[
https://issues.apache.org/jira/browse/DRILL-6295?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16437785#comment-16437785
]
ASF GitHub Bot commented on DRILL-6295:
---------------------------------------
Github user ilooner commented on a diff in the pull request:
https://github.com/apache/drill/pull/1208#discussion_r181488592
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/partitionsender/PartitionerDecorator.java
---
@@ -262,68 +280,122 @@ public FlushBatchesHandlingClass(boolean
isLastBatch, boolean schemaChanged) {
}
@Override
- public void execute(Partitioner part) throws IOException {
+ public void execute(Partitioner part) throws IOException,
InterruptedException {
part.flushOutgoingBatches(isLastBatch, schemaChanged);
}
}
/**
- * Helper class to wrap Runnable with customized naming
- * Exception handling
+ * Helper class to wrap Runnable with cancellation and waiting for
completion support
*
*/
- private static class CustomRunnable implements Runnable {
+ private static class PartitionerTask implements Runnable {
+
+ private enum STATE {
+ NEW,
+ COMPLETING,
+ NORMAL,
+ EXCEPTIONAL,
+ CANCELLED,
+ INTERRUPTING,
+ INTERRUPTED
+ }
+
+ private final AtomicReference<STATE> state;
+ private final AtomicReference<Thread> runner;
+ private final PartitionerDecorator partitionerDecorator;
+ private final AtomicInteger count;
- private final String parentThreadName;
- private final CountDownLatch latch;
private final GeneralExecuteIface iface;
- private final Partitioner part;
+ private final Partitioner partitioner;
private CountDownLatchInjection testCountDownLatch;
- private volatile IOException exp;
+ private volatile ExecutionException exception;
- public CustomRunnable(final String parentThreadName, final
CountDownLatch latch, final GeneralExecuteIface iface,
- final Partitioner part, CountDownLatchInjection
testCountDownLatch) {
- this.parentThreadName = parentThreadName;
- this.latch = latch;
+ public PartitionerTask(PartitionerDecorator partitionerDecorator,
GeneralExecuteIface iface, Partitioner partitioner, AtomicInteger count,
CountDownLatchInjection testCountDownLatch) {
+ state = new AtomicReference<>(STATE.NEW);
+ runner = new AtomicReference<>();
+ this.partitionerDecorator = partitionerDecorator;
this.iface = iface;
- this.part = part;
+ this.partitioner = partitioner;
+ this.count = count;
this.testCountDownLatch = testCountDownLatch;
}
@Override
public void run() {
- // Test only - Pause until interrupted by fragment thread
- try {
- testCountDownLatch.await();
- } catch (final InterruptedException e) {
- logger.debug("Test only: partitioner thread is interrupted in test
countdown latch await()", e);
- }
-
- final Thread currThread = Thread.currentThread();
- final String currThreadName = currThread.getName();
- final OperatorStats localStats = part.getStats();
- try {
- final String newThreadName = parentThreadName + currThread.getId();
- currThread.setName(newThreadName);
+ final Thread thread = Thread.currentThread();
+ Preconditions.checkState(runner.compareAndSet(null, thread),
+ "PartitionerTask can be executed only once.");
+ if (state.get() == STATE.NEW) {
+ final String name = thread.getName();
+ thread.setName(String.format("Partitioner-%s-%d",
partitionerDecorator.thread.getName(), thread.getId()));
+ final OperatorStats localStats = partitioner.getStats();
localStats.clear();
localStats.startProcessing();
- iface.execute(part);
- } catch (IOException e) {
- exp = e;
- } finally {
- localStats.stopProcessing();
- currThread.setName(currThreadName);
- latch.countDown();
+ ExecutionException executionException = null;
+ try {
+ // Test only - Pause until interrupted by fragment thread
+ testCountDownLatch.await();
+ iface.execute(partitioner);
+ } catch (InterruptedException e) {
+ if (state.compareAndSet(STATE.NEW, STATE.INTERRUPTED)) {
+ logger.warn("Partitioner Task interrupted during the run", e);
+ }
+ } catch (Throwable t) {
+ executionException = new ExecutionException(t);
+ }
+ if (state.compareAndSet(STATE.NEW, STATE.COMPLETING)) {
+ if (executionException == null) {
+ localStats.stopProcessing();
+ state.lazySet(STATE.NORMAL);
+ } else {
+ exception = executionException;
+ state.lazySet(STATE.EXCEPTIONAL);
+ }
+ }
+ if (count.decrementAndGet() == 0) {
+ LockSupport.unpark(partitionerDecorator.thread);
+ }
+ thread.setName(name);
+ }
+ runner.set(null);
+ while (state.get() == STATE.INTERRUPTING) {
--- End diff --
I think we can still leave a dirty interrupt flag. Consider the following
case.
1. We create a PartitionerTask **Task A**.
2. **Task A** executes and reaches `thread.setName(name)` in the run
method.
3. Now a cancel is called.
4. `thread.interrupt()` and `state.lazySet(STATE.INTERRUPTED)` is executed.
5. Now **Task A** can continue executing
6. **Task A** checks `state.get() == STATE.INTERRUPTING` which is false.
So Thread.sleep is never called and the interrupt flag is never cleared.
> PartitionerDecorator may close partitioners while CustomRunnable are active
> during query cancellation
> -----------------------------------------------------------------------------------------------------
>
> Key: DRILL-6295
> URL: https://issues.apache.org/jira/browse/DRILL-6295
> Project: Apache Drill
> Issue Type: Bug
> Reporter: Vlad Rozov
> Assignee: Vlad Rozov
> Priority: Critical
> Fix For: 1.14.0
>
>
> During query cancellation, in case
> {{PartitionerDecorator.executeMethodLogic()}} is active (waiting on the
> {{latch}}), the wait will be interrupted and {{Futures}} cancelled, but there
> is no guarantee that all {{CustomRunnable}} terminate before returning from
> {{PartitionerDecorator.executeMethodLogic()}}. On exit, both income and
> outgoing batches are cleared, leading to clearing of underlying {{Vectors}}
> and {{DrillBufs}}. This eventually causes unallocated memory access and JVM
> crash as {{CustomRunnable}} may execute after income/outgoing batches are
> cleared.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
