Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

NTP synch issue

Dear All,

The Unix box keeps on losing sync to the router, the message as below:

Mar 31 07:24:33 micros1 xntpd[5412]: synchronisation lost

Mar 31 07:30:57 micros1 xntpd[5412]: synchronisation lost

Mar 31 07:37:21 micros1 xntpd[5412]: synchronisation lost

Mar 31 07:43:45 micros1 xntpd[5412]: synchronisation lost

Mar 31 07:50:09 micros1 xntpd[5412]: synchronisation lost

Mar 31 07:56:33 micros1 xntpd[5412]: synchronisation lost

Mar 31 08:02:57 micros1 xntpd[5412]: synchronisation lost

Mar 31 08:09:21 micros1 xntpd[5412]: synchronisation lost

Problem happened when router rebooted, the configuration for the ntp never changed on the router:

APBASES1#sh ntp asso

address ref clock st when poll reach delay offset disp

+~135.89.101.229 135.89.249.44 3 303 512 377 109.4 7.10 8.2

*~135.89.100.102 135.89.1.141 3 408 512 377 83.4 1.29 43.1

+~172.24.200.1 135.89.100.102 4 173 512 377 62.8 0.99 1.2

+~172.24.200.3 135.89.100.102 4 489 512 164 58.3 1.30 37.1

* master (synced), # master (unsynced), + selected, - candidate, ~ configured

APBASES1#

APBASES1#sh ntp stat

Clock is synchronized, stratum 4, reference is 135.89.100.102

nominal freq is 250.0000 Hz, actual freq is 250.0210 Hz, precision is 2**18

reference time is CB9C240E.ADEB560F (03:01:34.679 GMT Tue Apr 1 2008)

clock offset is 1.2861 msec, root delay is 181.64 msec

root dispersion is 70.94 msec, peer dispersion is 43.06 msec

APBASES1#sh run | i ntp

ntp clock-period 17178422

ntp source Loopback0

ntp server 135.89.101.229

ntp server 135.89.100.102 prefer

ntp peer 172.24.200.1

ntp peer 172.24.200.3

Any suggestion for this issue will be appreciated.

Thank You,

Best Regards,

Trina

10 REPLIES
Hall of Fame Super Silver

Re: NTP synch issue

Trina

I think that there must be some aspect of your post that I am not understanding. the show ntp association and the show ntp status show that APBASES1 has successfully synced with 135.89.100.102 which is its preferred NTP server. Your post says something about the router rebooting and the configuration changing, but I can not tell what you think changed.

HTH

Rick

New Member

Re: NTP synch issue

From router end, I can see the ntp always synced, even enabled debug, I couldn't see anything wrong also, however, customer unix box keep on poping error message like:

Mar 31 09:35:03 micros1 xntpd[19005]: synchronisation lost

Mar 31 09:41:27 micros1 xntpd[19005]: synchronisation lost

Mar 31 09:47:51 micros1 xntpd[19005]: synchronisation lost

Mar 31 09:54:15 micros1 xntpd[19005]: synchronisation lost

Mar 31 10:00:39 micros1 xntpd[19005]: synchronisation lost

I am not sure what could be the good troubleshoot step here.

Re: NTP synch issue

Hi Trina,

Need to ask a lot of questions :)

Can you post your NTP config (i.e. ntp.conf) and the "ntpq -pn" command output?

Is there a firewall in-between the UNIX box and the NTP Server? If there is a firewall, are the necessary ports open (TCP/UDP 123) and is there NAT (Dynamic Static)?

What is the UNIX box IP Address use to sync with the NTP Server and what is the NTP Server IP Address use to sync with the UNIX box?

Are both NTP Server and UNIX box able to reach each other (i.e. PING maybe)?

From your original post, it seems to be that the router is your NTP Server? Correct me if I'm wrong.

I never use a router as NTP Server - because I like to use NTP Pool : ). However, I think if the router will be configured as NTP Server, this command line should be present "ntp master stratum_number"

Regards,

Dandy

Hall of Fame Super Silver

Re: NTP synch issue

I would be cautious about configuring ntp master stratum on the router. As long as the router has learned authoritative time it will act as an NTP server without needing the ntp master command. And if it has not learned authoritative time do you want it acting as a master?

I agree that answering the questions that Dandy asks will help find the problem.

HTH

Rick

Re: NTP synch issue

Hi Rick,

Good point.

This means the router is acting as NTP Relay?

Regards,

Dandy

Hall of Fame Super Silver

Re: NTP synch issue

Dandy

In IOS once the router has learned authoritative time it will act as an NTP server (or NTP peer) to other devices without needing any additional NTP configuration.

