Deployment Architecture

Possible to do failover rather than load balancing with the forwarder?

gryz
Explorer

Given our network architecture, we would like to do failover rather than load balancing.

We would want a primary and a secondary indexer.

Under normal conditions, all forwarders should sent data to the primary indexer.

If the primary indexer goes down, data should then go to the secondary indexer.

When the primary indexer comes back up, data should go to the primary again.

Thanks!

Tags (1)

bandit
Motivator
0 Karma

mcronkrite
Splunk Employee
Splunk Employee
0 Karma

sowings
Splunk Employee
Splunk Employee

To the best of my knowledge, reading over docs for outputs.conf, I don't think what you're looking for is possible. You could try changing the autoLBFrequency to a high value, but that just means that the forwarder would develop a long-lasting affinity for whatever indexer it's talking to. In a failover event, it would hang on to its connection to the secondary for a long time.

Let's dig deeper, however. I'm curious to know why you'd want data to go only to a primary indexer, and not both. You'll get the redundancy you request, as well as increased search performance if you utilize both indexers together. Is there some reason why you wouldn't use both? Is it simply a matter of bandwidth? If so, note that you can enable compression on both sides of the connection (forwarder and indexer) to decrease overall bandwidth usage.

0 Karma
Register for .conf21 Now! Go Vegas or Go Virtual!

How will you .conf21? You decide! Go in-person in Las Vegas, 10/18-10/21, or go online with .conf21 Virtual, 10/19-10/20.