Page 3 of 4
Re: Linux SNMP: Process name table No response from remote h
Posted: Mon Oct 07, 2013 3:03 pm
by msbensonstk
We are having this problem on about 10% of our systems. The rest it works just fine, so I don't believe the plug-in is completely to blame. If we could get the plug-in to respect the extended timeout values I am fairly confident that would fix the issue, but as I demonstrated earlier in the ticket, that part of the script appears to not be working correctly.
Re: Linux SNMP: Process name table No response from remote h
Posted: Tue Oct 08, 2013 12:36 pm
by sreinhardt
On the note of those 10% systems, are they on a separate vlan, another datacenter/location, or anything else of that nature that would introduce latency. Also is this a continual issue with all of those systems or just sporadic? I would tend to agree that the timeout symptoms seem to fit your issue, I will see about that portion of the code and how the flag is handled.
Re: Linux SNMP: Process name table No response from remote h
Posted: Wed Oct 09, 2013 8:33 am
by msbensonstk
They are in the same VLAN and security zone (and physical switches) as the rest of the servers that are working normally. There is no difference between them security-wise or latency-wise.
The issue on them is sporadic, but down is the 'normal' state. They sporadicly are successful, but never twice in a row.
Re: Linux SNMP: Process name table No response from remote h
Posted: Wed Oct 09, 2013 4:57 pm
by abrist
Have you tried changing the timeout directly in the script?
At line #60, change:
Code: Select all
my $o_timeout= 5; # Default 5s Timeout
To:
Code: Select all
my $o_timeout= 20; # Default 5s Timeout
Re: Linux SNMP: Process name table No response from remote h
Posted: Tue Oct 15, 2013 10:10 am
by msbensonstk
Setting the timeout in the script appears to have no effect either:
vim /usr/local/nagios/libexec/check_snmp_process_wizard.pl
my $o_timeout = 50
date; /usr/local/nagios/libexec/check_snmp_process_wizard.pl -H <host> --login=nagiosxi --passwd=<pass> --privpass=<pass> --protocols=sha,aes -n '/usr/local/contego/ContegoSPOP/' -f -w '0,1' -c '0,1'; date
Tue Oct 15 10:08:29 CDT 2013
ERROR: Alarm signal (Nagios time-out)
Tue Oct 15 10:08:44 CDT 2013
Script still stopping after 15 seconds.
Re: Linux SNMP: Process name table No response from remote h
Posted: Tue Oct 15, 2013 11:10 am
by abrist
Out of curiosity, have you attempted to run the check without the fullpath?
Code: Select all
date; /usr/local/nagios/libexec/check_snmp_process_wizard.pl -H <host> --login=nagiosxi --passwd=<pass> --privpass=<pass> --protocols=sha,aes -n 'ContegoSPOP' -w '0,1' -c '0,1'; date
Is ContegoSPOP the actual process name you wish to monitor?
Re: Linux SNMP: Process name table No response from remote h
Posted: Tue Oct 15, 2013 4:25 pm
by msbensonstk
Yes. It's the process name as reported by the process table (this is the same check that works on other systems).
Re: Linux SNMP: Process name table No response from remote h
Posted: Tue Oct 15, 2013 5:07 pm
by abrist
Did the command work without the full path switch?
Re: Linux SNMP: Process name table No response from remote h
Posted: Wed Oct 16, 2013 1:59 pm
by msbensonstk
Sorry, you'll have to clarify that a bit. Can you write out the command with the 'switch' you are referring to?
Re: Linux SNMP: Process name table No response from remote h
Posted: Wed Oct 16, 2013 3:51 pm
by sreinhardt
your -n flag is always "-n '/usr/local/contego/ContegoSPOP/'" abrist was asking if you could run it successfully without the full path, like such:
Code: Select all
date; /usr/local/nagios/libexec/check_snmp_process_wizard.pl -H <host> --login=nagiosxi --passwd=<pass> --privpass=<pass> --protocols=sha,aes -n 'ContegoSPOP' -w '0,1' -c '0,1'; date