I have an instance where another group of people want to receive an exact copy of data that is indexed on my main Splunk server. I've modified the outputs.conf on my server and the second server receives the data and indexes accordingly.
I've come across a problem, though. My splunk environment is production, with a large number of sources being indexed on my server. The second server, though, has the tendency of being rebooted, services stopped, etc. This wreaks havoc in my production environment - when my server cannot forward it shuts down its listening ports, which causes all my forwarders to shut off their listening ports, and so on and so on.
Is there a way that I can tell my production system that if system #2 is down and forwarding is stopped to continue listening and indexing data?
Well, you're not going to have an exact copy if you do this, and there's no easy way to get them synced back up, but you can set dropEventsOnQueueFull and maxQueueSize in outputs.conf to control how much and how long the primary indexer holds and waits for the secondary.