Red Hat Bugzilla – Bug 77086
Unable to sync system time
Last modified: 2007-04-18 12:48:04 EDT
I'm unable to use ntpdate to sync time&date because too big offset:
# ntpdate tak.cesnet.cz
2 Nov 01:36:50 ntpdate: Can't adjust the time of day: Invalid argument
# ntptrace tak.cesnet.cz
tak.cesnet.cz: stratum 1, offset -93634.278879, synch distance 0.00157, refid 'GPS'
# date 10312339
# ntpdate tak.cesnet.cz
31 Oct 23:39:11 ntpdate: step time server 188.8.131.52 offset 6.853757 sec
Seems odd error message or stupid bug... I don not remember same problem with
old version of ntpdate (without libcap) but I could be wrong (I did not tested it).
I've noticed this problem with the RedHat 8.0 distribution. It's not a problem
with ntp but a problem with dhcpcd. Apparently, it fails to make the entry:
in the /etc/ntp.conf file (where NTP-SERVER is the ip-address or name provided
via DHCP). Since it does make the entry:
restrict default ignore
the end result is that ntp is effectively blocked from syncing with anyone.
I was about to submit it as a dhcpc bug but decided to check the ntp bugs to see
if someone else had encountered the issue.
Seems to be two different bugs. I had no ntp daemon running and I did not used
dhcpd to obtain IP address. I'm only using ntpdate to sync time every hour
(because ntp daemon was not so fail-proof).
What I did is try to use with ntpdate to change local time and it refused to do
because "bad argument". This is odd. The second problem is that script for ntpd
daemon is using ntpdate to sync time like I did, but this simply not work when
time offset is too big as I reported.
I'm reopening to point that there is an inconsistency.
ok, found the bug in the source.. .thx for reporting...
I had problems with ntp that relate to the default ignore entry in the
ntp.conf. they didn't come from dhcp, I think it's just the defaul entry in
the config file.
- ntp doesn't log that it's not synching. in fact there's no indication that
anything's wrong unless you happen to do a manual clock check
- there's no warning from ntp that it's got a stupid configuration. in fact
it's not trivial to find that it's getting the returned signals correctly and
then throwing them away. (I faffed for ages checking firewall settings, and
ethereal to be sure that something was going on at all.
should be fixed with ntp-4.1.1b-1
An errata has been issued which should help the problem described in this bug report.
This report is therefore being closed with a resolution of ERRATA. For more information
on the solution and/or where to find the updated files, please follow the link below. You may reopen
this bug report if the solution does not work for you.