Hp rp3440 rebooting every month

[Robert Peregrin]

One of our server is rebooting every month:
reboot system boot Mon Aug 23 04:28 still logged in
reboot system boot Thu Jul 29 08:49 - 04:28 (24+19:39)
reboot system boot Thu Jul 29 08:22 - 08:49 (00:27)

Found the following repeating alarms in the syslog and mail.log before the reboot:

Aug 23 02:53:00 s7s214 above message repeats 7 times
Aug 23 02:53:01 s7s214 tftpd[1024]: Access from 172.30.58.248 denied.
Aug 23 02:53:10 s7s214 tftpd[1024]: Access from 172.30.58.248 denied.

Aug 23 01:27:00 s7s214 sendmail[12511]: unable to qualify my own domain name (s7s214) -- using short name
Aug 23 01:27:00 s7s214 sendmail[12511]: be sure to set domain name in Dj macro in sendmail.cf
Aug 23 01:27:00 s7s214 sendmail[12511]: BAA12511: from=root, size=256, class=0, pri=30256, nrcpts=1, msgid=<201008222227.BAA12511@s7s214>, relay=root@localho


Then the server rebooted:
Aug 23 03:12:00 s7s214 sendmail[16745]: DAA16743: to=root, ctladdr=root (0/3), delay=00:00:00, xdelay=00:00:00, mailer=local, stat=Sent
Aug 23 04:29:26 s7s214 /sbin/init.d/sendmail[1145]: #### rebooted ####
Aug 23 04:29:26 s7s214 sendmail[1158]: unable to qualify my own domain name (s7s214) -- using short name

I need some help on finding out what is causing the server to reboot every month.

Regards,

Robert

[Matti_Kurkela]

Sendmail and tftpd log messages aren't normally related to rebooting at all

TFTP is normally used for network boot, network switch firmware upgrades and the like. Is this system supposed to be a netboot server? (Is it an Ignite-UX server, for example?)

Sendmail's complaints about name resolution indicate your DNS domain name might not be properly set. (Look into /etc/hosts and /etc/resolv.conf.) But this definitely won't cause the system to reboot.

Look into /etc/shutdownlog: if the system was rebooted in a controlled fashion (i.e. someone commanded it to reboot), the shutdown/reboot command should have written a log entry about it.

If there is no entry in /etc/shutdownlog, your system probably lost power, crashed or was shut down abruptly in some other manner.
Look into /var/adm/crash: any crash dumps in there? If you find crash dump files, do their modification times match the reboot times?

Also check the MP log: get on the system console (via serial port, or through the network console if it has been configured). If the MP is already in console mode, press Ctrl-B to access MP. The default MP login/password is Admin/Admin. In the MP main menu, the command "SL" displays the MP logs. You'll want the error log, and in Text mode for maximum human-readability (the default is Keyword mode).

The MP log should clearly indicate if the system lost power, crashed, overheated etc.

MK

[Pete Randall]

Start with "crontab -l" to see if there is a scheduled reboot. Check the /etc/shutdown.allow file to see if any user other than root can shutdown. If so, check their cron files as well.

I'm not sure which logs you're displaying, does /etc/shutdownlog offer any clues?


Pete

[madhuchakkaravarthy]

hi
!r u sure that server is rebooting once in a month.if so check user allowed to run cron jobs and check /etc/shutdownlog and check in syslog do any one logged as root somemight may be aware of root pwd.

!regards
Mc

[Prashanth.D.S]

Hi Robert,

If you say that its happening every month then i would defntly check the cron jobs.

Also check if there was any crash dump under /var/adm/crash/ with latest timestamp.

And check the shutdown logs, what message do you see for the latest reboot.

Best Regards,
Prashanth

[djoshi]

As per you the system is rebooting every month.... if it is always at the same time stamp.... the you should look cron first.

Secondly, faulty CPU would be the cause of rebooting as well. if you have support of that server, I would prefer check with your vendor.

[Viktor Balogh]

> As per you the system is rebooting every
> month.... if it is always at the same time
>stamp.... the you should look cron first.

It could also be the sign of the cleaning staff, doing a regular monthly cleaning of the datacenter... :) (just joking)

[Robert Peregrin]

Hi,

There was no file under /var/adm/crash. last entry in /etc/shutdownlog is for 2007. There is no cron job running every month to reboot the server.

Unfortunately, the lan1 for the Management IP is not configured when I checked it from SAM. But I also did not see a way to add the lan1 interface and its IP address from SAM. Can someone help me with the procedure to add the lan1 interface.


Regards,

Robert

[Jim Walls]

The console LAN connection is not configurable from HP-UX; The console is controlled from a separate management computer that has its own RS232 and LAN ports. These ports (physical sockets) are located at the rear of the server cabinet.

You need to either connect a serial console device (or suitably configured client application on a PC or laptop) to connect to one or the other. If you connect directly to the LAN port you need to use a cross-over cable - otherwise use an network switch or hub. And you need to know what IP address may have been used previously (unfortunately, I do not know what the default is)