George Yang created KAFKA-17547:
-----------------------------------
Summary: Write large number of mirror maker logs when Kafka crashes
Key: KAFKA-17547
URL: https://issues.apache.org/jira/browse/KAFKA-17547
Project: Kafka
Issue Type: Test
Components: mirrormaker
Affects Versions: 3.7.1
Environment: OS: AlmaLinux 9.3
CPU: 28cores
Mem: 128GiB
Kafka: v3.7.0
MirrorMaker2: v3.7.1
Reporter: George Yang
Attachments: connect.log
We have deployed 2 data centers, each with one node, and on each node, there is
a Kafka pod and a MirrorMaker2 pod. Currently, if one of the Kafka pods
crashes, the running MirrorMaker2 pod will continuously output a large number
of logs (please see the attachment). As long as the Kafka pod remains down, the
logs will keep accumulating, eventually filling up the disk. Besides stopping
all the MirrorMaker2 instances or getting the crashed Kafka pod back online,
are there any other solutions to prevent this excessive logging? For example,
can we configure any parameters in MirrorMaker2 to handle this?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
