you should use a Load Balancer, if you have to distribute load between Indexers and manage fail over.
If you haven't a Load Banancer, you can configure your DNS to do this role: in other words, you should associate to one logical address (e.g. splunk_syslogs) the two IP addresses of Indexers, in this way, DNS can distribute load between both the Indexers in normal conditions and send them to the up Indexer when there are problems on the other.
On additional hint: if you can it should be better to separate syslogs receiving phase from indexing phase, pratically to use two Heavy Forwarders to receive syslogs insteda Indexers (always with a Load Balancer), and then they send logs to the Indexers that index them.
Why this: because Heavy Forwarders have less maintenance activities than Indexers so you lose less syslogs.
Ciao and Merry Christmas.
... View more