HTH

Rick

Re: NTP synch issue

Thanks Rick.

Regards,

Dandy

New Member

Re: NTP synch issue

Thanks a lot for your reply! Since the unix is managed by customer, so I can not give you the output from unix box. However, there is no FW between router to unix box, the NTP server is router. I enabled the debug and observed that the peer is keep changing:

Apr 1 08:25:51: NTP: rcv packet from 135.89.100.102 to 172.24.197.105 on Loopback0:

Apr 1 08:25:51: leap 0, mode 4, version 3, stratum 3, ppoll 512

Apr 1 08:25:51: rtdel 1B8F (107.651), rtdsp 082A (31.891), refid 8759018D (135.89.1.141)

Apr 1 08:25:51: ref CB9C6E14.20BD8000 (08:17:24.127 GMT Tue Apr 1 2008)

Apr 1 08:25:51: org CB9C700E.EF3BFED8 (08:25:50.934 GMT Tue Apr 1 2008)

Apr 1 08:25:51: rec CB9C700E.FA32D000 (08:25:50.977 GMT Tue Apr 1 2008)

Apr 1 08:25:51: xmt CB9C700E.FA3C0000 (08:25:50.977 GMT Tue Apr 1 2008)

Apr 1 08:25:51: inp CB9C700F.78C5D570 (08:25:51.471 GMT Tue Apr 1 2008)

Apr 1 08:25:51: NTP: synced to new peer 135.89.101.229 <<<<<<<<

.Apr 1 08:26:28: NTP: xmit packet to 135.89.101.229:

.Apr 1 08:26:28: leap 0, mode 3, version 3, stratum 4, ppoll 1024

.Apr 1 08:26:28: rtdel 3691 (213.150), rtdsp 147C (80.017), refid 87596466 (135.89.100.102)

.Apr 1 08:26:28: ref CB9C6E0F.1C2E5587 (08:17:19.110 GMT Tue Apr 1 2008)

.Apr 1 08:26:28: org CB9C6EFD.FEA5C000 (08:21:17.994 GMT Tue Apr 1 2008)

.Apr 1 08:26:28: rec CB9C6EFE.115F0351 (08:21:18.067 GMT Tue Apr 1 2008)

.Apr 1 08:26:28: xmt CB9C7034.EE62A128 (08:26:28.931 GMT Tue Apr 1 2008)

.Apr 1 08:26:29: NTP: rcv packet from 135.89.101.229 to 172.24.197.105 on Loopback0:

.Apr 1 08:26:29: leap 0, mode 4, version 3, stratum 3, ppoll 1024

.Apr 1 08:26:29: rtdel 1FFB (124.924), rtdsp 0CCD (50.003), refid 8759F92E (135.89.249.46)

.Apr 1 08:26:29: ref CB9C6CAA.52845000 (08:11:22.322 GMT Tue Apr 1 2008)

.Apr 1 08:26:29: org CB9C7034.EE62A128 (08:26:28.931 GMT Tue Apr 1 2008)

.Apr 1 08:26:29: rec CB9C7034.F7726000 (08:26:28.966 GMT Tue Apr 1 2008)

.Apr 1 08:26:29: xmt CB9C7034.F77C5000 (08:26:28.966 GMT Tue Apr 1 2008)

.Apr 1 08:26:29: inp CB9C7035.0A7FE008 (08:26:29.041 GMT Tue Apr 1 2008)

Apr 1 08:26:39: NTP: rcv packet from 172.24.200.3 to 172.24.197.105 on Loopback0:

=================================

Below is the NTP configuration:

scheduler allocate 20000 1000

ntp clock-period 17178382

ntp source Loopback0

ntp server 135.89.101.229

ntp server 135.89.100.102 prefer

ntp peer 172.24.200.1

ntp peer 172.24.200.3

Please let me know what kind of other information can be provided. thank you very much.

Re: NTP synch issue

Hi Trina,

I think the router is having connection problem with NTP Source (AT&T DNS/NTP) - 135.89.100.102 and 135.89.101.229 from your debug log . Might be very far (RTT).

I experienced this with AT&T before. In you case, your router NTP packet reach AT&T DNS/NTP but it might have close the connection before it received a return packet from them. You can ask the AT&T DNS/NTP admin to perform a TCPDUMP.

[EDIT] Alternatively, you can try using NTP Source which is nearer to you. This fixed my problem also.

Regards,

Dandy

New Member

Re: NTP synch issue

Thanks a lot for your suggestion, I will check it out and get back to you the finding.

228
Views
4
Helpful
10
Replies