Hi, i have a replica set with 2 inMemory (primary and secondary) and 1 WiredTiger (Hidden) and i was expecting a huge performance impact without indexing all queries with this setup because of the in-memory engine but when i compare my old whole WiredTiger setup with this there is no better performance with in-memory engine, do you think that i’m missing something?
If this is a read-only workload then you might not see significant differences if those reads are already in memory. You might see a greater impact on writes. If that’s a wrong assumption, then version and environment information might be needed. Meanwhile, though, this webinar might be of interest as it covers use cases: