data storage in the
> event
> > of failover. The same cannot be say for the 200GB virtual disk that
> > contains the xen guest images because the virtual guests will freeze
> in
> > the event of failover.
> 
> you mean in the event of failover of the path or the dom0?
> 
> johannes
> 
> >
> > I'm still trying to access if I need the automated failover of
> > virtualized guests feature. If not, I'll probably keep 2 virtual
> guests
> > on each physical host instead.
> >
> > Regards,
> > Bernard Chew
> >
> >
> > --
> > Linux-cluster mailing list
> > Linux-cluster at redhat.com
> > https://www.redhat.com/mailman/listinfo/linux-cluster
> 
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster

Hi,

Sorry for the confusion here. I was saying the failover of the MPIO
works nicely for the 1TB virtual disk (domU cluster) but not the 200MB
virtual disk (dom0 cluster) which contains all the virtual guests.

As for the second part, I was saying such a setup allows the failover of
the domU machines as they are stored in the central virtual disk. The
failover here refers to using the RHEL cluster to control the guests. 

But as the MPIO failover does not work nicely for the dom0 cluster, I
may just keep the domU machines locally.

Regards,
Bernard