SERVER-46721 Secondary readers should read at the no-overlap time instead of lastApplied

The no-overlap time, ReadSource::kNoOverlap, is the minimum of replication's lastApplied timestamp
and WiredTiger's all_durable time. This time is independent of replication state and ensures
queries do not see oplog holes after state transitions from secondary to primary.

1 files changed, 0 insertions, 13 deletions

diff --git a/src/mongo/db/db_raii.h b/src/mongo/db/db_raii.h
index edf32cd0010..51208643c6a 100644
--- a/src/mongo/db/db_raii.h
+++ b/src/mongo/db/db_raii.h
@@ -133,19 +133,6 @@ private:
     // This field is optional, because the code to wait for majority committed snapshot needs to
     // release locks in order to block waiting
     boost::optional<AutoGetCollection> _autoColl;
-
-    // Returns true if we should read at the last applied timestamp instead of at "no" timestamp
-    // (i.e. reading with the "latest" snapshot reflecting all writes).  Reading at the last applied
-    // timestamp avoids reading in-flux data actively being written by the replication system.
-    bool _shouldReadAtLastAppliedTimestamp(OperationContext* opCtx,
-                                           const NamespaceString& nss,
-                                           repl::ReadConcernLevel readConcernLevel) const;
-
-    // Returns true if the minSnapshot causes conflicting catalog changes for either the provided
-    // lastAppliedTimestamp or the point-in-time snapshot of the RecoveryUnit on 'opCtx'.
-    bool _conflictingCatalogChanges(OperationContext* opCtx,
-                                    boost::optional<Timestamp> minSnapshot,
-                                    boost::optional<Timestamp> lastAppliedTimestamp) const;
 };
 
 /**