[ntp:questions] Flash 400 on all peers; can't get ntpd to be happy

Chuck Swiger cswiger at mac.com
Wed Mar 9 18:10:24 UTC 2011


On Mar 8, 2011, at 5:56 PM, Steve Kostecke wrote:
> On 2011-03-08, Chuck Swiger <cswiger at mac.com> wrote:
[ ... ]
>>> NTP disciplines the system (i.e. kernel) clock, not the hardware
>>> clock on the mother board.
>> 
>> That's right, although in reasonably common for platforms to
>> periodically write the system clock time back to the hardware
>> clock-- variously called the RTC/TOD/TOY clock which is in the
>> BIOS/EFI/firmware and keeps time when the system is off.
> 
> The RTC is _updated_, not synced, by the kernel.

Right.  Perhaps I'm mistaken, but I don't recall suggesting otherwise...?

[ ... ]
>>> I have a Debian 6.0 system running as a VMWare guest. ntpd on this
>>> system has no problem disciplining the clock.
>> 
>> OK. Does it do any better than using VMWare's "tools.syncTime = true"?
> 
> I don't have access to the host.

OK.  If I was going to compare two things, well, I would actually try them
both while measuring relevant factors to my comparison.  If I had never
tried one of the two, and I didn't have any other data available, then
I wouldn't try to draw a conclusion about which alternative works better.

Note [1].

>> Your jitter values are well over an order of magnitude worse than that
>> of ntpd running on a non-virtualized machine, and your offsets are
>> nearly an order of magnitude worse:
> 
> You're comparing apples and oranges.

Absolutely.  In fact, this is exactly the point I was making.

>> For all of that, your VM is doing pretty well running ntpd compared to
>> others I'd seen. I'd imagine the host running the VM isn't especially
>> busy; if it was, I wouldn't be surprised if ntpd can't manage to
>> discipline the clock without "tinker panic 0".
> 
> The default panic threshold is 1024 seconds. 

Right.  Depending on which VM technology is being used, it's entirely
possible to suspend a guest OS for longer than 1024 seconds.

Again, there's a point lurking here about the quality of timekeeping which
is possible within a VM.  A real machine, or the host ESX/Dom 0/whatever,
ought to be able to keep good time without this option, and I would
be looking for hardware failure, interrupt routing problems, etc if it
could not.

>>>> You are better off running ntpdate (or sntp) periodically via cron
>>>> in the DomUs.
>>> 
>>> Perhaps in certain cases, but not across the board.
>> 
>> I'd be happy to review counterexamples to my generalization...
> 
> There's my example.

See note [1] above.  To my mind, you don't have enough data to
draw a conclusion from your example.

Regards,
-- 
-Chuck




More information about the questions mailing list