[ntp:questions] ntp execution starts two processes?

Joshua Coombs jcoombs at gwi.net
Thu Mar 17 16:34:02 UTC 2005


I've installed a dev snapshot of ntpd on a FreeBSD 4.11 box, and noticed
some odd behavior, when I start the process, I get two processes listed.  I
can kill the second process and NTP seems fine, kill the first, they both
die.  Leave them both running and my time keeping drifts terribly, with a
pps adjustment rate twice what it normally settles at.

rassilon# ps auxwww | grep ntp
root       85  0.0  0.3  3104 1508  ??  Ss   11:19AM   0:00.16 
/usr/local/bin/ntpd -g -p /var/run/ntpd.pid -N
rassilon# kill 85
rassilon# ntpdate ntp.gwi.net
17 Mar 11:24:19 ntpdate[161]: adjust time server 207.5.128.20 
offset -0.000921 sec
rassilon# /usr/local/bin/ntpd -g -p /var/run/ntpd.pid
rassilon# ps auxwww | grep ntp
root      163  1.1  0.3  3104 1612  ??  Ss   11:24AM   0:00.12 
/usr/local/bin/ntpd -g -p /var/run/ntpd.pid
root      164  0.0  0.3  3104 1612  ??  S    11:24AM   0:00.00 
/usr/local/bin/ntpd -g -p /var/run/ntpd.pid


Mar 17 11:22:15 rassilon ntpd[85]: kernel time sync enabled 2001
Mar 17 11:24:15 rassilon ntpd[85]: ntpd exiting on signal 15
Mar 17 11:24:31 rassilon ntpd[162]: ntpd 4.2.0a at 1.1345-o Sat Mar 12 00:14:14 
GMT 2005 (1)
Mar 17 11:24:38 rassilon ntpd[163]: kernel time sync disabled 2041
Mar 17 11:24:48 rassilon ntpd[163]: kernel time sync enabled 2001
Mar 17 11:24:49 rassilon ntpd[163]: clock LOCAL(1) event 'clk_fault' (0x03)

I built from ntp-dev-4.2.0a-20050310

My conf:
### Remote servers
server ntp.colby.edu            iburst          maxpoll 12
server ntp0.jensenresearch.com  iburst          maxpoll 12
server timeserver.cs.umb.edu    iburst          maxpoll 12
server nist1-ny.glassey.com     iburst          maxpoll 12
server time.no-such-agency.net  iburst          maxpoll 12

### Local Peers
peer omega.gwi.net

### Local Clock - Last Resort!
server 127.127.1.1
fudge 127.127.1.1 stratum 12

### Driftfile
driftfile /var/db/ntp.drift

This is occuring on two separate FreeBSD boxen, both built from the same dev 
snap.  Should I just try a later dev snap or did I badly misconf something?

Joshua Coombs 





More information about the questions mailing list