-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
Atlas Stream Processing Engine
-
ALL
-
200
-
None
-
None
-
None
-
None
-
None
-
None
-
None
related to https://jira.mongodb.org/browse/BF-39421
We should avoid throwing an uncaught ProducerEndClosed exception when failing to send an object to a writer thread's queue.
Lets add a try/catch here and other push() calls. ProducerEndClosed exceptions should be caught. In the catch block, read the SPStatus of the writer thread and rethrow an SPException using that SPStatus. Also change consumeLoop to ensure queue->closeConsumerEnd() is called after the consumerStatus is set.
- is related to
-
SERVER-110718 Insert only 1 document when running kinesis retry tests to avoid test flakiness
-
- Closed
-