[SPARK-26042] KafkaContinuousSourceTopicDeletionSuite may hang forever

• Type: Test
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.4.0
• Fix Version/s: 2.4.1, 3.0.0
• Component/s: Structured Streaming, Tests
• Assignee: Shixiong Zhu
• Reporter: Shixiong Zhu
• Created: 13/Nov/18 19:26
• Updated: 14/Nov/18 18:20
• Resolved: 14/Nov/18 18:20

Description

Saw the following thread dump in some build:

"stream execution thread for [id = 1c13482e-1edf-4b5c-b63a-d652738c8a48, runId = 10667ce9-7eac-4cef-a525-f1bd08eb50f1]" #4406 daemon prio=5 os_prio=0 tid=0x00007fab1d3c5000 nid=0x7f4b waiting on condition [0x00007fa96efcb000]
   java.lang.Thread.State: WAITING (parking)
	at sun.misc.Unsafe.park(Native Method)
	- parking to wait for  <0x000000070a904cf8> (a scala.concurrent.impl.Promise$CompletionLatch)
	at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:836)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedInterruptibly(AbstractQueuedSynchronizer.java:997)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireSharedInterruptibly(AbstractQueuedSynchronizer.java:1304)
...
	at org.apache.spark.sql.execution.SparkPlan.executeQuery(SparkPlan.scala:180)
	at org.apache.spark.sql.execution.SparkPlan.execute(SparkPlan.scala:131)
	at org.apache.spark.sql.execution.QueryExecution.toRdd$lzycompute(QueryExecution.scala:109)
	- locked <0x000000070a913ee8> (a org.apache.spark.sql.execution.streaming.IncrementalExecution)
	at org.apache.spark.sql.execution.QueryExecution.toRdd(QueryExecution.scala:109)
	at org.apache.spark.sql.execution.streaming.continuous.ContinuousExecution$$anonfun$runContinuous$3$$anonfun$apply$1.apply(ContinuousExecution.scala:270)
	at org.apache.spark.sql.execution.streaming.continuous.ContinuousExecution$$anonfun$runContinuous$3$$anonfun$apply$1.apply(ContinuousExecution.scala:270)
,,,


"pool-1-thread-1-ScalaTest-running-KafkaContinuousSourceTopicDeletionSuite" #20 prio=5 os_prio=0 tid=0x00007fabc4e78800 nid=0x23be waiting for monitor entry [0x00007fab3dbff000]
   java.lang.Thread.State: BLOCKED (on object monitor)
	at org.apache.spark.sql.execution.QueryExecution.executedPlan$lzycompute(QueryExecution.scala:100)
	- waiting to lock <0x000000070a913ee8> (a org.apache.spark.sql.execution.streaming.IncrementalExecution)
	at org.apache.spark.sql.execution.QueryExecution.executedPlan(QueryExecution.scala:100)
	at org.apache.spark.sql.kafka010.KafkaContinuousSourceTopicDeletionSuite$$anonfun$3$$anonfun$apply$mcV$sp$12$$anonfun$apply$15.apply(KafkaContinuousSourceSuite.scala:210)
	at org.apache.spark.sql.kafka010.KafkaContinuousSourceTopicDeletionSuite$$anonfun$3$$anonfun$apply$mcV$sp$12$$anonfun$apply$15.apply(KafkaContinuousSourceSuite.scala:209)
...

It hung forever because the test main thread was trying to access `executedPlan` but the lock was held by the streaming thread.

This is a pretty common issue when using lazy vals as all lazy vals share the same lock.

Issue Links

• links to: [Github] Pull Request #23023 (zsxwing)