[Tfug] ntpd problems - server losing an hour or more each day

[Chad Woolley]

On 10/30/06, Bexley Hall <bexley401 at yahoo.com> wrote:
> What does this machine normally *do*?
> I assume the RTC (if you reboot the machine and
> examine it *before* any ntpdate(1) calls) *is*
> maintaining proper time.
>

It is a Continuous Integration box.  We run the free version of VMWare
Server.  It tests and builds all our (Ruby) projects, all day long,
whenever any change is checked into Subversion.

If by RTC, you mean the actual host OS clock, then yes, it's correct.
Other VMs on the same host are also correct.

> ntpd will "give up" if the kernel's idea of the
> "current time" differs from *it's* idea of the
> current time.  Presmably, a discrepancy comes into
> existence at some point during normal operation.
> Then, ntpd takes itself out of the loop and the
> discrepancy *grows*.
>
> Usually, resulting in time being *lost* (not
> gained).

Yes.  We've tried to specify command line options to tell ntpd NOT to
give up, but this seemed to have no effect.

>
> If the machine sees heavy I/O, this can result in
> the jiffy being missed -- hence the time "loses"
> one jiffy.  Do this a few hundred times in each period
> of heavy disk I/O (e.g., burning CD's!) and the
> effects are cumulative.

Yes, the VM does heavy disk IO.  It checks out projects, builds them,
writes multiple logs, deletes them, and starts over.  It does this all
day long for multiple projects.  I could check iostat, but... (see
below)

>
> I don't know how much leeway you have regarding how
> the machine is used.  Can you leave it quiescent
> and verify that everything works fine?  With and
> *without* ntpd running?

Actually, I'm just going to stick with the crontab workaround for now.
 The bigger plan is to move this VM entirely off the VMWare box to a
Xen virtual server, which hopefully won't have these problems.

The VMWare machine is overloaded anyway, so we are going to keep it
just to run the non-Linux VMs that we can't run under Xen.

Thanks anyway for your help, I'm just punting and keeping my
workaround.  Too many other things to do.  I just wanted to report
back to the list for clo-sure.

-- Chad