[Linux-cluster] How to monitor secondary link?

[Rafael Micó Miranda]

Hi Jakov

El mar, 13-10-2009 a las 14:42 +0200, Jakov Sosic escribió:
> On Tue, 13 Oct 2009 11:37:23 +0200
> Jakov Sosic <jakov.sosic at srce.hr> wrote:
> 
> > I don't like this because it's a "hack" and not a real solution :-/
> > But it seems to be the best choice for the moment... Problem is I also
> > didn't test this one, and this might resolve in IP relocating, and
> > Xen's staying in place :-/
> 
> Currently I've solved it with qdisk heuristics, but I wonder, how would
> you do this if you don't have a qdisk? Quorumd without label
> specified, that only leans on heuristics?
> 
> 

I cant' check it now but i think filesystem resources have a "monitor"
or "status" method, in which they try accessing the filesystem and would
return error code if they can't access the filesystems.

So, if your secondary link fails and it is used to access your iSCSI NAS
and you have filesystems mounted from that NAS you should be able to
detect it and the cluster should fail the service. I imagine the problem
here is the Xen Service, in which (correct me if i'm wrong) you do not
use any kind of filesystem (you use a RAW device).

I'm very interested in this situation because i think in a short time i
would be deploying a cluster in a similar situation, but without Xen so
i hope i will get the benefits of this "monitor" method (if it exists
and works). Please, keep me in touch.

Cheers,

Rafael


-- 
Rafael Micó Miranda