I’m currently using check_mk raw with Distributed Monitoring in a “hub and spoke” setup. One master server, and several servers underneath.
To help lighten the load on the main server, I’ve built a new server to act as a monitoring spoke, and need to move several hundred hosts over to that new server. The actual “moving” is working as expected, but the one negative is that I am losing perfrormance data as part of the move.
Is there a reliable way / script / command to move performance data from one instance of check_mk to another one so as to keep continuity and not lose data?
I don’t know if there is a script existing but it is fairly easy.
Inside your site copy the complete content from “~/var/pnp4nagios/perfdata” to the new empty instance.
You can do this at a running instance but will loose around 1-2 hours of performance data. If this is ok then you can move your hosts after the copy job is complete to the new site.
If you need all performance data then you need first to flush the rrdcached.
This can be done with
echo "FLUSHALL" | unixcat ~/tmprun/rrdcached.sock
Wait some time and then copy the content as before. Loose here is only some minutes of performance data.
Thanks for the update! Yeah, if I have a small gap of a couple of hours, I’m okay with that generally, since we’re checking OS variables that don’t change with that level of frequency (or if they have, we can look as past days for trends).