In the picture, which is referred in the [[master1_fqdn]] and [[master2_fqdn]] on the setup part? Is master1 “Master B” and master2 “Master A2” ?
I assume Master A2 is the Icinga2-corp one ?
Not at all.
On of my infrastructure as 2 master (master A1 and masterA2) wich supervise master B
So when I run the node setup I specify the 2 host. If you have only one master you just set this one.
So in my context « icinga2-corp » on master B is the name of the second instance which run icinga2 as agent which report to master A1 and masterA2
1 Like
Regarding the second instance icinga2-corp, is that just the same complete installation as the the first one with separate DB etc ?
you don’t need to install anything else, all the necessary binaries are in place (if your server is already configured like icinga master).
You just have to create a second systemd service (call icinga2-corp in my case) and start this service with the agent configuration (which is generated by running icinga2 node wizard with all the parameter linked to the file of that second.
Do you try to follow the steps describe in my previous post ?
If yes what happen on your server ?
NB : If you can create a snapshot before playing in case of emergency
Yes, I did follow but it fails in the last step to actyually start the service icinga2-corp saying that some file (or path) is missing… I reverted the snapshot to start over so I don’t have the actual error right now, but will do it over again and see if I just missed or misspelled some foldername or similar… Thanks for all the help so far!
Just rerunned the setup and there was indeed just a small error in the config for the service… So it works now! Thanks!
1 Like
Just a small update… I’m using Dashing (GitHub - mocdaniel/dashing-icinga2: Dashing dashboard for Icinga 2 using the REST API) that’s using Icinga API to get information on hosts/services. It seems to be failing some API requests if I have the second instance (icinga-2-corp) running and that results in some cases doesn’t show alerts. If I stop the second instance it works perfectly. Does the second instance somehow responds to API requests or what am I missing?
On witch port your second instance is listening ?
If you follow my previous post, you changed the default one (5665) to 5664 so there is no reason for there to be any interference…
1 Like
You are again completly correct, the second API bind port was not set so it listened to 5665 DOH 
. Thanks for pointing that out to me!
no problem 
have a good day