Hi and thanks for the reply.

Not really what we're after.  We want the stateless VMs but we want to be able to work on the template (patch, install addtional software, etc) and have those changes flow through to the existing VMs based on these templates.  In addition, when the VMs are refreshed based on the updated template, we need the MAC addresses to persist (as we define all our leases in DHCP).


On Sun, Jan 31, 2016 at 5:27 AM, Sigbjorn Lie <sigbjorn@nixtra.com> wrote:

We've had the same experience as you with vm pools and have moved away from vm pools to non-pool vms and manual assignment for Windows vms. We have an open ticket with RH support for enabling stateful vm pool. This is the bugzilla entry if you’d like to add your interest for the stateful vm pools feature: https://bugzillaredhat.com/show_bug.cgi?id=1234394


On 29 Jan 2016, at 13:20, Colin Coe <colin.coe@gmail.com> wrote:

Hi all

We're using RHEV 3.5.7 and have about 40 desktop VMs that are all clones from a single original.

In the very beginning we used desktop pools but this was torn down when we started getting "The trust relationship between this workstation and the primary domain failed" error messages.

We use DHCP for all desktops however every VM is defined on the DHCP server and unknown MACs are rejected.

I'd like to revisit desktop pools but I'd like to know if other are using this and what their experiences have been.

Kind of thing I'd like to know is
- how is patching done
- what is the performance like
- what tuning if any is done on the desktops
- GPO vs LPO

Any comments?


Users mailing list