Skip to content

Latest commit

 

History

History
95 lines (68 loc) · 3.13 KB

start_process.md

File metadata and controls

95 lines (68 loc) · 3.13 KB

======= Restart process - check_multi as event handler =======

Description

This is a short example for a very simple event handler. In Nagios this is a little bit more complicated ;).

The example makes use of the check_multi feature to execute child checks sequently.

  1. First check if a specific process is running
  2. Evaluate the result of check 1 and restart process if its not running
  3. Second check for the process, it checks the result of step 2. (If step 2 was not necessary, it's redundant :-))

Simple, isn't it?

Call / Parameters

check_multi -f start_process.cmd -s PROC=`<process>` -s ARGS=`<args>`
  • Parameters:
    • PROC - name of the process to monitor and restart
    • ARGS - arguments to start the process (and to distinguish it from others)

Download

The start_proc.cmd file is available in the check_multi package or here.

State evaluation

  • OK

    • if process is running

  • UNKNOWN:

    • none

  • WARNING

    • if process was NOT running and was successfully started by start_process.cmd

  • CRITICAL

    • if process was NOT running and start_process.cmd was not able to restart it.

Code: start_process.cmd

:::perl
#
# start_process.cmd
#
# (c) Matthias Flacke, 5.4.2008
# 
# starts process if not already started and checks result
#
# check_multi -f start_proc.cmd -s PROC=`<process>` -s ARGS=`<args>`
#
# 1. check for process
command [ proc_before ] = check_procs -c 1: -C "$PROC$" -a "$ARGS$"
#
# 2. start process in background
eeval [ start_proc ] = \
        ( $STATE_proc_before$ != 0 ) \
                ? ( system("$PROC$ $ARGS$ &") != 0 ) \
                        ? "- failed: $?" \
                        : "- done" \
                :  "- not necessary, already running"
#
# 3. check process again (ok, maybe redundant ;-))
command [ proc_after ] = check_procs -c 1: -C "$PROC$" -a "$ARGS$"
#
# 4. state evaluation
state [ OK       ] = proc_before == OK     && proc_after == OK
state [ WARNING  ] = proc_before != OK     && proc_after == OK
state [ CRITICAL ] = start_proc =~/failed/ || proc_after == CRITICAL

Example Output

  • nsca is not running - note the WARNING overall state:

    :::bash nagios@thinkpad:~> check_multi -f etc/start_process.cmd -s PROC=nsca -s ARGS="-c /usr/local/nagios/etc/nsca.cfg" WARNING - 3 plugins checked, 1 critical (proc_before), 2 ok [ 1] proc_before PROCS CRITICAL: 0 processes with command name 'nsca', args '-c /usr/local/nagios/etc/nsca.cfg' [ 2] start_proc - done [ 3] proc_after PROCS OK: 1 process with command name 'nsca', args '-c /usr/local/nagios/etc/nsca.cfg'

  • nsca is already running:

    :::bash nagios@thinkpad:~> check_multi -f etc/start_process.cmd -s PROC=nsca -s ARGS="-c /usr/local/nagios/etc/nsca.cfg" OK - 3 plugins checked, 3 ok [ 1] proc_before PROCS OK: 1 process with command name 'nsca', args '-c /usr/local/nagios/etc/nsca.cfg' [ 2] start_proc - not necessary, already running [ 3] proc_after PROCS OK: 1 process with command name 'nsca', args '-c /usr/local/nagios/etc/nsca.cfg'