cpqthreshold_mib

[Mohamed K Ahmed]

Dear all,
I came this morning to be told that there was problems connecting toone of my servers by others.
The server is an ES40 running Tru64 UNIX V 5.1 PK#5.
Yesterday, some users lost connection to th system, and I can't find anything except on the console log (BTW, Where are the console messages stored???)
The console read as follows:

cpqthresh_mib[651]: cpqthresh_mib Warning esnmp_init.c line 3441: recv returned, Successful.
cpqthresh_mib[651]: cpqthresh_mib ""Error cpqThresholdMgmt_mib.c line 377: esnmp error: -4 lost connection with Master Agent, restart.
cpqthresh_mib[651]: cpqthresh_mib ""Error esnmp_init.c line 3318: Could not connect to master: connection refused.
cpqthresh_mib[651]: cpqthresh_mib ""Error cpqThresholdMgmt_mib.c line 377: esnmp error: -1 cannot send OPEN to Master Agent, restart after delay
mountd[58534]: startup
statd[59878]: startup
Initializing the threshold structure


Do these messages mean anything??
I researched it and found out that the cpqthresh_mib is kind of a threshold value for Compaq Insight, and since I don't use insight, I disabled it, but I want to know why dod I loose connection to the server?

Mohamed

[Joris Denayer]

Mohamed,

These messages from the snmpd and it's subagents are caused by something going wrong in your system.
If you don't use snmp, you can simply
# /sbin/init.d/snmpd stop
# /sbin/init.d/insightd stop

Of course, there is still something else happening on your system.
Investigate the usual suspects. CPU, Memory, Network (vmstat, netstat, collect, etc...)

Joris

[Mohamed K Ahmed]

Well, there is no any messages any where (dia or messages log) on anything problem with the system. However, there was a message on the console saying that the Legato Networker file is full and then these messages started coming.
I don't use Legato Networker, it came with the preinstalled system and it used to get a line everyday saying that the Legato Licince is expired. What I think is this is logged in some log file with a max value, and when it is full, it gave a signal to the cpqthreshold_mib that there is an error.
The problem is that I don't know where is this file at, and also the console messages file. If you know, please let me know.

Mohamed

[Joris Denayer]

Mohamed,

Can you post us the exact Legato error in the console messages ?

BTW: I forgot to mention that most console messages, logged by the syslogd, are also copied in /var/adm/messages or in /var/adm/syslog.dated/current.


Joris

[Mohamed K Ahmed]

Ok,
Attached is the whole sequence of messages, and that is it. There is no any logs in the errorlog or messages file

Hope this have been seen before and someone can tell me why did it happen.

Mohamed

[Joris Denayer]

I do not see anything wrong in these messages.
I don't see where the "Networker file" is full.

It should be clear that the cpqtresh_mib messages are not the cause of the fact that noone could login on your system. It is the other way around. There must have been a resource shortage, that crippled the performance of your system. (f.i. a program that looped and consumed all CPU cycles, or not enough free memory or heavy swapping or network errors or .....) and due to this shortage, the snmp processes logged these messages.

If nothing is logged in kern.log or daemon.log (see /var/adm/syslog.dated/current) then very little can be done.


You can start collect during boot with following command.
# rcmgr set COLLECT_AUTORUN 1

Then you'll be able to verify resource problems.

Joris

[Ralf Puchner · Accepted solution]

the message given indicates a snmpd problem on startup and is not the cause of the other problems you have.

To deaktivate networker execute the command /sbin/init.d/NSRstartstop stop
and delete the link within /sbin/rc3.d

If networker is installed the log files can be found on /nsr/logs.

As Joris said, check process table for stuck processes.