[ovirt-users] Ovirt and Gluster

[Jorick Astrego]

On 07/29/2015 02:24 PM, Joop wrote:
> On 29-7-2015 13:59, Jorick Astrego wrote:
>> For RHEV I don't have an answer.
>>
>> But for anyone that cares, we've been running oVirt 3.5 and GlusterFS
>> 3.6 for about three months in production now. We run a Gluster
>> cluster and a seperate virtualization hosts cluster. The Gluster
>> cluster is installed, configured and managed seperately from oVirt as
>> there is no way to have a seperate storage network in oVirt 3.5
>>
> I run something comparable but using a split dns to get a separate
> storage network but also a storage cluster and virt cluster. The
> storage is managed through ovirt though because of the split-dns.
>
>> There have been no specific incompatibility issues, so we are very
>> happy. But as we don't use every feature mileage may vary.
> Question: do you use fuse mounts or libgfapi?
>
> I use libgfapi in test, using special vdsm packages and need to test
> if rebooting gluster nodes has an impact on running VMs. Select
> GlusterFS storage domain in oVirt isn't guaranteed  to give you libgfapi!
>
> Joop
>
It's production, so I'm using regular FUSE mounts. Although slower, it
works at least with "supported" packages.

We did spend some extra on hardware and are conservatively staying close
to the Red Hat Storage hardware requirements. So that means RAID10 for
now and a seperate SSD stroge volume.

For our enterprise customers we are waiting for oVirt 3.6 and will spend
even more on hardware (RDMA, Datacenter bridging, PCIe SSD's, etc) but
we have to do a lot more testing before rolling it out.







Met vriendelijke groet, With kind regards,

Jorick Astrego

Netbulae Virtualization Experts 

----------------

	Tel: 053 20 30 270 	info at netbulae.eu 	Staalsteden 4-3A 	KvK 08198180
 	Fax: 053 20 30 271 	www.netbulae.eu 	7547 TA Enschede 	BTW NL821234584B01

----------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20150729/5cbc2ff9/attachment-0001.html>