Description of problem: ntpd essentially loses synchronisation Version-Release number of selected component (if applicable): kernel-2.4.9-e.4, ntp-4.1.2-1.AS21.2 Build date Build Date: Wed Feb 4 05:04:16 2004, glibc-2.2.4-32.15 How reproducible: Steps to Reproduce: 1.Problem occurs whilst ntpd is running - the build date is also recent. /etc/init.d/ntpd start 2. Observe enteries in /var/log/hosts/[hostname]/ subdirectories (using egrep -r -e ntp * > /tmp/ntpd_errs.txt ) 3. Actual results: Best shown in log file when I'm presented with that option Expected results: No time drift or lost sync. Additional info: Synchronisation should be maintained with this recent build of ntp and time should be maintained to a few milli-seconds as other applications like kerberos misbehaves when the clock drift is 'too large'. Furthermore, the ntp build date shows that this bug has reappeared, as there was a previous fix to the same problem.
Created attachment 101923 [details] ntp log of lost synchronisation with time servers This is associated to bug # 127893
Hi Harald ! Can you please investigate why we are having these lost synchronization entries in most of our RH ES 2.1 servers running kernel release 41 and all latest updates from RH ?
CC
Why are these messages present?? Do you start the ntp daemon twice? Do you use ntpdate? 07/daemon-07-07:Jul 7 15:00:46 aklim905 ntpd[28281]: bind() fd 4, family 2, port 123, addr 0.0.0.0, in_classd=0 flags=1 fails: Address already in use 07/daemon-07-07:Jul 7 15:00:46 aklim905 ntpd[28281]: bind() fd 4, family 2, port 123, addr 127.0.0.1, in_cla07/daemon-07-07:Jul 7 15:00:46 aklim905 ntpd[28281]: bind() fd 4, family 2, port 123, addr 0.0.0.0, in_classd=0 flags=1 fails: Address already in use 07/daemon-07-07:Jul 7 15:00:46 aklim905 ntpd[28281]: bind() fd 4, family 2, port 123, addr 127.0.0.1, in_classd=0 flags=0 fails: Address already in use 07/daemon-07-07:Jul 7 15:00:46 aklim905 ntpd[28281]: bind() fd 4, family 2, port 123, addr 10.65.16.33, in_classd=0 flags=1 fails: Address already in usessd=0 flags=0 fails: Address already in use 07/daemon-07-07:Jul 7 15:00:46 aklim905 ntpd[28281]: bind() fd 4, family 2, port 123, addr 10.65.16.33, in_classd=0 flags=1 fails: Address already in use
Frank, do you see similar messages? Could you attach your logfiles, too? Also interesting would be the output of: # /usr/sbin/ntpdc > peers
Hi Harald ! Here is the output you requested. let me know if you need anything else, since the logfiles would come from the same server. [root@aklim905 ntp]# uname -a Linux aklim905.airnz.co.nz 2.4.9-e.41smp #1 SMP Wed Jun 16 23:41:56 EDT 2004 i686 unknown [root@aklim905 ntp]# rpm -q glibc glibc-2.2.4-32.15 [root@aklim905 ntp]# ntpdc ntpdc> peers remote local st poll reach delay offset disp ======================================================================= =LOCAL(0) 127.0.0.1 10 64 377 0.00000 0.000000 0.00093 =aklim903.nz.air 10.65.16.33 3 1024 377 0.00052 0.000455 0.01483 =aklim904.nz.air 10.65.16.33 3 1024 377 0.00038 0.000620 0.01485 *aklft901.airnz. 10.65.16.33 3 1024 377 0.00038 0.000557 0.01483 =aklft902.airnz. 10.65.16.33 3 1024 377 0.00060 0.000325 0.01483 ntpdc> listpeers client LOCAL(0) client aklim903.nz.airnz.com client aklim904.nz.airnz.com client aklft901.airnz.co.nz client aklft902.airnz.co.nz
Bump!
I cannot reproduce it... do you have connection problems sometimes?
Closing due to inactivity.