Hello, fellow timekeepers,

> You really want your time sources to be "real" physical
> systems as much as possible, as virtualization inherently
> offers much worse timing than an OS running on bare metal.

Oh well!

Thanks to Ryan for what may well be good advice (I am not
sure and I cannot tell) - provided you have the choice to
follow it.

On the other hand, if you don't, don't worry, don't bother.

If a virtual machine is what you have, do not hesitate to
put an ntpd on it.  And do not hesitate to add that to the
pool, either.

At least that's what I do.  A virtual machine is the only
thing I run that's connected to the internet and has a
stable IP number.  From the monitoring page at
http://www.pool.ntp.org/scores/109.75.188.245 , it seems to
consistently stay be within +/- 10 ms.  And that is only
within the precision of the monitoring process, which is
known to be limited.

Upon being asked, my ntpd itself has this to say:

  $ ntpdc -c kerninfo
  pll offset:           0.000145 s
  pll frequency:        -0.526 ppm
  maximum error:        0.069376 s
  estimated error:      0.001136 s
  status:               0001  pll
  pll time constant:    10
  precision:            1e-06 s
  frequency tolerance:  500 ppm

This may be worse than what you could do with bare iron.  Or
maybe I already get the precision that's possible, given the
machine's network connection.  I am not sure.  But then, it
does not matter all that much to me.

What does matter to me is what I hand out to the world:

  $ ntpdc -c sysstat
  time since restart:     12349870
  time since reset:       12349870
  packets received:       613732613
  packets processed:      52763
  current version:        312114698
  previous version:       295896455
  declined:               611
  access denied:          14085
  bad length or format:   5747856
  bad authentication:     41186
  rate exceeded:          96461802

Some 43 times per second (pun obviously intended).

Regards, Andreas

P.S.: For the record: I don't want to talk against Ryan's
warning regarding interconnecting various ntpd instances
that _run on the same underlying hardware_.  That warning
stands utterly unaffected by anything I've written.  I just
want to make the point that an ntpd on a virtual machine can
be a valuable pool member.

_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool

Reply via email to