Tugger the SLUGger!SLUG Mailing List Archives

Re: [SLUG] Virtualisation and DRBD

While the setup you describe is possible there are a number of drawbacks:

- DRBD disks are very slow, depending on the protocol you set them to use
but still there is a performance penalty. If you DRBD the entire image of a
running xen guest, your going to suffer performance hits for everything the
OS does - not just the HA important data

- Live migration is fantastic but you lose the ability to independently
upgrade either side of the cluster. This makes rollback harder and more

- Likewise, you kinda put all your HA eggs into one basket since this one
DRBD setup might be the basis for a number of clusters. Can create some
complex failure scenarios, and be hard to maintain e.g. should you wish to
resize the DRBD disk maybe

- Finally you now make a decision and describe the VM's themselves as
cluster resources, or you have clusters on-top of clusters and have a dom0
cluster and domU cluster.

For these reasons I would lean towards running DRBD inside each VM, with a
minimum amount of shared state of the disk


On Wed, Aug 25, 2010 at 12:46 PM, Nigel Allen <dna@xxxxxxxxxxx> wrote:

> Hi All
> We're investigating both virtualisation of servers and High Availability
> at the same time.
> Currently looking at Linux-HA and DRBD (amongst others).
> The idea of DRBD appeals to both me and the client as it means (or seems
> to at least) that we could add a third (off-site) machine into the
> equation for "real" DR.
> What happens when we then introduce Virtualisation into the equation
> (currently have 4 x servers running Centos & Windoze Server - looking at
> virtualising them onto one single box running Centos-5).
> I suppose the (first) question is: If we run 4 virtualised servers (B,
> C, D, and E) on our working server "A" (complete with it's own storage),
> can we also use DRBD to sync the entire box and dice onto server A1
> (containing servers B1, C1, D1, and E1) or do we have to sync them all
> separately? Will this idea even float? Can we achieve seamless failover
> with this. If not, how would you do it
> Any input (as ever) gratefully accepted.
> Confused at the Console
> Nigel.
> --
> SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
> Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html