[ovirt-users] GlusterFS performance questions

[Jayme]

I am currently in the process of researching converting an existing SMB
infra to virtual.  Ovirt/RHEV is a strong  contender and checks off a lot
of boxes on our list.  GlusterFS is appealing but I am finding it very
difficult to find any answers or stats/numbers regarding how well it can
perform as backend storage for VMs, especially VMs with medium to high
workloads, and possibly some or many in a dev environment where many small
files are interacted with very frequently.

>From my research on GlusterFS in general some people tend to say that it is
not the best performer and it's not the best choice for handling small
files.  How true is this?

I understand that there are many variables such as hardware, drives, proper
configuration, internal network etc. but what I'm trying to avoid here is
going through the whole process of spendng a lot of money building out an
ovirt gluster with a proper SSD glusterFS configuration w/ 10gbe internal
network only to find out some day down the road that it cannot meet our I/O
needs.

Can anyone comment on how I can best determine the performance level to
expect from a typical ovrt w/ glusterfs replica 3 arbiter 1 configuration.
Or perhaps any use cases with people/companies running a similar setup in
production with high VM workloads?

Right now I'm leaning toward using a SAN due to the simplicity and
predictable performance.  Although this will definitely cost a lot more
money for something that glusterFS *may* be able to handle.

Thanks!

James
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20171016/7ec5b917/attachment.html>