But I am leveraging the tags system where i get multiple filebeat.conf files down.
But because it starts the process with “default” --path.data & --path.logs then it can’t start another process.
I simply get:
Exiting: /var/lib/filebeat/filebeat.lock: data path already locked by another beat. Please make sure that multiple beats are not sharing the same data path (path.data)
on my 2nd config file.
My log collector config looks like this:
it’s funny, because my graylog setup is brand new.
I tried to do migration from old collector to new, but didn’t get it to work and then i waited and got an opportunity to deploy brand new setup.
my configs does have the ${sidecar.spoolDir} in data and logs.
But clearly the runtime of filebeat (in my case) doesn’t updated variables from the config file.
I believe (without knowing filebeat), that if you don’t define --path.data for example to runtime, then filebeat will just read the paths from the config file.
But the variables are not defined anywhere in my configs, so it must come straight from go code or somewhere inside graylog server.
I can see the correct data and log paths are used:
I can also see i have 2 unique instances of filebeat running at once (one is the test i’m doing here the other is my existing config unrelated to this test):