Bug#599161: xen-linux-system-2.6.32-5-xen-amd64: Clock moved forward 50 minutes, caused Xen HVM domU restart

[Olivier Hanesse <olivier.hanesse@gmail.com>]

Hello,

I am got this issue on several (10+) Xen 4.0 servers without HVM: 

linux-image-2.6.32-bpo.5-xen-amd64          2.6.32-29~bpo50+1          Linux 2.6.32 for 64-bit PCs, Xen dom0 suppor

xen-hypervisor-4.0-amd64                    4.0.1-1                    The Xen Hypervisor on AMD64

Each times there is this strange "50min" : 

Nov  4 13:33:55  kernel: [591694.109052] Clocksource tsc unstable (delta = -2999660340870 ns)

Nov 16 19:04:27 kernel: [102888.814677] Clocksource tsc unstable (delta = -2999660352101 ns)

Nov 28 08:16:52 kernel: [2671840.236281] Clocksource tsc unstable (delta = -2999660333313 ns)

Nov 29 08:59:54 kernel: [171581.195202] Clocksource tsc unstable (delta = -2999660341178 ns)

Dec  8 12:58:09  kernel: [3108143.298526] Clocksource tsc unstable (delta = -2999660353020 ns)

Dec 27 02:34:16 kernel: [1012551.748589] Clocksource tsc unstable (delta = -2999660334211 ns)

Jan 12 09:12:53  kernel: [6537645.820016] Clocksource tsc unstable (delta = -2999660339286 ns)

Jan 28 11:04:54  kernel: [5352834.035048] Clocksource tsc unstable (delta = -2999660330184 ns)

Feb  9 21:04:03  kernel: [6415408.244988] Clocksource tsc unstable (delta = -2999660342333 ns)

Feb 10 04:19:24 kernel: [4306193.416722] Clocksource tsc unstable (delta = -2999660332510 ns)

after that ntp in Dom0 fails with : ntpd[6834]: time correction of -3000 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time

Same things with ntp in domU.

I didn't have this kind of issue with Xen Lenny.

Regards

Olivier