The plan is to scale up a current distributed search framework — from one search head (SH) + one indexer to one SH + two indexers.
We are not planning to use an indexer cluster, so each indexer will have different indexes.
Will the SH be smart enough to search only one indexer based on a query?
if not, I assume SH will go through all indexers for searches and return results from all of them.
Will this impact query performances?
Each Indexer just needs to be a search peer of the search head and it will distributed the search to both Indexers. No performance impact on asking for something that doesn't exist on an Indexer, you'll be fine with that.
I'd go for clustering anyway, even if with RF=SF=1 and so the cluster master wouldn't be impacting that at all and you'd be ready for future improvements. But again, you'll be fine with that Architecture too
Each Indexer just needs to be a search peer of the search head and it will distributed the search to both Indexers. No performance impact on asking for something that doesn't exist on an Indexer, you'll be fine with that.
I'd go for clustering anyway, even if with RF=SF=1 and so the cluster master wouldn't be impacting that at all and you'd be ready for future improvements. But again, you'll be fine with that Architecture too