Axon 4.0.3 +JPA + Aurora serverless Token entry table is blocked

Imported from Google Groups
1

HI Team,

We have issue from Consumer side where TEP is implemented.

Application Yml:
axon:
eventhandling:
processors:
“[kafkaProcessor]”:
source: kafkaMessageSource
mode: TRACKING
threadCount: 1
initialSegmentCount: 1
batchSize: 500

When we try to process the huge load kafka then i see that some the database connection opened and hanging there in sleep state, because this tracking table is locked and other Database threads can’t access the token entry table ( tracking token) .

2

i meant to say the connections are not released once it’s done with processing, so other thread not able to get access to the token entry table. is it a bug or do i need to change any configs?

3

The processor keeps a lock on the token entry table to prevent other threads from claiming it. That’s expected behavior. However, it shouldn’t lock the entire table.

Which DB do you use?

4

We are using AWS aurora serverless.

Sorry for the confusion. It’s not locking enter table. It locking only the row, but as it not closing next thread can’t get a hold on this row and not able to process the events from Kafka.

JFYI: We are using hikariCP and with jpa.

5

this issue happening only when we run the same axon consumer in multiple instances. Did anyone faced same issue? Is it config issue or bug in axon framework?