inetd and syslogd

[tigm7103105q1]

Hi!, I'm Marcos and I need your help.
Case 1:
I've a HP-SD3200 with HPUX11.i, and I want to know how can get register's about inetd daemon when sombody kill it.
Case 2:
Why is there many message "restart" in "/var/adm/syslog/syslog.log" file like these?
Dec 4 09:21:25 adu820 syslogd: restart
Dec 4 10:21:39 adu820 syslogd: restart
Dec 4 11:21:40 adu820 syslogd: restart

Thank's.

[Steven E. Protter]

inetd -k stopes the inet deaemon

inetd -c restarts in place

inetd -l toggles logging.

The restart message is because for some reason the syslogd daemon is being restarted.

This could be due to strange settings in /etc/inetd.conf

It could be due to putting that daemon in /etc/inittab

Check it out, post inetd.conf

SEP

[Paula J Frazer-Campbell]

Hi

Notice the restart times each one hour apart - Looks like the system is restarting it and not a user.


Paula

[Paula J Frazer-Campbell]

As Stephen suggests check inittab.

Paula

[Fabio Ettore · Accepted solution]

Hi Marcos,

it already happens (with just inetd running too, that is without -l option)!
When inetd is killed, a message appears into /var/adm/syslog.log:
I have just tested it on my system:

- open a first shell:
# tail -f /var/adm/syslog/syslog.log

- open a second shell:
# inetd -k

In the first shell you will see something as

Dec 4 18:36:28 pippo inetd[5205]: Going down on signal 15

Then you will know that inetd was killed!

Anyway inetd -l to restart will trace IP addresses and/or hostname which try a connection. Just attention about this -l option; it could fill /var/adm/syslog/syslog.log quickly. Otherwise just type

# inetd

to restart.

About 'syslogd: restart' I don't know valid reasons or known problems about it anyway I'd guess that you can ignore that message but other replies about this could be more useful.

I hope this helps you.

Best regards,
Ettore

[Elmar P. Kolkman]

I think the inetd problem is covered, but as for the syslogd: check out your crontab's ('crontab -l' when logged in as root) for commands started every hour at about 20 minutes past the whole hour. It will look something like:
20 * * * *
(the 20 could also be 21, depending on the command being run)

I think there is a script that is getting called that restarts the log file after clipping the header of the file.