[ntp:questions] searching for doc: "ntp as an indicator ofimpending hardware failures"

Justin Pryzby jpryzby+qptpnt at quoininc.com
Fri Aug 31 14:40:43 UTC 2007


On Fri, Aug 31, 2007 at 09:09:57AM -0500, Jason Rabel wrote:
> > I've been searching for a doc I saw previously regarding how high
> > jitter/dispersion can indicate impending hardware failures, but in
> > my last 2 attempts haven't been able to find it.  I'm also interested
> > in docs regarding NTP and power quality.
> 
> I think it would just be sheer luck if NTP showed some anomalies before you
> had a serious hardware failure. Some pieces of hardware you can sometimes
> physically hear before they die (like hard drives), others can have partial
> failure (bad chunks of RAM), while others die instantly with no prior
> warning.
> 
> Can you elaborate on your question about power quality? I've seen unstable
> systems due to poor power quality, adding a UPS or other line conditioning
> device can usually resolve the issue. I've also seen unstable systems due to
> poor quality power supplies (in the PC). Swapping them out with a quality
> PSU can work wonders.
I have two issues:

 . I've seen histogram graphs of our NTP jitter/dispersion as it
   exceeds some thresholds; it varies by a factor of 16.  I was
   initially concerned that something was actually wrong but the
   problem seems to have resolved itself (which doesn't rule out the
   possibility that a real problem didn't exist)

 . I'm certain that I've seen this doc before and it bugs me that I
   can't find it :)

Thanks
Justin

PS. Please CC me.



More information about the questions mailing list