xntp problem

[cl_5]

Hi all
I've got a problem with xntpd on a L2000. I got the following message:
xntpd[16843]: xntpd version 3.5f: Fri Dec 10 10:27:26 GMT 1999 PHNE_19711
xntpd[16843]: tickadj = 625, tick = 10000, tvu_maxslew = 61875
xntpd[16843]: precision = 7 usec
xntpd[16843]: bind() fd 12, family 2, port 123, addr c053e5ff, in_classd=0 flags=0 fails: Address already in use

I don't have an idea, what does this mean.

Thanks for help

Carsten

[T G Manikandan]

NTP port usage is 123.

Looks like some other application is already using it.

Just use lsof to find out the process using the port

#lsof -i udp:123

Revert

[Elmar P. Kolkman]

Or, if you don't have lsof, try
'netstat -anf inet | grep 123'

[cl_5]

Thanks for your rapid answer.

Nether lsof nor netstat gave any output about Port 123. So it seems that the port isn't in use.

Carsten

[T G Manikandan]

do you have this patch PHNE_19711 installed.
Just do a
#swlist -l fileset|grep PHNE_19711

can you post your /etc/ntp.conf file


Thanks

[Elmar P. Kolkman]

If it still doesn't run (sometimes waiting a while and retry help, because all residu of the old run need to clean up before a new bind can be done) I think the problem is with your configuration file.
Could you give is an extract of the ntp.conf file, like the server/peer/broadcast lines ?

[RolandH]

Hi Carsten,

I think your port is already in use. Have you started before any xntpd process before? Only one is allowed to run at time.

Perhaps you have a died ntpdate process anywhere? Check that and kill them if exist.

Have you lsof on your system. Do an
# lsof -i UDP:123

and wich process is running on that port.

HTH
Roland

[cl_5]

The Patch PHNE_19711 is already installed.
Also lsof -i UDP:123 gave no answer.

Here the extract from ntp.conf:

peer rasrv2.cms.fra.dlh.de version 3 prefer

broadcastclient yes
driftfile /etc/ntp.drift
requestkey 65535
controlkey 65535

Thanks for help

Carsten

[T G Manikandan]

Try doing a

broadcastclient no

in your /etc/ntp.conf file.

Make sure that you update the latest NTP patch

http://www1.itrc.hp.com/service/patch/patchDetail.do?patchid=PHNE_27223&context=hpux:800:11:00

[RolandH]

Hi Carsten,

have you a ntpq on your system?
If yes do a "ntpq -p".
What is the output.

HTH
Roland

[RolandH]

Hi Carsten,

i have take a deeper look on your output and i think i now why this happens.

Could it be that you have configured a value
for NTPDATE_SERVER in /etc/rc.config.d/netdaemons.

Set this value to null if you have configured xntpd.

If you have set it you get following message if you start xntpd

# /sbin/init.d/xntpd start
17 Oct 15:46:52 ntpdate[18858]: the NTP socket is in use, exiting

and in syslog you get this message.
bind() fd 4, family 2, port 123, addr 00000000, in
_classd=0 flags=1 fails: Address already in use

You see same message!!!

So stop your xntpd daemon and set this in /etc/rc.config.d/netdaemons

export NTPDATE_SERVER=""
export XNTPD=1
export XNTPD_ARGS="-l /var/adm/syslog/ntp.log"

Restart your xntpd
/sbin/init.d/xntpd start

I have set a own log for the xntpd, because he sends a lot of messages and I did'nt want it in syslog.

HTH
Schoenes Wochenende !!!
Roland

[RolandH]

And as I see in your profile you never assign points to those who try to help you !

Please read this:

http://forums1.itrc.hp.com/service/forums/helptips.do?#28

Do it, please!!!

=|:-]

[cl_5]

I've installed the newest ntp patch, but it has no effect.
ntpq -p gave the following output:
ntpq: read: Can't assign requested adddress

And the settings in /etc/rc.config.d/netdaemons didn't really help.

And if I do a telnet to port 123 it is possible to connect. So it seems the port isn't really in use.

Have anybody some other ideas.

regards
Carsten

[RAC_1]

Does it have something to do wit keys that are in your /etc/ntp.conf file?

[RolandH]

Hi carsten,

have you checked that your /etc/services file is correct.

ntp 123/udp # Network Time Protocol

Only a idea !

Roland

[RAC_1]

Is /etc/ntp.keys in place and have appropriate perms?

[cl_5]

Yes, /etc/services is correct. But there's no /etc/ntp.keys. No client in our network has a ntp.keys file and they running xntpd anyway.

[RolandH]

Hi Carsten,

I have attached how I have configured my DCF-77 Clock.

Sorry the document is in german, but I think that is Ok for you

HTH
Roland

[Ron Kinner]

If you can telnet to the port and it is not refused then some process is listening on the port so it is in use even if there is no current TCP/IP connection.

netstat -an |grep 123

should show it listening.

Ron

[cl_5]

Hi forumers!

Everything is right with my configuration. And no port was occupied. The xntpd is running now. I think it was the executable /usr/sbin/xntpd. Maybe corrupt.

I've copied a xntpd from the second cluster node to tmp and execute without any error.

Now the question is how to repair xntp on this system without injuring the package management.

Thanks for your help

Best regards
Carsten

[Elmar P. Kolkman]

I would suggest a swremove and then an swinstall of the package containing the xntpd. Or a reinstall. Remember that there could also be some patches that are installed and needed, though they could also be the cause of the problem.

[Rgomes]

Hi Carsten and all,

FYI, we had the same problem of xntpd on 11i system. We also got "Address already in use" in syslogWe have overcome it by removing the latest ARPA patch PHNE_28089 and by installing the previous one PHNE_25664. PHNE_28089 has some tcp/udp open port problem.

best regards,
Richard