[KAFKA-5745] Partition.makeLeader() should convert HW to OffsetMetadata before becoming the leader

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.11.0.0
• Fix Version/s: 0.11.0.1, 1.0.0
• Component/s: core
• Assignee: huxihx
• Reporter: Jun Rao
• Created: 17/Aug/17 01:06
• Updated: 18/Aug/17 15:07
• Resolved: 18/Aug/17 15:07

Description

Saw the following uncaught exception in the log.

ERROR [KafkaApi-411] Error when handling request Name: FetchRequest; Version: 3; CorrelationId: 708572; ClientId: client-0; ReplicaId: -1; MaxWait: 500 ms; MinBytes: 1 bytes; MaxBytes:52428800 bytes; RequestInfo: ([topic1,3],PartitionFetchInfo(953794,1048576)) (kafka.server.KafkaApis)
org.apache.kafka.common.KafkaException: 953793 [-1 : -1] cannot compare its segment info with 953794 [829749 : 564535961] since it only has message offset info
at kafka.server.LogOffsetMetadata.onOlderSegment(LogOffsetMetadata.scala:48)
at kafka.server.DelayedFetch$$anonfun$tryComplete$1.apply(DelayedFetch.scala:93)
at kafka.server.DelayedFetch$$anonfun$tryComplete$1.apply(DelayedFetch.scala:77)
at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59)
at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:48)
at kafka.server.DelayedFetch.tryComplete(DelayedFetch.scala:77)
at kafka.server.DelayedOperation.safeTryComplete(DelayedOperation.scala:104)
at kafka.server.DelayedOperationPurgatory.tryCompleteElseWatch(DelayedOperation.scala:196)
at kafka.server.ReplicaManager.fetchMessages(ReplicaManager.scala:516)
at kafka.server.KafkaApis.handleFetchRequest(KafkaApis.scala:534)
at kafka.server.KafkaApis.handle(KafkaApis.scala:79)
at kafka.server.KafkaRequestHandler.run(KafkaRequestHandler.scala:60)
at java.lang.Thread.run(Thread.java:745)

Issue Links

• links to: GitHub Pull Request #3682