[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Xen Failover - Best Pratice

On Sat, 2014-02-22 at 12:45 +0100, ml ml wrote:
> Hello List,
> i have two nodes with xen and drbd.
> Is there a easy way to make sure my xen instances are either running
> on node 1 or node 2?

Unless you are using a cluster-aware file system with dual-primary mode
DRBD, you are also going to need to make sure that the DRBD resource is
also primary on the node where you want to run a Xen image stored on it.
You probably did something like this in your old v1 configuration.

> corosync, pacemaker, etc seem to be a overkill.

But I think that's probably what you need. It is still possible to run
heartbeat in v1 mode like you used to, but it never did have monitoring
(meaning, it would make sure the resource was initially started
somewhere, but if it died, there wasn't any mechanism for detecting that
or restarting it). And hearbeat is no longer being developed.

>  Its just xml hell.

It has been quite a while since it has been necessary to work directly
with XML. There is now the "crm shell" (which uses a much less hideous
than XML, although still non-trivial, configuration language). There is
even a GUI now I believe, although I haven't personally tried it.

I use heartbeat/pacemaker with the crm shell to run Xen VMs. If I were
starting out fresh now, I would use corosync/pacemaker instead. The crm
shell configuration language is the same either way.

Xen also has its own high availability system called "Remus", but I have
not tried that either.


Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.