[ntp:questions] what happens when sys.peer turns stratum 16?
unruh at wormhole.physics.ubc.ca
Mon May 31 03:42:58 UTC 2010
On 2010-05-30, bombjack <bombjack99 at gmail.com> wrote:
> I am fairly new to the ntp arena, but have to learn fast as part of my
> work is trying to fix ntp problems. Lately I had a customer that
> complained that there servers went bananas due to time being wrong.
> (The customer claims that ntp had changed the time 20 years ahead).
> Looking at some wire shark logs it seems that sys.peer for some reason
> had turned into a stratum 16 server (i.e. unsynched) and had "leap"
> set to 3 -"Alarm condition (clock not synchronized"
> the setup: klient syncs to 2 servers, one stratum 1 (sys.peer) and one
> stratum 2 server. If sys.peer all of a sudden says it's a stratum 16
> server, how does ntp react to this? instant outvote of the sys.peer?
> or are these packets filtered for some time as being errornous? or
> anything else?
> I have told the customer to use more than two servers, but they insist
> that they want answers to the above scenario.
Tell them that the source code available, and they can read it.
But I think it may also depend on which version of ntp they are running.
Which version are they using?
ntpd should just ignore a stratum 16 server-- ie act as if it did not
Note, that this is one of the reasons why your customers should never
use 2 servers. You have no way to know which one is crazy. Use 3 or 5.
(4 can be as bad as 2 if two of the servers go nuts in exactly the same
way-- eg they are both tied to a single server which has gone nuts).
Now if the clients ntpd is going nuts because a server went to stratum
16 that suggests a bug in ntpd.
> Any help welcome
More information about the questions