Ricardo Oliveira
2009-08-24 14:31:40 UTC
Hi,
I'm using snmptrapd v5.1.2 on CentOS 5.2, and I'm having some issues with traps being ignored or dropped.
I have several devices logging to the server, so I'm concerned this could be a performance issue of snmptrapd.
Otherwise, it could very well be some kind of configuration issue on my part, and that's why I'm asking for your help.
My startup options are:
-Ls6 -p /var/run/snmptrapd.pid
My snmptrapd.conf has the following:
--- start ---
disableAuthorization yes
traphandle SNMPv2-SMI::enterprises.9.9.41.2.0.1 /usr/local/bin/TrapHandler
--- end ---
The TrapHandler is a C++ executable I made which basically receives a trap, forks and dumps the trap to a file, to try and return to the snmptrapd as fast as possible (before, I had a perl script, but it was way too slow).
The symptoms I have are:
- The trap is generated on the device
- The trap is received at the interface (seen by tcpdump)
- The traps never shows on any log file.
I tried:
- Adding -d to the options: successfully logs the number of bytes received from each host. The lost traps do NOT show on this log.
- Adding -D ALL to the options: snmptrapd inexplicably crashes/dies after a few seconds and logs nothing.
- Running snmptrapd logging to stdout in a shell: again, the traps never show on the log
I'm not sure how many traps per second this should be able to handle, but right now I consider I'm way bellow the number I will need to handle (I'm only receiving from less than 10 devices, although some generate a large amount of traps) and I'm concerned this daemon will not be enough for my requirements.
Can anyone give me any help/tips with this one?
Thanks in advance,
Ricardo Oliveira.
_________________________________________________________________
Show them the way! Add maps and directions to your party invites.
http://www.microsoft.com/windows/windowslive/products/events.aspx
I'm using snmptrapd v5.1.2 on CentOS 5.2, and I'm having some issues with traps being ignored or dropped.
I have several devices logging to the server, so I'm concerned this could be a performance issue of snmptrapd.
Otherwise, it could very well be some kind of configuration issue on my part, and that's why I'm asking for your help.
My startup options are:
-Ls6 -p /var/run/snmptrapd.pid
My snmptrapd.conf has the following:
--- start ---
disableAuthorization yes
traphandle SNMPv2-SMI::enterprises.9.9.41.2.0.1 /usr/local/bin/TrapHandler
--- end ---
The TrapHandler is a C++ executable I made which basically receives a trap, forks and dumps the trap to a file, to try and return to the snmptrapd as fast as possible (before, I had a perl script, but it was way too slow).
The symptoms I have are:
- The trap is generated on the device
- The trap is received at the interface (seen by tcpdump)
- The traps never shows on any log file.
I tried:
- Adding -d to the options: successfully logs the number of bytes received from each host. The lost traps do NOT show on this log.
- Adding -D ALL to the options: snmptrapd inexplicably crashes/dies after a few seconds and logs nothing.
- Running snmptrapd logging to stdout in a shell: again, the traps never show on the log
I'm not sure how many traps per second this should be able to handle, but right now I consider I'm way bellow the number I will need to handle (I'm only receiving from less than 10 devices, although some generate a large amount of traps) and I'm concerned this daemon will not be enough for my requirements.
Can anyone give me any help/tips with this one?
Thanks in advance,
Ricardo Oliveira.
_________________________________________________________________
Show them the way! Add maps and directions to your party invites.
http://www.microsoft.com/windows/windowslive/products/events.aspx