Network Time Protocol
Matt Sillitoe
I'm trying to confiure xntpd to run on some 9000/712's running 9.05
and 9.07. The code (from the liverpool archive) compiles OK. When I
configure the ntp.confs on the two test machines, and run in debug mode,
I see messages passing between the two, but the time is not
synchronized.
Has anybody got this running ? I'm using one woirkstation as the local
time server, but if I configure its ntp.conf to have 127.127.1.0 as the
server, which should be use local clock, I get a message in the log file
saying attenmp to configure an invalid address. There is another line
in the short ntp.conf saying broadcast 224.0.1.1
TRhe client is the started with an ntp.conf which has one line (server
<server ip address> in it. Its started with the commands line option of
-b to listen for the broadcast message and sync to them.
Lastly, adjtime does not appear to do anytyhing. I start it before xntpd
on both the client and the server. Using the -v or -d options does not
appear to produce any debug output. I simply get the message adjtimed
started or stopped in the syslog.
Anybody any ideas ?
Paul Medcalf
John Venneman
> Hi,
> I'm trying to confiure xntpd to run on some 9000/712's running 9.05
> and 9.07.
[...snip...]
>
> Has anybody got this running ? I'm using one woirkstation as the local
> time server, but if I configure its ntp.conf to have 127.127.1.0 as the
^^^^^^^^^^^
This is no longer the proper syntax for specifying a local clock. Instead try:
server 127.127.1.1
fudge 127.127.1.1 stratum 0
[...snip...]
--
John Venneman
Tom Lane
>> Has anybody got this running ? I'm using one woirkstation as the local
>> time server, but if I configure its ntp.conf to have 127.127.1.0 as the
> ^^^^^^^^^^^
> This is no longer the proper syntax for specifying a local clock. Instead
> try:
> server 127.127.1.1
> fudge 127.127.1.1 stratum 0
Furthermore, you should never, never, ever specify a local clock (or any
other less-than-perfect source of time) as stratum zero. Stratum 10 is a
good choice. In an isolated setup, the results will be just the same,
but if your machines ever end up exchanging NTP packets with *real*
timeservers, there won't be any question whose time is more accurate.
comp.protocols.time.ntp is the proper place to be asking about NTP
stuff, btw.
regards, tom lane
Dave Podolske
(a white paper?) on how to do it in a step-by-step format. I dont want
to
become an expert on the protocols and endless options for configuring
it.
Any help (email or newsgroup) would be appreciated.
Bill Hassell
: Has anyone here set up NTP on HPUX 10.01 yet? Im looking for information
: (a white paper?) on how to do it in a step-by-step format. I dont want
: to
: become an expert on the protocols and endless options for configuring
: it.
Just use SAM. It's at the bottom of the task list.
--
__?__
/ \
/ ^ ^ \
( (o) (o) )
\ /
*=oOOO======(_)=====OOOo====*=====================================*
| Bill Hassell | Hewlett-Packard Response Center |
| System Administrator | b...@atl.hp.com / Atlanta, GA. |
*=============Oooo.=========*=====================================*
.oooO ( )
( ) ) / "There are two types of computer users in the
\ ( (_/ world...those that have lost data, and those
\_) that are going to." (blh, circa 1972)
Neelesh Rangnekar
-- Neelesh
In article <31517D...@ecg.csg.mot.com>, Dave Podolske <da...@ecg.csg.mot.com> writes:
|> Has anyone here set up NTP on HPUX 10.01 yet? Im looking for information
|> (a white paper?) on how to do it in a step-by-step format. I dont want
|> to
|> become an expert on the protocols and endless options for configuring
|> it.
|> Any help (email or newsgroup) would be appreciated.
--
<!-- Neelesh Rangnekar --> ////
(. .)
*----------------------------o00-(_)-00o-----------------------------*