I have an application which writes the same log to two servers. They act as each other's backup during maintenance schedules. We plan to run a forwarder on both the log servers. Is it possible to configure a Splunk indexer in such a way that it will index the events from only one source & ignore the other? This will ensure that when one of the servers is under maintenance, events from the 2nd server keep getting processed, ensuring no interruption to Splunk monitoring.
There is a crazy solution I can think of :
Lets say we have 2 servers : Server 1 and Server 2, both having cloned logs.
Create a script in Server 2 to the outputs.conf to start forwarding when Server 1 is under maintenance(no heartbeat), as soon as you get a heartbeat for Server1 the script will in the outputs.conf in Server2.
So this script should reside on Server 2 and should have the below capabilities:
1. Should be able to check heartbeat of Server1
2. Should be able to make edit outputs.conf of Server2
smart solution 🙂 depending on the requirements the split brain situation (communication loss between the two servers) has to taken into account.
why do you wane filter out this events at index time, not just during search time be removing duplicates. You would need some logic to decide from which node you have to disable the events.