Details
-
Bug
-
Resolution: Done
-
Major - P3
-
None
-
2.4.2
-
None
-
None
-
ALL
Description
Relevant piece of code:
if ( secondaryThrottle && numDeleted > 0 ) {
|
if ( ! waitForReplication( c.getLastOp(), 2, 60 /* seconds to wait */ ) ) {
|
warning() << "replication to secondaries for removeRange at least 60 seconds behind" << endl;
|
}
|
millisWaitingForReplication += secondaryThrottleTime.millis();
|
}
|
Note that 2 (num of secondaries to wait) is always passed to waitForReplication. This can cause the delete to needlessly wait for 60 seconds if there are less than 2 secondaries (and worse, if the node is not a member of a replica set at all!)
Attachments
Issue Links
- is related to
-
SERVER-14708 A mixed single node & replicaset set sharded cluster can wait for replication on the single node when moving chunks, if secondaryThrottle is enabled
-
- Closed
-