On Fri, Mar 18, 2016 at 5:27 PM, Clint Boggio <clint@theboggios.com> wrote:

Greetings all, and Happy Friday;

I'm running oVirt 3.6 management engine that is currently overseeing 4
compute nodes, attached to iSCSI storage over Infiniband. The compute
nodes have 8 hot-swap hard drive slots, with each only having 2 slots
occupied in a hardware RAID array level one for the hypervisor to live
on.

Would it be possible for me to set up each compute node with a second
array in the remaining 6 hot-swap slots, and utilize that new space for
a nice fault tolerant GlusterFS storage array ?

That schema is called hyper-converged.

Please see:

https://fosdem.org/2016/schedule/event/virt_iaas_ovirt_hyperconverge/attachments/slides/1119/export/events/attachments/virt_iaas_ovirt_hyperconverge/slides/1119/presentation_fosdem.pdf

Take care that currently it's not production ready.

1. Would the system allow me to add the Gluster storage to the existing
datacenter/cluster that is currently using iSCSI ?

2. Would I be able to configure all aspects of the Gluster
environment(except the hardware RAID array)through the GUI ?

3. What hardware RAID level would be optimal for this configuration ?
(RHEV documentation says RAID 6 is "Mandatory")

4. Will the system support live migration between storage domains ?

I'm reading through the gluster documentation now to get a better
understanding of it's inner workings. Is there a good source for Gluster
on oVirt that i can reference as well ?

_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users