Nagios Support Forum

Just curious regarding the use of mod gearman and checking for mrtg data. The service checks is checking for rrd information in /var/lib/mrtg. How/If does mod gearman handle the checks for bandwidth data? or does network bandwidth can only be done with the XI server?

Since checks do not always go to the same worker, you would need to run those locally or ensure that MRTG is running pretty much in lock-step on each of your gearman workers so the RRD data is the same. That, or create a single hostgroup for all MRTG checks and send them all to one worker.

Personally I would keep them local to XI since other components in the system like reports and graphing will need those MRTG RRDs.

Thanks. Since mod gearman is used to help with lightening the work load of XI, what's the best way to make sure that XI isn't overloaded with mrtg-related checks?

OptimusB wrote:Thanks. Since mod gearman is used to help with lightening the work load of XI, what's the best way to make sure that XI isn't overloaded with mrtg-related checks?

Only do those checks that are needed?
Otherwise, if mrtg becomes too resource intensive for the XI server, you will need to offload it. This usually means creating a share for /etc/mrtg/* and /var/lib/mrtg/*.
And then moving the mrtg cron to another box and mounting those directories on it (as well as the XI server).
This allows the remote server to access the configs and mrtg rrds necessary for the cron to run, and also allows the XI server to access the rrds for check_rrdtraf checks and the /etc/mrtg directory for editing/creating the mrtg configs.
Does that make sense?