Snmptt trap handling limit.

[prak2010]

Hi All,
I was testing the trap handling limit of snmptt.
Setup:
snmtt: snmptt_1.4
mysql: mysql Ver 14.14 Distrib 5.1.71
net-snmp : net-snmp-5.5

I sent 200 traps within a minute from a script to my setup .
Surprising only first 80 traps came in sequence and after that it starts skipping the traps. The total traps I received is only about 130.

Can any body tell me that
Is 80 traps in a minute is the trap handling limit of snmptt.


Thank You in advance.

Prakash Ranjan

[Box293]

Is the device sending the traps and your nagios server in the same subnet or does it have to go through a router / firewall?

If it is going through a router / firewall then it may have a limit imposed that is causing traps to be dropped.

[prak2010]

Hi Box293,
First of all thank you for your response.
Regarding your question, The device and nagios server is in the same subnet.
Also. on the machine where nagios is installed, I run wireshark and able to see all the traps send by the device is reached on nagios server machine.
So the packet drops happens on between snmptt and snmptrad application.

Flow of the trap is as below.

Device --------> snmptrapd (Trap is dropping here)---------> snmptt ------nagios and mysql and log_file.


Thank You
Prakash

[tmcdonald]

Please run the following command from the Nagios server command line as root and post the results:

Code: Select all

ls -l /var/spool/snmptt | wc -l
service snmptt status

[sreinhardt]

I do not know of a hard limit in snmptrapd or snmptt. I would suggest stopping the snmptt service but leaving snmptrapd running. Send in your traps and check the /var/spool/snmptt folder for all the traps that should be coming in. Once that is verified to be correct, start the snmptt service and see if all are reaped properly. We can also take the path of enabling debug logging and see why and if it is actually skipping things within snmptt, as it should not be.

[prak2010]

HI sreinhardt,

I checked this option. application snmptrapd is able to handle the traps more than 1000 in one minute. but as soon as I am attaching the application with snmptt application. both the application starts skipping the trap.
I got one information from the snmptrapd man page as
" The daemon blocks on the executing traphandle commands. (This should be fixed in the future with an appropriate signal catch and wait() combination). "

http://www.tin.org/bin/man.cgi?section= ... trapd.conf

I guess this is the reason of skipping the traps by snmptrapd application.

Prakash Ranjan

[sreinhardt]

Could you post your current snmptrapd.conf and snmptt.conf please? I'd like to have a look, and maybe I can provide some suggestions to help speed it up. One thing I would note, is that snmptrapd should never be set to call snmptt, the new way of using spooled files is far more efficient and resolves any fork\call issues that were present before.