check_snmp_cisco improved version

Project Notes

The plugin is a enhanced and modified version of check_cisco.pl from R3dL!GhT.

19.05.2009 - New version. Corrects a minor bug. In case the cache file was deleted a new one was not generated automatically. Fixed. Now - in case the cached indexfile is wrong (for example you have reconfigured the cisco switch) delete the file. The next time the script runs the cachefile will be generated.

15.07.2008 - First release of the improved version.

The changes are mainly related on performance. The script was rewritten in several parts and the code was cleaned up.

On bigger switches the plugin was too slow due to the fact that it did multiple snmpwalks. On a switch with over 350 ports that caused a timeout in Nagios.

The problem was solved by implementing a cache mechanism. This was done in the following manner.

You need a directory for the cache files. It is best to do it with a tmpfs because a tmpfs is
opposite to a ramdisk swapable.

Sample entry from /etc/fstab:

tmpfs /var/nagios_plugin_cache tmpfs defaults 0 0

Please change variable $CachePath for a different location

Unforunately I use some systemcalls for executing system commands. This is a little bit dirty
but every snmp call in this script was done be a system call instead of using perl. A complete
rewrite was too much work for the time I had to solve my problems.

The function snmpwalkgrep, snmpwalk and some lines of code were kicked out because they were
no longer necessary

The script does a find in the cache directory for the cache file. If it is too old (-mtime) or
not existent it will be generated. OIDs, interface names and descriptions are store in the cache
That means the double walk will be done only every 2 days instead of every run

Synopsis:

check_cisco_snmp -H ip -C community -I interface -S state (optional)