[Xen-users] Re: In the year 2038! Is Xen messing with my clock?

[Michael Brancato <mike@xxxxxxxxxxxxxxxx>]

This is an update on my problem.  Still trying to solve this after 
almost a month.
On 1/29/09 8:29 PM, Michael Brancato wrote:
> I've started experiencing some weird issues with my Xen machine at home.
> It has been very stable until a few weeks ago, but it only ran an
> OpenSolaris PV system other than the Ubuntu Dom0. The build of
> OpenSolaris was relatively old, nv79.
>
> So crazy me did two things not too far apart from one another. First was
> I manually P2V'd the only Windows box in the house in an attempt to
> reduce power consumption for a computer that was always on but rarely
> used. That went smooth. That HVM now runs 24/7 in case my wife needs it.
In an attempt to isolate the problem, I have kept this domU offline 
since late January.
> The second thing I did was an in-place upgrade of my OpenSolaris box
> which has my ZFS pool on it. I also switched from Samba to the Solaris
> CIFS Service, but I doubt that has anything to do with this.
This is still running, and is definitely not causing the problem


> I should have kept better records, but that doesn't matter now. At some
> point, the Dom0 became unresponsive over the network. A trip to the
> basement showed the computer unresponsive to input. I powered off and
> rebooted, it worked. A week or two later, same thing. Then last night,
> it did it again. I gave up and went to bed and when I woke up, I noticed
> I had new mail on my phone (mail is received on my Dom0). I ssh'd in and
> it was up. Only, it was June 9th, 2038 with an uptime of 19499 days(I
> took a screenshot for bragging purposes =p). This itself was wrong
> because it would have been the year 2062 by those numbers. The output of
> 'xm list' showed my VM's running but none were reachable. This lasted
> for only an hour or so before it again was unresponsive and never came
> back.
>
> I can't find any logs showing what caused it, but here are some
> interesting things. If you have any idea why this may be happening,
> please respond. I'm just hoping its not hardware.
>
> Jan 23 00:10:02 fry kernel: [305645.199106] clocksource/0: Time went
> backwards: ret=11707a938ea09 delta=-31783517 shadow=117078e4777f3
> offset=1cd6c78e
> Jan 26 23:14:36 fry kernel: [646712.625002] clocksource/1: Time went
> backwards: ret=24e581b176a1e delta=-11315340 shadow=24e57e320d5b6
> offset=38a355be
So here is the gist... with HVM domains, my clock blows up after a week. 
 Without them, it is stable.  Does anybody have any idea what I could 
do to further isolate the problem?
I am running Xen 3.3.0 on Ubuntu.

Regards,

--
Mike

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users