CernVM-FS Nagios Check

Description

We suggest the following Nagios check for CernVM-FS clients. The check requires CernVM-FS 2.0. Under heavy load, it might take a little while (perhaps up to 30 seconds) until all information is gathered.  By itself however, it puts only little load on cvmfs. It works as follows:
check_cvmfs <repository> [version], like
check_cvmfs atlas.cern.ch 2.0.0

If /cvmfs/atlas.cern.ch is not available: CRITICAL
If /cvmfs/atlas.cern.ch/.cvmfsdirtab is not readable: CRITICAL
If the currently active host/proxy is unavailable: CRITICAL

If there have been I/O errors recorded by cvmfs: WARNING
If the no. open file descriptors is greater than 80% of the max: WARNING
If the resident memory is more than 50M: WARNING
If the no. loaded nested catalogs is greater than 750: WARNING
If the installed version and the loaded version mismatch: WARNING
If the installed version and the expected version mismatch: WARNING
If the cvmfs cache partition is filled more than 95%: WARNING

Note: If the Nagios test runs more frequently than the maximum idle time defined by automount, the cvmfs volume will never be unmounted.

Source Code (Bash)

Source code is available in the downloads section.