NTP not working.

[Arun Vijay V C]

Hii,
My server is not getting synchronized with the ntp server. The configuration is exactly same as of a working server. The connectivity to the ntp server also Okay.

But in ntpq -p the server is not in sync with any ntp server.

root@aucom1:(root)> ntpq -p
remote refid st t when poll reach delay offset disp
==============================================================================
brwgwn03.melosd 0.0.0.0 16 u 12 64 0 0.00 0.000 16000.0
192.170.84.15 0.0.0.0 16 u 11 64 0 0.00 0.000 16000.0
omsmeodns2.melo 0.0.0.0 16 - - 64 0 0.00 0.000 16000.0
omcmeodns1.melo 0.0.0.0 16 - - 64 0 0.00 0.000 16000.0

Ntp port is opened and Okay,

root@aucom1:(root)> lsof -i :123
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
xntpd 3608 root 4u inet 0x4867a868 0t0 UDP *:ntp (Idle)
xntpd 3608 root 5u inet 0x41dd8468 0t0 UDP ar-aucom.melosdbb.aus.hp.com:ntp (Idle)
xntpd 3608 root 6u inet 0x5af54e68 0t0 UDP aucom1:ntp (Idle)
xntpd 3608 root 7u inet 0x432a5868 0t0 UDP aucom1:ntp (Idle)
xntpd 3608 root 8u inet 0x55601868 0t0 UDP localhost:ntp (Idle)

But, When I try to do the update with ntp server,

root@aucom1:(root)> ntpdate -d 192.170.84.233
28 Oct 18:36:23 ntpdate[4560]: ntpdate version 3.5f: Fri Dec 10 18:32:44 GMT 1999 PHNE_19711
transmit(192.170.84.233)
transmit(192.170.84.233)
transmit(192.170.84.233)
transmit(192.170.84.233)
transmit(192.170.84.233)
server 192.170.84.233, port 123
stratum 0, precision 0, leap 00, trust 000
refid [0.0.0.0], delay 0.00000, dispersion 64.00000
transmitted 4, in filter 4
reference time: 00000000.00000000 Thu, Feb 7 2036 17:28:16.000
originate timestamp: 00000000.00000000 Thu, Feb 7 2036 17:28:16.000
transmit timestamp: ce9272fa.94d6f000 Wed, Oct 28 2009 18:36:26.581
filter delay: 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000 0.00000
filter offset: 0.000000 0.000000 0.000000 0.000000
0.000000 0.000000 0.000000 0.000000
delay 0.00000, dispersion 64.00000
offset 0.000000

28 Oct 18:36:27 ntpdate[4560]: no server suitable for synchronization found

[TTr]

You have to check if you NTP server is configured to use any trusted keys or has restrictions based on network segments. Look in the ntp server's config file for details.

[BUPA IS]

Hello,
Since you are not getting a response from any of your time servers do check that you have a reversible route to each of them .

I hope this idea is of some use

Mike

[Arun Vijay V C]

Hi..
I am able to ping to the configured NTP servers and getting the reply. Also, network team has confirmed that there is no firewall or block between the NTP servers and the client servers.
More over, there are many servers working with this NTP server in same n/w.

Thanks,
Arun

[Steven E. Protter]

Shalom,

telnet servername 123

Lets run diagnosis and see if the port is open.

Most corporate networks do not permit connections to outside ntp servers.

Windows NTP, may not be good enough for Unix.

SEP

[Bill Hassell]

The responses indicate that the NTP servers are not responding. Start by using ntpq -p to test local servers on the same subnet. If you have another HP-UX box running NTP with a simple ntp.conf file, you can ask that server what ntp server it is using:

ntpq -p

That will verify that your server can talk to an unrestricted NTP server. As mentioned before, your company may have setup NTP restrictions. For the working systems, cat the ntp.conf file without the comments:

grep -v "^#" /eetc/ntp.conf

That may reveal specialized settings needed for your NTP servers.

[Matti_Kurkela]

NTP uses UDP, not TCP.

Only TCP-based protocols can be tested with telnet. Attempting to test NTP with telneting to port 123 will lead to incorrect conclusions.

With NTP it's important that each NTP packet reaches its destination as fast as possible.

If some packets are lost, NTP can still use the rest. The TCP protocol would the hold the rest of packets and wait for re-transmission of missing packets. This is counter-productive for NTP, as any delay makes it harder to achieve a time synchronization.

MK

[Arun Vijay V C]

Correct Matti...

checking the connectivity by telnet is not possible here...

Any other ideas?..

Thanks,
Arun

[Johnson Punniyalingam]

traceroute

telnet ntp_server 123 --> should be fine,

not meaning need to telnet direct to ntp_server

if telnet doesn't show's up connected --> ntp port seems to open

if telnet turns with trying.....

I suspect we need check with "Network Team" to open the ntp port

[BUPA IS]

Hello,

Since you are sure the network link is open from your client to your server.
you may want to try these.

- As suggested by TTr checking that the restrict statements in the server and client are correct may help.

- On the server use the : xntpdc -c monlist command to see if the sever tried to respond to your client's request ( assumes NTP v3 running on unix/linux for an NTP v4 server use ntpdc )
If there is an entry in there then the packet arrived at the server and can't find its way back. If there is no entry, then the packet either never arrived or has been rejected at the server end. You should review the log files of the NTP server.

- unlikely but you never know : Your NTP client patch is a bit out of date , It used to be possible to start a second ntp daemon before the old one stopped completely.
Just check that you do not have two copies of xntpd or something like an ntpdate -B running on your client (meaning adjust by slew only which may take days to finish) if so kill them all off and try again.

- The latest patch which supercedes PHNE_19711 is PHNE_31821

Mike