The failure in https://jira.mongodb.org/browse/BF-4726 occurred when executing a readConcern=majority read on a document that was written with writeConcern=majority. Thus, it appears that the failure was likely caused by a mongod taking a long amount of time to read a document from local storage (not making any network requests to other nodes).

The functions `doDirtyRead` and `doCommittedRead` should have the `maxTimeMS` values in their find commands increased to 10 minutes, so as to account for any potentially slow test machines.