Hi Splunkers,
Question about replication factors and search factor in cluster environment.
If I have 8 indexers and 8 search heads with search and replication factors of 2 each, would I benefit in search speed increase by setting replication and/or search factors higher?
How much higher? What are the recommended replication/search factors for number of indexers, search heads?
Currently we are running into issues where search queries that take long time to execute, >10min, get reading errors while waiting for peers and Search results are incomplete.
Thank you for suggestions
Hello
Rep factor and search factor will define only "how many" indexers can you lose without impact to the service. With rep factor and search factor 2, you can lose only 1 of the 8 indexers, Increasing it to 3, will allow to lose 2 indexers before affecting the service.
So, you won´t get any performance gain increasing those settings. That´s because there is only one primary bucket of each copy, and that indexer is the one returning the results.
Regards
Hello
Rep factor and search factor will define only "how many" indexers can you lose without impact to the service. With rep factor and search factor 2, you can lose only 1 of the 8 indexers, Increasing it to 3, will allow to lose 2 indexers before affecting the service.
So, you won´t get any performance gain increasing those settings. That´s because there is only one primary bucket of each copy, and that indexer is the one returning the results.
Regards