Re: Is high jitter contagious?
- From: Harlan Stenn <stenn@xxxxxxx>
- Date: Mon, 10 Sep 2007 22:42:52 +0000
For the situation you describe I would recommend running the local refclock
on your master server at 2 stratum values "worse" than the worst upstream
server it would be talking to.
If you leave the local refclock stratum at 10 and this machine usually talks
to an S2 server, that means your master node will ordinarily offer time to
the rest of your systems as S3. If this master node ever stops getting
access to the outside time sources, it will jump straight down to S10, and
the other machines will slowly "wind down" to that point.
I recommend you take a look at
http://support.ntp.org/Support/OrphanMode
for the new recommended way of handling your case, or if you want to stick
with the "old" way:
https://support.ntp.org/Support/ConfiguringLocalRefclocks
I would also recommend you have all the machines in your group (it's about
11 machines, right?) peer with each other. Doing so will:
- make it easy to monitor them (run 'ntpq -p' on any machine)
- keep them all in-step with each other
- something else that I forgot
H
.
- Follow-Ups:
- Re: Is high jitter contagious?
- From: dromedaryl
- Re: Is high jitter contagious?
- References:
- Is high jitter contagious?
- From: dromedaryl
- Is high jitter contagious?
- Prev by Date: Re: NTP settings for machine with irregular, short connections to the Net
- Next by Date: Re: NTP settings for machine with irregular, short connections to the Net
- Previous by thread: Re: Is high jitter contagious?
- Next by thread: Re: Is high jitter contagious?
- Index(es):
Relevant Pages
|
