On Wed, 1 Jun 2011, Jan van Eldik wrote:

>> > > >  I'm running a couple of CentOS 5.6 instances under Hyper-V. Horrible
>> > > >  clock drift
>> > > >  issues, but otherwise okay.
>> > > 
>> > >  You may want to add:
>> > > 
>> > >  divider=10 clocksource=acpi_pm
>> > 
>> >  Doesn't help me.
>
> Just to add that we different parameters on 32bit and 64bit SLC5 systems.
> From https://cern.service-now.com/kb_view.do?sysparm_article=KB0000278
>
>     * SLC5/x86_64: "divider=10 clock=pmtmr"
>     * SLC5/i386: "divider=10 clocksource=acpi_pm"
>     * SLC4/x86_64: "divider=10 notsc"
>     * SLC4/i386: "divider=10 notsc"
>
> These parameters seem to work fine on 500+ VMs

Interesting, that's what we did:

  - EL5/x86_64: elevator=noop divider=10 notsc
  - EL5/i386:   elevator=noop divider=10 clocksource=acpi_pm
  - EL4/x86_64: elevator=noop divider=10 notsc
  - EL4/i386:   elevator=noop divider=10 clock=pmtmr

So it's different for EL5/x86_64 and EL4/i386. This was based on:

     http://kb.vmware.com/kb/1006427

-- 
-- dag wieers, [log in to unmask], http://dag.wieers.com/
-- dagit linux solutions, [log in to unmask], http://dagit.net/

[Any errors in spelling, tact or fact are transmission errors]