[
https://issues.apache.org/jira/browse/KAFKA-6870?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16468538#comment-16468538
]
Rajini Sivaram commented on KAFKA-6870:
---------------------------------------
[~ijuma] This is a different issue from KAFKA-6765, there is missing
synchronization.
[~chia7712] Yes, I think it makes sense to add `synchronized(this.lock)` in
`KafkaMetric#measurableValue`.
> Concurrency conflicts in SampledStat
> ------------------------------------
>
> Key: KAFKA-6870
> URL: https://issues.apache.org/jira/browse/KAFKA-6870
> Project: Kafka
> Issue Type: Bug
> Reporter: Chia-Ping Tsai
> Priority: Major
> Fix For: 2.0.0, 1.1.1
>
>
> The samples stored in SampledStat is not thread-safe. However,
> ReplicaFetcherThreads used to handle replica to specified brokers may update
> (when the samples is empty, we will add a new sample to it) and iterate the
> samples concurrently, and then cause the ConcurrentModificationException.
> {code:java}
> [2018-05-03 13:50:56,087] ERROR [ReplicaFetcher replicaId=106, leaderId=100,
> fetcherId=0] Error due to (kafka.server.ReplicaFetcherThread:76)
> java.util.ConcurrentModificationException
> at java.util.ArrayList$Itr.checkForComodification(ArrayList.java:909)
> at java.util.ArrayList$Itr.next(ArrayList.java:859)
> at
> org.apache.kafka.common.metrics.stats.Rate$SampledTotal.combine(Rate.java:132)
> at
> org.apache.kafka.common.metrics.stats.SampledStat.measure(SampledStat.java:78)
> at org.apache.kafka.common.metrics.stats.Rate.measure(Rate.java:66)
> at
> org.apache.kafka.common.metrics.KafkaMetric.measurableValue(KafkaMetric.java:85)
> at org.apache.kafka.common.metrics.Sensor.checkQuotas(Sensor.java:201)
> at org.apache.kafka.common.metrics.Sensor.checkQuotas(Sensor.java:192)
> at
> kafka.server.ReplicationQuotaManager.isQuotaExceeded(ReplicationQuotaManager.scala:104)
> at
> kafka.server.ReplicaFetcherThread.kafka$server$ReplicaFetcherThread$$shouldFollowerThrottle(ReplicaFetcherThread.scala:384)
> at
> kafka.server.ReplicaFetcherThread$$anonfun$buildFetchRequest$1.apply(ReplicaFetcherThread.scala:263)
> at
> kafka.server.ReplicaFetcherThread$$anonfun$buildFetchRequest$1.apply(ReplicaFetcherThread.scala:261)
> at
> scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59)
> at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:48)
> at
> kafka.server.ReplicaFetcherThread.buildFetchRequest(ReplicaFetcherThread.scala:261)
> at
> kafka.server.AbstractFetcherThread$$anonfun$2.apply(AbstractFetcherThread.scala:102)
> at
> kafka.server.AbstractFetcherThread$$anonfun$2.apply(AbstractFetcherThread.scala:101)
> at kafka.utils.CoreUtils$.inLock(CoreUtils.scala:250)
> at
> kafka.server.AbstractFetcherThread.doWork(AbstractFetcherThread.scala:101)
> at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:82)
> {code}
> Before
> [https://github.com/apache/kafka/commit/d734f4e56d276f84b8c52b602edd67d41cbb6c35|https://github.com/apache/kafka/commit/d734f4e56d276f84b8c52b602edd67d41cbb6c35]
> the ConcurrentModificationException doesn't exist since all changes to
> samples is "add" currently. Using the get(index) is able to avoid the
> ConcurrentModificationException.
> In short, we can just make samples thread-safe. Or just replace the foreach
> loop by get(index) if we have concerns about the performance of thread-safe
> list...
>
>
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
