On Fri, May 12, 2017 at 05:08:50PM +0200, Paul de Weerd wrote:
> On Fri, May 12, 2017 at 04:35:40PM +0200, Paul de Weerd wrote:
> | Thanks for your reply. Yours and an off-list reply both pointed at
> | the HZ parameter. I built a kernel with HZ=1000, and things now work
> | much better in the vm.
>
> I spoke too soon :-(
>
> Things do work better, but the clock is still slow:
>
> [weerd@vm1] $ rdate -npv ${N}; uptime
> Fri May 12 16:50:14 CEST 2017
> rdate: adjust local clock by 124.798956 seconds
> 4:48PM up 16 mins, 1 user, load averages: 0.00, 0.00, 0.00
>
> So, in 16 minutes, the clock is 2 minutes behind. So, a lot better,
> but still falling behind.
>
> [weerd@vm1] $ rdate -npv ${N}; time sleep 600; rdate -npv ${N}
> Fri May 12 16:51:47 CEST 2017
> rdate: adjust local clock by 135.359214 seconds
> 10m03.01s real 0m00.00s user 0m00.00s system
> Fri May 12 17:03:08 CEST 2017
> rdate: adjust local clock by 213.572802 seconds
>
> Which turns out to be almost 8 seconds per minute.
>
> Paul 'WEiRD' de Weerd
>
> --
> >++++++++[<++++++++++>-]<+++++++.>+++[<------>-]<.>+++[<+
> +++++++++++>-]<.>++[<------------>-]<+.--------------.[-]
> http://www.weirdnet.nl/
>
Known issue.
