When the test function ReplSetTest.checkDBHashesForReplSet() detects inconsistencies between nodes in a replica set, the output can get very verbose. It would be helpful to get a hint on where to look within the output, especially where the dbhash/collection stats differ.

Current output:

FiveNodeDoubleRollbackTest, the primary and secondary have different stats for the collection createIndexes_0.test:

Proposed output (using differences in number of indexes as an example):

FiveNodeDoubleRollbackTest, the primary and secondary have different stats for the collection createIndexes_0.test: indexes