Run CheckMK 2 Server on AWS

[coofercat (Ralph Bolton)]

I’m running a CheckMK 2 server in AWS - it’s just a single server running on a VM. I’d like to be in a position that if the VM is destroyed, that I can recover it quickly. Some other systems I have use an EFS volume (ie. an NFS mounted disk), and store all their config files and run-time information on there. If the server is destroyed, the new one mounts the disk, installs the software and starts up as if nothing really happened.

Just to be clear: I’m looking to move to an NFS mount all the files that can vary because someone added a new host to be monitored, or fiddled with the settings in the UI. Things that come with the system (eg. the checkmk package contents, my Apache certs and configs, etc) don’t count.

I looked around and found suggestion that /opt/omd was the location of most of the files that CheckMK writes during operation. I had a look in there, and it seems to have a lot of code in there (in /opt/omd/versions) - which ideally I wouldn’t be storing on EFS.

Is there a tidy way to do what I’m aiming at here, or do I have to take a load of code along with the configs? Are all the configs in /opt/omd or are there other locations too?

Any help much appreciated

[andreas-doehler (Andreas)]

The config and history data is inside the folder /opt/omd/sites there you have for every created site one folder.
The docker container does it the same way is it mounts the persistent volume to the folder /omd/sites.
If you use NFS you have to pay attention for the user rights as every site uses it’s own user and all the files inside the site folder must belong to this user.

[coofercat (Ralph Bolton)]

Ahh, okay - that makes sense, and I can solve it with just one directory.

You make a good point about UIDs - I didn’t really pay attention this time around, but will do a destroy/rebuild so get it right then. Thanks for your help