We are trying to Configure Forwarding using Load balancing and data cloning.
By making these two entries at forwarder.
Data will be cloned between two sets of indexers and load balanced amongst the individual sets.
I want to know how is this handling at searcher level and how does a search know which set to use during fail over.
I have read about whiteList and BlackList but that does not seem to be relevant to this scenario.
For the architecture you describe, there is currently no way to force a search head to automatically switch over to the second set of indexers. This is typically a manual process, where you only need to change the distsearch.conf file.
Something to consider, is leveraging the index and forward method. This will allow you to keep data consistency across each set of indexers.
It's a good question and something I also would like a definitive answer to.
As far as I am aware , you could set up your 2 sets of Indexers in distsearch.conf.
Lets call Indexer Set A your "Active" set and Indexer set B your "Failover" set.
You could use the "servers" property of the distributedSearch stanza to declare the peers of your Active set and the "disabled_servers" property to declare the peers in your Failover set.
But I don't see how an automatic switch could be configured, so in a DR/Failover scenario, such as 1 node within a set goes down or the entire set goes down, it looks as if it would require a manual intervention to switch distributed search from the Active set to the DR set.
In lieu of having the facility to automatically do this set switch in distsearch.conf , it wouldn't be all that tricky to rattle up an automated solution using the REST API ie: have a script that checks the status of the peers in the active set and if the set has 1..n failures, change to the DR set via a REST call, and then back to the Active set when it is restored.