[KAFKA-9016] Warn when log dir stopped serving replicas

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.4.0
• Fix Version/s: 2.5.0
• Component/s: log, log cleaner
• Labels: easy newbie
• Assignee: kumar uttpal
• Reporter: Viktor Somogyi-Vass
• Created: 10/Oct/19 13:42
• Updated: 16/Sep/20 11:58
• Resolved: 08/Nov/19 19:06

Description

Kafka should warn if the log directory stops serving replicas as usually it is due to an error and it's much visible if it's on warn level.

2019-09-19 12:39:54,194 ERROR kafka.server.LogDirFailureChannel: Error while writing to checkpoint file /kafka/data/diskX/replication-offset-checkpoint
java.io.SyncFailedException: sync failed
..
2019-09-19 12:39:54,197 INFO kafka.server.ReplicaManager: [ReplicaManager broker=638] Stopping serving replicas in dir /kafka/data/diskX
..
2019-09-19 12:39:54,205 INFO kafka.server.ReplicaFetcherManager: [ReplicaFetcherManager on broker 638] Removed fetcher for partitions Set(test1-0, test2-2, test-0, test2-2, test4-14, test4-0, test2-6)
2019-09-19 12:39:54,206 INFO kafka.server.ReplicaAlterLogDirsManager: [ReplicaAlterLogDirsManager on broker 638] Removed fetcher for partitions Set(test1-0, test2-2, test-0, test3-2, test4-14, test4-0, test2-6)
2019-09-19 12:39:54,222 INFO kafka.server.ReplicaManager: [ReplicaManager broker=638] Broker 638 stopped fetcher for partitions test1-0,test2-2,test-0,test3-2,test4-14,test4-0,test2-6 and stopped moving logs for partitions  because they are in the failed log directory /kafka/data/diskX.

Issue Links

• links to: GitHub Pull Request #7563