You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After running successfully for weeks at a time we sometimes get a Nullpointer in the DelegatingInvocableHandler at return new InvocationResult(result, replyTo, (Boolean)this.handlerReturnsMessage.get(handler)) when looking up one of the handlers.
For example when looking up the onA handler the "handlerReturnsMessage" map contains the onB handler and null so the lookup of onA returns null.
We have seen the lookup fail for both the onA and onB handlers, never both at once.
When this fails it usually only happens for one or two Kubernetes pods out of the bunch.
After restarting the failing pod the handlers can both consume again.
Spring Kafka version: 2.7.1
KAFKA_CONCURRENCY: 12
The number of topic partitions is higher than the number of running pods * concurrency.
Any ideas on why this happens?
The text was updated successfully, but these errors were encountered:
Discussed in #1947
Originally posted by hannybNdq September 20, 2021
Affects Version(s): \2.7.1
Nullpointer in DelegatingInvocableHandler, lookup in "handlerReturnsMessage" map fails.
Two handlers (annotated with @KafkaHandler) within the same class (annotated with @KafkaListener) listening for different payload types:
After running successfully for weeks at a time we sometimes get a Nullpointer in the
DelegatingInvocableHandler
atreturn new InvocationResult(result, replyTo, (Boolean)this.handlerReturnsMessage.get(handler))
when looking up one of the handlers.For example when looking up the
onA
handler the "handlerReturnsMessage" map contains theonB
handler and null so the lookup ofonA
returns null.We have seen the lookup fail for both the
onA
andonB
handlers, never both at once.When this fails it usually only happens for one or two Kubernetes pods out of the bunch.
After restarting the failing pod the handlers can both consume again.
Spring Kafka version: 2.7.1
KAFKA_CONCURRENCY: 12
The number of topic partitions is higher than the number of running pods * concurrency.
Any ideas on why this happens?
The text was updated successfully, but these errors were encountered: