It seems like from replication queries (during getmores) we can see an open tailing oplog cursor returning repeated segments of the oplog to clients.
- is duplicated by
-
SERVER-21985 WiredTiger range scans can see out of order keys
- Closed
- is related to
-
WT-2307 Internal page splits can corrupt cursor iteration
- Closed
- related to
-
SERVER-21930 Restart oplog query if oplog entries are not monotonically increasing
- Closed
-
SERVER-21962 Add testing hook for detecting anomalous log entries to resmoke.py
- Closed