I just deployed Splunk in an indexer cluster deployment, and I've noticed that my indexers have a different number of buckets. Shouldn't they have the same number of buckets since the data is replicated? Or not all buckets get replicated between indexers?
↧