Details
-
Bug
-
Resolution: Fixed
-
Major
-
7.1.2
-
0
-
KV 2023-2
Description
When a DCP Producer or Consumer hits the DCP noop idle timeout (no message received after N seconds), we log details of when the last message was seen as part of disconnecting - e.g. Consumer side:
WARNING 287: (Bucket) DCP (Consumer) eq_dcpq:replication:ns_1@node1->ns_1@node2:Bucket - Disconnecting because a message has not been received for the DCP idle timeout of 360s. Received last message (e.g. mutation/noop/StreamEnd) 361s ago. DCP noop interval is 1s.
|
|
However, we do not log any noop details on the other side of the connection when it is closed. This makes it difficult to determine the exact cause of a DCP noop timeout - did the other side fail to send a message, or was there some issue with the network?
Add additional logging whenever we destroy a connect to record this information.