[ovirt-users] Single node hyper converged deployment best practices.
Hello all,
On Mon, May 11, 2020 at 7:10 PM Sandro Bonazzola
<sbonazzo(a)redhat.com> wrote:
Can you please detail what you're trying to achieve? sort of
hyperconverged
deployment based on NFS instead of GlusterFS?
If so, no, it's not officially supported.
Long story short: I've got a number of oVirt setups, a couple of "big"
GlusterFS based multi-node clusters (used for production) and a couple
of small single node ones (via localhost NFS) that are used to backup,
staging, testing, etc.
As I'm planning a (slow...) upgrade path to 4.4, what is the supported
/ best practice method to install oVirt on a single node? single-node
GlusterFS, localhost NFS or local file system (which I never tried for
no good reason)? Which tends to perform better?
I must admit that even with 20-30 active VMs (and a lot of IO), thus
far, localhost NFS served me (very) well.
BTW, per oVirt documentation [1], these no much information about the
preferred method.
[1] https://www.ovirt.org/documentation/admin-guide/chap-Storage.html
Thanks,
Gilboa