Today, after starting Kafka, it was found that only one node was started successfully and the other two failed. Check the log as follows

After multi-party inspection, it turned out that the service was not normally shut down before

Solution:

According to the log prompt, we need to manually delete the two index files under each partition, restart the cluster, and Kafka will automatically rebuild the index files

find /opt/module/kafka/logs/ -name "*.timeindex" |xargs rm -f
find /opt/module/kafka/logs/ -name "*.index" |xargs rm -f

Just restart the service.