On Thu, May 16, 2019 at 6:32 PM Lucie Leistnerova <lleistne@redhat.com> wrote:

Hi Rik,

On 5/14/19 2:21 PM, Rik Theys wrote:

Hi,

It seems VM pools are completely broken since our upgrade to 4.3. Is anybody else also experiencing this issue?

I've tried to reproduce this issue. And I can use pool VMs as expected, no problem. I've tested clean install and also upgrade from 4.2.8.7.
Version: ovirt-engine-4.3.3.7-0.1.el7.noarch with ovirt-web-ui-1.5.2-1.el7ev.noarch

Only a single instance from a pool can be used. Afterwards the pool becomes unusable due to a lock not being released. Once ovirt-engine is restarted, another (single) VM from a pool can be used.

What users are running the VMs? What are the permissions?
Each VM is running by other user? Were already some VMs running before the upgrade?
Please provide exact steps.


Hi, just an idea... could it be related in any way with disks always created as preallocated problems reported by users using gluster as backend storage?
What kind of storage domains are you using Rik?

Gianluca