Re: [Xen-devel] Possible hint for "Clocksource tsc unstable" problem

[Andreas Kinzler <ml-xen-devel@xxxxxx>]

On 01.12.2011 08:49, Olivier Hanesse wrote:
> good news !
> Is your "test system" the same hardware as before ?
> For example, I don't have time issue with processor having the
> 'nonstop_tsc' flag.
> I got only time issue on older processor (Xeon L5420 for example)
Sorry, but my report was on a machine with Xeon E5620 (has nonstop_tsc). 
I did some tests to reproduce the behavior by blocking NTP traffic via 
iptables but the problem did not reproduce in a lab setting.
Regards Andreas

>     *From:*Olivier Hanesse [mailto:olivier.hanesse@xxxxxxxxx
>     <mailto:olivier.hanesse@xxxxxxxxx>]
>     *Sent:* Wednesday, October 19, 2011 2:48 PM
>     *To:* Thomas Pöhler
>     *Cc:* Simonet Philippe, ITS-SDL-EIS-CNV-DLE-TLC; ml-xen-devel@xxxxxx
>     <mailto:ml-xen-devel@xxxxxx>
>     *Subject:* Re: [Xen-devel] Possible hint for "Clocksource tsc
>     unstable" problem____
>
>     __ __
>
>     Hi,____
>
>     __ __
>
>     Same here.____
>
>     I am using ntpd in dom0 too (lenny version so 4.2.4)____
>
>     __ __
>
>     Regards____
>
>     __ __
>
>     Olivier____
>
>     __ __
>
>     2011/10/19 Thomas Pöhler <tp@xxxxxxxxxxxxxxxxxxxxxxx
>     <mailto:tp@xxxxxxxxxxxxxxxxxxxxxxx>>____
>
>     Hi Philippe,
>
>     we are using ntpd in Dom0 too. (also ntpd - NTP daemon program -
>     Ver. 4.2.6p2).
>     We are monitoring ntp over a nagios plugin, check_ntp
>
>     Regards
>     Thomas
>
>     -----Ursprüngliche Nachricht-----
>     Von: Philippe.Simonet@xxxxxxxxxxxx
>     <mailto:Philippe.Simonet@xxxxxxxxxxxx>
>     [mailto:Philippe.Simonet@xxxxxxxxxxxx
>     <mailto:Philippe.Simonet@xxxxxxxxxxxx>]
>     Gesendet: Mittwoch, 19. Oktober 2011 14:40
>     An: olivier.hanesse@xxxxxxxxx <mailto:olivier.hanesse@xxxxxxxxx>;
>     Thomas Pöhler
>     Cc: ml-xen-devel@xxxxxx <mailto:ml-xen-devel@xxxxxx>
>     Betreff: RE: [Xen-devel] Possible hint for "Clocksource tsc
>     unstable" problem____
>
>
>     Hi Olivier and Thomas
>
>     andreas ist using ntpd in DOM0, myself too (ntpd - NTP daemon
>     program - Ver. 4.2.6p2)
>
>     and you ?
>
>     @andreas : how do you turn on your ntp logging ?
>
>     Philippe
>
>
>      > -----Original Message-----
>      > From: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
>     <mailto:xen-devel-bounces@xxxxxxxxxxxxxxxxxxx> [mailto:xen-devel-
>     <mailto:xen-devel->
>      > bounces@xxxxxxxxxxxxxxxxxxx <mailto:bounces@xxxxxxxxxxxxxxxxxxx>]
>     On Behalf Of Andreas Kinzler
>      > Sent: Tuesday, October 18, 2011 5:01 PM
>      > To: xen-devel@xxxxxxxxxxxxxxxxxxx
>     <mailto:xen-devel@xxxxxxxxxxxxxxxxxxx>
>      > Subject: [Xen-devel] Possible hint for "Clocksource tsc unstable"
>     problem
>      >
>      > Hello,
>      >
>      > I made an interesting observation related to the "Clocksource tsc
>      > unstable (delta = -2999660320319 ns)" problem. In the log of ntpd
>     I found:
>      >
>      > Oct  5 03:46:35 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 04:03:41 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 05:29:03 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 05:46:09 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 06:54:30 greenville-dom0 ntpd[4020]: synchronized to
>      > 192.53.103.104 <tel:192.53.103.104>, stratum 1
>      > Oct  5 06:54:30 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 07:28:22 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 08:19:35 greenville-dom0 ntpd[4020]: synchronized to
>      > 192.53.103.108 <tel:192.53.103.108>, stratum 1
>      > Oct  5 10:53:26 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 11:27:32 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 12:01:41 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 12:18:44 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 13:09:58 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 13:27:04 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 15:26:37 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 15:43:41 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 17:43:11 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 6001
>      > Oct  5 19:08:31 greenville-dom0 ntpd[4020]: synchronized to
>      > 192.53.103.104 <tel:192.53.103.104>, stratum 1
>      > Oct  5 19:08:31 greenville-dom0 ntpd[4020]: kernel time sync status
>      > change 2001
>      > Oct  5 21:23:58 greenville-dom0 ntpd[4020]: no servers reachable
>      > Oct  5 21:40:58 greenville-dom0 ntpd[4020]: synchronized to
>      > 192.53.103.104 <tel:192.53.103.104>, stratum 1
>      > Oct  5 21:40:58 greenville-dom0 ntpd[4020]: time correction of -3000
>      > seconds exceeds sanity limit (1000); set clock manually to the
>     correct
>      > UTC time.
>      >
>      > I had the problem twice on different machines but everytime I saw the
>      > line "no servers reachable". Could ntpd and/or linux kernel make some
>      > stupid stuff when ntp upstream servers are unreachable?
>      >
>      > Regards Andreas
>      >
>      >
>      > _______________________________________________
>      > Xen-devel mailing list
>      > Xen-devel@xxxxxxxxxxxxxxxxxxx <mailto:Xen-devel@xxxxxxxxxxxxxxxxxxx>
>      > http://lists.xensource.com/xen-devel____
>
>     __ __
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel