2013-06-28 12:44:40 -04:00
|
|
|
Documention for Zabbix Scripts
|
|
|
|
|
|
|
|
|
|
Work-in-Progress
|
|
|
|
|
|
2013-06-28 12:55:28 -04:00
|
|
|
The agents provided in this bundle are intended to be used for Zabbix Trapper
|
|
|
|
|
agents, run on the server and batched together and sent as bulk. Documentation
|
|
|
|
|
covering how these work will be provided further in this documentation.
|
|
|
|
|
|
|
|
|
|
Suggestion installation path is to place the scripts/ directory into:
|
|
|
|
|
/etc/zabbix/scripts/
|
|
|
|
|
|
|
|
|
|
And then create a directory:
|
|
|
|
|
/etc/zabbix/trap.d/
|
|
|
|
|
|
|
|
|
|
If you need more granularity, you can make a trap.d/daily, trap.d/hourly, etc
|
|
|
|
|
as needed and adjust your crons to run the runtrap with the path accordingly.
|
|
|
|
|
|
|
|
|
|
Inside your trap.d path should be either symlinks directly back to the
|
|
|
|
|
scripts/trappername or a script that runs the trapper agent with neccessary
|
|
|
|
|
arguments as needed. You can number the links/scripts in the trap.d so they
|
|
|
|
|
execute in a specific order.
|
|
|
|
|
|
|
|
|
|
One thing to note, it's best to make sure everything in your trap.d is uniquely
|
|
|
|
|
named. The runtrap script is designed to prevent stalling trappers by killing
|
|
|
|
|
them after a number of execution attempts (default of 3). Also, when executing
|
|
|
|
|
runtrap from cron, it's best to always insure it's run by the full pathname so
|
|
|
|
|
it can kill itself accordingly when deadlocks are detected. This will allow
|
|
|
|
|
the next run to attempt running the trappers again.
|
|
|
|
|
|
