[
https://issues.apache.org/jira/browse/SPARK-15725?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15311567#comment-15311567
]
Ryan Blue commented on SPARK-15725:
-----------------------------------
I'm linking to a work-around that ensures the AM thread that drives the
YarnAllocator sleeps its minimum interval. While there are still problems
because of the YarnAllocator's locking strategy and the driver threads waiting
on a kill RPC to the AM, this avoids calling {{allocateResources}} too often
and the calls return quickly enough that the application can progress.
The logs from running with this patch confirm that the allocator is called too
often:
{code:title=AM log with work-around}
16/06/01 18:30:42 INFO yarn.ApplicationMaster$AMEndpoint: Driver requested to
kill executor(s) 24.
16/06/01 18:30:50 INFO yarn.YarnAllocator: Driver requested a total number of
266 executor(s).
16/06/01 18:30:50 INFO yarn.ApplicationMaster: Number of pending allocations is
0. Slept for 1/3000.
16/06/01 18:30:50 INFO yarn.ApplicationMaster: Going back to sleep for 199 ms
16/06/01 18:30:50 INFO yarn.ApplicationMaster$AMEndpoint: Driver requested to
kill executor(s) 134.
16/06/01 18:30:50 INFO yarn.YarnAllocator: Driver requested a total number of
265 executor(s).
16/06/01 18:30:50 INFO yarn.ApplicationMaster$AMEndpoint: Driver requested to
kill executor(s) 74.
16/06/01 18:30:50 INFO yarn.YarnAllocator: Driver requested a total number of
264 executor(s).
16/06/01 18:30:50 INFO yarn.ApplicationMaster$AMEndpoint: Driver requested to
kill executor(s) 89.
16/06/01 18:30:50 INFO yarn.YarnAllocator: Driver requested a total number of
263 executor(s).
16/06/01 18:30:50 INFO yarn.ApplicationMaster$AMEndpoint: Driver requested to
kill executor(s) 122.
16/06/01 18:30:50 INFO yarn.YarnAllocator: Driver requested a total number of
262 executor(s).
{code}
> Dynamic allocation hangs YARN app when executors time out
> ---------------------------------------------------------
>
> Key: SPARK-15725
> URL: https://issues.apache.org/jira/browse/SPARK-15725
> Project: Spark
> Issue Type: Bug
> Components: Spark Core
> Affects Versions: 1.6.1, 2.0.0
> Reporter: Ryan Blue
>
> We've had a problem with a dynamic allocation and YARN (since 1.6) where a
> large stage will cause a lot of executors to get killed around the same time
> and RPC calls to the application master time out. This can happen even with a
> small number of executors (~100).
> When executors are killed by the driver, the [network connection to the
> driver
> disconnects|https://github.com/apache/spark/blob/master/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnSchedulerBackend.scala#L201].
> That results in a call to the AM to find out why the executor died, followed
> by a [blocking and retrying `RemoveExecutor` RPC
> call|https://github.com/apache/spark/blob/master/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnSchedulerBackend.scala#L227]
> that results in a second `KillExecutor` call to the AM. When a lot of
> executors are killed around the same time, the driver's AM threads are all
> taken up blocking and waiting on the AM (see the stack trace below, which was
> the same for 42 threads). I think this behavior, the network disconnect and
> subsequent cleanup, is unique to YARN.
> {code:title=Driver AM thread stack}
> sun.misc.Unsafe.park(Native Method)
> java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:215)
> java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedNanos(AbstractQueuedSynchronizer.java:1037)
> java.util.concurrent.locks.AbstractQueuedSynchronizer.tryAcquireSharedNanos(AbstractQueuedSynchronizer.java:1328)
> scala.concurrent.impl.Promise$DefaultPromise.tryAwait(Promise.scala:208)
> scala.concurrent.impl.Promise$DefaultPromise.ready(Promise.scala:218)
> scala.concurrent.impl.Promise$DefaultPromise.result(Promise.scala:223)
> scala.concurrent.Await$$anonfun$result$1.apply(package.scala:190)
> scala.concurrent.BlockContext$DefaultBlockContext$.blockOn(BlockContext.scala:53)
> scala.concurrent.Await$.result(package.scala:190)
> org.apache.spark.rpc.RpcTimeout.awaitResult(RpcTimeout.scala:81)
> org.apache.spark.rpc.RpcEndpointRef.askWithRetry(RpcEndpointRef.scala:102)
> org.apache.spark.rpc.RpcEndpointRef.askWithRetry(RpcEndpointRef.scala:78)
> org.apache.spark.scheduler.cluster.YarnSchedulerBackend$YarnSchedulerEndpoint$$anonfun$receiveAndReply$1$$anonfun$applyOrElse$2.apply$mcV$sp(YarnSchedulerBackend.scala:286)
> org.apache.spark.scheduler.cluster.YarnSchedulerBackend$YarnSchedulerEndpoint$$anonfun$receiveAndReply$1$$anonfun$applyOrElse$2.apply(YarnSchedulerBackend.scala:286)
> org.apache.spark.scheduler.cluster.YarnSchedulerBackend$YarnSchedulerEndpoint$$anonfun$receiveAndReply$1$$anonfun$applyOrElse$2.apply(YarnSchedulerBackend.scala:286)
> scala.concurrent.impl.Future$PromiseCompletingRunnable.liftedTree1$1(Future.scala:24)
> scala.concurrent.impl.Future$PromiseCompletingRunnable.run(Future.scala:24)
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> java.lang.Thread.run(Thread.java:745)
> {code}
> The RPC calls to the AM aren't returning because the `YarnAllocator` is
> spending all of its time in the `allocateResources` method. That class's
> public methods are synchronized so only one RPC can be satisfied at a time.
> The reason why it is constantly calling `allocateResources` is because [its
> thread|https://github.com/apache/spark/blob/master/yarn/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala#L467]
> is [woken
> up|https://github.com/apache/spark/blob/master/yarn/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala#L686]
> by calls to get the failure reason for an executor -- which is part of the
> chain of events in the driver for each executor that goes down.
> The final result is that the `YarnAllocator` doesn't respond to RPC calls for
> long enough that calls time out and replies for non-blocking calls are
> dropped. Then the application is unable to do any work because everything
> retries or exits and the application hangs for 24+ hours, until enough errors
> accumulate that it dies.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
