Re: [Users] Timezone Hypervisor/VM
On Fri, Feb 07, 2014 at 08:01:11AM +0000, Markus Stockhausen wrote:
> Von: Dan Kenigsberg [danken(a)redhat.com]
> Gesendet: Donnerstag, 6. Februar 2014 19:43
> An: Bob; Martin Polednik
> Cc: Markus Stockhausen; ovirt-users
> Betreff: Re: [Users] Timezone Hypervisor/VM
>
> On Thu, Feb 06, 2014 at 12:46:38PM -0500, Bob wrote:
> > I have the same issue. I have assumed it was a Windows configuration
> > issue. I have relied on sync'ing to Internet Time upon each bootup.
> > It would be nice to find a solution to this. My Windows VM
> > consistently comes up exactly 5 hours off, and although NTP is
> > configured the time is never corrected until I manually sync to
> > Internet Time. It certainly appears to be using the wrong TZ (GMT).
>
> I suppose you are seeing
>
> https://bugzilla.redhat.com/show_bug.cgi?id=956741 When RHEL VMs are
> powered off/on time is off by as much as 3 hrs when system comes
> back up
>
> which http://gerrit.ovirt.org/14750 should fix. Which ovirt version are
> you using? Would you agree to verify the patch once Martin backports it?
>
> Dan.
I implemented the above patch on one of our hypervisor nodes. But this
does not change anything. The VM has always the time difference after reboot
that is set in table vm_dynamic.
That patch is not intended to change this fact. What it should do is
report a correct timeOffset once your guest OS syncs its time.
We are on GMT+1. if utc_diff = 0 Windows VM time is on hour back. If
utc_diff = 3600 Windows VM time matches.
Should I file a new BZ?
I'd prefer to make sure we have all the facts first. Would you boot your
VM, and report the vmCreate line from vdsm.log?
The, fix the timezone withing the guest, and report the rtc events from
the log, and the timeOffset reported by `vdsClient -s 0 getAllVmStats`.
Stop the VM and check if Engine's utc_diff has changed.
A detailed BZ report is worth its weight in gold ;-)
Dan.