topic CORRECTION - SNMP problems on a D350 (NOT A K420) in Operating System - HP-UX

CORRECTION - SNMP problems on a D350 (NOT A K420)

Cathy Squires — Wed, 03 Sep 2003 16:11:10 GMT

******Sorry my brain slipped a digit *****

After working like a champ for several years and having been rebuilt from the OS up in October 2002, last Friday one of our D350s decided to lock up hard requiring me to power down the machine to reboot it. When it was rebooted we noticed errors in the boot up, upon reviewing the /etc/rc.log I found the following:

Start SNMP Master Network Management daemon
Output from ???/sbin/rc2.d/S560SnmpMaster start???:
----------------------------------------------
/sbin/rc2.d/S560SnmpMaster[2]: 761 Bus error (coredump)
EXIT CODE: 138
/sbin/rc2.d/S560SnmpMaster started??? FAILED

Start SNMP HP-UNIX Network Management subagent
Output from ???/sbin/rc2.d/S565SnmpHpunix start???:
--------------------------------------------
Master agent not responding.
EXIT CODE: 255
???/sbin/rc2.d/S565SnmpHpunix start??? FAILED

Start SNMP MIB-2 Network Management subagent
Output from ???/sbin/rc2.d/S565SnmpMib2 start???:
--------------------------------
Master agent not responding.
EXIT CODE: 255
???/sbin/rc2.d/S565SnmpMib2 start??? FAILED

***************
When I ran /sbin/init.d/SnmpMaster start I got the following message
???Pid 1400 killed due to test modification or page I/O error???

This machine is locking up after just a couple of hours from the reboot. Is this a hardware problem with the D350, a software problem with the OS load or both???

Any suggestions on how I can handle this would be greatly appreciated.
tks
Cathy Squires

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

Pete Randall — Wed, 03 Sep 2003 16:18:21 GMT

Cathy,

My first thought lean toward either hardware or a patch issue. What version of HP-UX are you running?

Pete

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

Cathy Squires — Wed, 03 Sep 2003 16:25:30 GMT

I'm running HPUX 10.20, I'm more inclined to believe Hardware as none of the software has changed since we reloaded in October.

tks
ecs

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

Pete Randall — Wed, 03 Sep 2003 16:27:22 GMT

Cathy,

Me, too but I wanted to see what I could find. I'll let you know.

Pete

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

Cathy Squires — Wed, 03 Sep 2003 16:30:43 GMT

Thanks I'd appriciate ANY help you could give me.

ecs

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

Pete Randall — Wed, 03 Sep 2003 16:34:06 GMT

Cathy,

I'm turning up nothing for patches - a generic search on SNMP turns up tons of them but anything more specific comes back empty handed.

Have you check /var/adm/syslog/syslog.log and OLDsyslog.log, demsg output, any diagnostic logs, crash dumps (if any).

Do you have a hardware contract on this box?

Pete

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

Cathy Squires — Wed, 03 Sep 2003 16:54:19 GMT

Don't have a service contract (any more) but I do have an on site spare configured for the system. I'm bringing the bad machine up one more time to try to make sure I have a good copy of the data stored on the RAID (just incase I can't recover my link) and then we'll put the spare on line. I'll also look at those logs either while online or once they ship it back to me.

tks again
ecs

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

Steven E. Protter — Wed, 03 Sep 2003 17:00:18 GMT

The software integrity was probably affected by the failure that locked the system.

dmesg

Other hardware checks in order.

You should check the configuration files for snmp and consider re-installation or copying of the software from a working system.

SEP

Re: CORRECTION - SNMP problems on a D350 (NOT A K420)

doug mielke — Wed, 03 Sep 2003 17:53:23 GMT

Since the system has locked a few times, It might be time to run some file system checks. They will run automatically opon boot if the filesystem was not shutdown cleanly. But, if there is damage, it may take more than one pass to clean, even though the state was set to okay after the 1st pass.
( man fsck to determine your options)

Also, if fsck finds a 'broken' file, it will place it in that filesystems lost+found directory under a new name, it's inode number.
If you have any files there, it means they are missing from the system. If you can determine what they are, you can sometimes simply move them back.