NTP- syslog message

[Jeeshan]

I have found the following repetative error message fro NTP client. The server is 7440 PA-RISC and 11iv2 OS
xntpd[21798]: synchronized to 172.16.10.5, stratum=11
xntpd[21798]: time reset (step) 0.473885 s
xntpd[21798]: synchronisation lost

Anyone have any suggestions why its giving?

[TTr · Accepted solution]

This message is normal telling you that the time adjustment in the previous line was greater than 128ms. In your case it was 473ms. Your client is drifting a little bit higher than usual between NTP adjustments. In some environments 473ms is acceptable in others it is not.

You can play with the "minpoll" setting in your ntp.conf file in the server line but you may cause unecessary network traffic.

Also the fact that you are at stratum 11, does not guarantee that the source time is accurate either. The drifting may be happening in previous strata and not between your client and the 7440 server.

As an alternative in some clients, I schedule a cronjob to run "ntpdate" as many times as needed (in my case I ran it once every 24 hours) and NOT use the ntpd at all on the client.

[Bill Hassell]

If you have databases or other data that must be in sync with other servers, using ntpdate is not a good idea as it can jump the current time and either drop or add seconds to the 24 hour period. ntp is designed to avoid this type of problem but the stratum=11 message is troublesome. Normally, stratum levels more than 5 to 9 are reserved for internal clocks in case all your NTP servers are unreachable.

Post your /etc/ntp.conf file (without comments) like this:

grep -v ^\# /etc/ntp.conf

Then post the output of this command:

ntpq -p

That will help explain the current situation.

[Jeeshan]

Dear all

I have made a mistake in my thread.
NTP server is in 3440 machine(OS 11i v1) and client is 7440 (OS 11i v2).
I have posted the syslog message of client machine.

My client ntp.conf is
broadcastclient yes
driftfile /etc/ntp.drift

Client ntpq -p output
remote refid st t when poll reach delay offset disp
==============================================================================
172.16.10.5 LOCAL(1) 11 b 28 64 37 0.15 315.023 1984.28

My server ntp.conf is

server 127.127.1.1 prefer
fudge 127.127.1.1 stratum 10
broadcast 172.16.10.255

remote refid st t when poll reach delay offset disp
==============================================================================
*LOCAL(1) LOCAL(1) 10 l 42 64 377 0.00 0.000 10.01
172.16.10.255 0.0.0.0 16 - - 64 0 0.00 0.000 16000.0

[TTr]

I personally do not use broadcast in NTP and prefer the direct connection in the ntp.conf by using the "server" statement. Not sure if this is the cause of your problems.

[TTr]

Have you also looked at the server and see what kind of adjustments are taking place there? Maybe the drifting is not a problem with your client.

[Tim Nelson]

Also...

If you have an NTP server and it is serving up stratum 11 to your clients then you need to rework your environment.

I suggest starting over and make sure you use at least stratum 3 NTP servers or get yourself a hard clock resource.

[Jeeshan]

thanks guyz, i think i need to do some work around again to resolve this issue.