big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

YARN-5999. AMRMClientAsync will stop if any exceptions thrown on allocate call. Contributed by Jian He

Browse Source
This commit is contained in:
Xuan 2016-12-14 14:33:23 -08:00
parent f5e0bd30fd
commit 64a2d5be91
2 changed files with 13 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/main/java/org/apache/hadoop/yarn/client/api/async/impl/AMRMClientAsyncImpl.java
View File

@ -61,7 +61,7 @@ public class AMRMClientAsyncImpl<T extends ContainerRequest>
private final HeartbeatThread heartbeatThread;
private final CallbackHandlerThread handlerThread;
private final BlockingQueue<AllocateResponse> responseQueue;
private final BlockingQueue<Object> responseQueue;
private final Object unregisterHeartbeatLock = new Object();
@ -70,8 +70,6 @@ public class AMRMClientAsyncImpl<T extends ContainerRequest>
private volatile String collectorAddr;
private volatile Throwable savedException;
/**
*
* @param intervalMs heartbeat interval in milliseconds between AM and RM
@ -90,7 +88,6 @@ public AMRMClientAsyncImpl(AMRMClient<T> client, int intervalMs,
handlerThread = new CallbackHandlerThread();
responseQueue = new LinkedBlockingQueue<>();
keepRunning = true;
savedException = null;
}
/**
@ -111,9 +108,8 @@ public AMRMClientAsyncImpl(AMRMClient<T> client, int intervalMs,
super(client, intervalMs, callbackHandler);
heartbeatThread = new HeartbeatThread();
handlerThread = new CallbackHandlerThread();
responseQueue = new LinkedBlockingQueue<AllocateResponse>();
responseQueue = new LinkedBlockingQueue<Object>();
keepRunning = true;
savedException = null;
}
@Override
@ -265,7 +261,7 @@ public HeartbeatThread() {
public void run() {
while (true) {
AllocateResponse response = null;
Object response = null;
// synchronization ensures we don't send heartbeats after unregistering
synchronized (unregisterHeartbeatLock) {
if (!keepRunning) {
@ -280,10 +276,7 @@ public void run() {
return;
} catch (Throwable ex) {
LOG.error("Exception on heartbeat", ex);
savedException = ex;
// interrupt handler thread in case it waiting on the queue
handlerThread.interrupt();
return;
response = ex;
}
if (response != null) {
while (true) {
@ -316,19 +309,20 @@ public void run() {
return;
}
try {
AllocateResponse response;
if(savedException != null) {
LOG.error("Stopping callback due to: ", savedException);
handler.onError(savedException);
return;
}
Object object;
try {
response = responseQueue.take();
object = responseQueue.take();
} catch (InterruptedException ex) {
LOG.info("Interrupted while waiting for queue", ex);
continue;
}
if (object instanceof Throwable) {
progress = handler.getProgress();
handler.onError((Throwable) object);
continue;
}
AllocateResponse response = (AllocateResponse) object;
String collectorAddress = response.getCollectorAddr();
TimelineClient timelineClient = client.getRegisteredTimelineClient();
if (timelineClient != null && collectorAddress != null

2
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/test/java/org/apache/hadoop/yarn/client/api/async/impl/TestAMRMClientAsync.java
View File

@ -213,7 +213,7 @@ private void runHeartBeatThrowOutException(Exception ex) throws Exception{
asyncClient.stop();
// stopping should have joined all threads and completed all callbacks
Assert.assertTrue(callbackHandler.callbackCount == 0);
Assert.assertTrue(callbackHandler.callbackCount > 0);
}
@Test (timeout = 10000)