Re: timekeeping regression?

[Brad Spencer <brad%anduin.eldar.org@localhost>]

"Greg A. Woods" <woods%planix.ca@localhost> writes:

> [1:text/plain Hide]
>
> At Tue, 18 Jun 2024 21:11:50 -0400, Brad Spencer <brad%anduin.eldar.org@localhost> wrote:
> Subject: Re: timekeeping regression?
>>
>> The drifts and instability I have seen with respect to time have always
>> been in the DOMUs, especially pure PV DOMUs with more than one vcpu.
>
> Do you have your dom0 vcpus pinned?  I only saw dom0 drift when I
> un-pinned their vcpus, otherwise it's rock-solid.

I do not pin CPUs for the DOM0...  I never give more than one vcpu to
DOM0, as that didn't work in the past.  I do run the DOM0s at 1000HZ and
they keep perfect ntpd time (at least with 10.0_BETA and before, don't
know about after that).

>> I suppose the 7.5 days of uptime could not be due to a cron job that
>> happens to run at about that time??
>
> No, nothing like that.
>
> I'm thinking wild thoughts right now about how there might be some
> adjustment to the vCPU scheduling in the hypervisor after that much time
> elapses, and that somehow upsets the RDTSC emulation.  Or maybe there's
> just a plain old bug in Xen's RDTSC emulation.  Either seem highly
> unlikely though.
>
> After all it does have to have something to do with the Xen hypervisor
> as I didn't see time warps in domUs either until I upgraded to 4.18.

That would be another difference for me...  I am still running 4.15, but
have plans to move to 4.18 at some point.  Just don't know when.

> --
> 					Greg A. Woods <gwoods%acm.org@localhost>
>
> Kelowna, BC     +1 250 762-7675           RoboHack <woods%robohack.ca@localhost>
> Planix, Inc. <woods%planix.com@localhost>     Avoncote Farms <woods%avoncote.ca@localhost>





-- 
Brad Spencer - brad%anduin.eldar.org@localhost - KC8VKS - http://anduin.eldar.org