RHS 2.0 and RHEV 3.2 .... anyone using this in the real world?

Latest response

We currently run RHS as a scaleout filer, and as a totaly seperate system, run RHEV on our old, old, old HP EVA (fc san).

I'd like to reconfigure my RHS system to serve RHEV VMs. We tried this a while back and, while the performance was pretty good, the vm locking issue killed us.

I'm wondering if anyone is now with the new versions of both, running this config in the real world yet, are you using the POSIX filesystem to mount the gluster volume?

rhev-h or rhel hosts?

Does anyone know if you can do geo-replication of a VM store? (for cold site DR, not hot failover)

Cheers,

Jonathan Collingridge

Responses

Hi Johnathan,

Are you running RHS 2.0 Update 5 per chance[1]? A lot of fixes went into that async release, but more feature enhancements are on the way with the next release of RHS coming soon.

I would highly recommend you work with your sales team (including your Solution Architect) to better evaluate your use case and perform an Architecture Review. There are certain situations where RHS can do something, but in some cases Red Hat doesn't recommend certain workloads.

Hope this helps,

Andrius.

[1] http://rhn.redhat.com/errata/RHBA-2013-1064.html

Hi Andrius,
We haven't updated to Update 5 yet.
I plan on speaking with the sales guys, before we implement anything, but i just wanted to see if rhs and rhev have finally matured to a point where this can be done.

Clearly the roadmaps of both products have been converging for some time. But whenever I talk to someone at RH about this, we seem to get a guarded response on whats possible (or "supportable")...

So i was hoping that someone might've done this in production, and also might be hanging around on this forum, what are the chances eh?

Cheers,
Jono

Hi Jonathan,

I've did a bit more checking, and you should be fine running RHEV 3.2 + RHS 2.0 Update 5 with the use case you described. We are interested in the "locking issue" you describe above - we're not quite sure what you are seeing and would like to understand more of this, maybe more details in a support case.

Regardless, we currently do support RHS run as a guest on a RHEV host and can also be used as backing store for VMs running on the same host.

Regards,

Andrius.

Hi Andrius,
Thanks for that confirmation.
The locking issue i was referring to was the lack of granular locking for files which caused issues, for rhev and the vms running on it. This was a bleeding edge implementation back then though, we were using the gluster storage appliance (pre RH purchase of gluster) with RHEV 2.x...

Thanks again,
Jonathan

Jonathan,

Wow, great to hear about those that have been there since the beginning! Hard to believe upstream gluster is at 3.4 already! Crazy!

Thanks again, and keep the questions coming! :-)

Andrius.

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.