This repository has been archived by the owner on Dec 14, 2022. It is now read-only.
[BUG] [PULSAR-12693] 1.13 FlinkPulsarSink couldn't complete checkpoint when kill one of the Pulsar Cluster using kill -9 #465
Labels
Describe the bug
Original Report : apache/pulsar#12693
Quick Description:
FlinkPulsarSink can't complete checkpoint if one of pulsar broker goes down. And when it reaches checkpoint restart limit, it will trigger a full restart and then see a traffic spike which cause the Bookie to go down as well.
To Reproduce
Steps to reproduce the behavior:
Check the original Bug Report: apache/pulsar#12693
Expected behavior
After one broker goes down, the checkpoint should complete before retry limit.
The text was updated successfully, but these errors were encountered: