I am getting OutOfMemory error very frequently.
Sample Code:
Flowable.fromIterable(req.getAllClaims()).filter(Objects::nonNull)
.flatMap(data -> {
ClaimStatusCollector collector = statsCollector.get(data.getName());
collector.setVehicleHelper(vehicleHelper);
collector.setMedicalHelper(medicalHelper);
return Flowable.fromCallable(() -> collector.verifyStatus(data)
);
}, 5)
.blockingIterable().forEach(data -> {
claims.add(data.blockingFirst());
});
VehicleClaimStatusCollector.java
#Override
public Flowable<Claim> verifyStatus(Claim claim)
{
return Flowable.create(emitter -> {
try
{
//external http call
emitter.onNext(claim);
}
catch (Exception e)
{
emitter.onNext(claim);
}
emitter.onComplete();
}, BackpressureStrategy.BUFFER);
}
MedicalClaimStatusCollector.java
#Override
public Flowable<Claim> verifyStatus(Claim claim)
{
return Flowable.create(emitter -> {
try
{
//external http call
emitter.onNext(claim);
}
catch (Exception e)
{
emitter.onNext(claim);
}
emitter.onComplete();
}, BackpressureStrategy.BUFFER);
}
RxCachedThreadScheduler-6034" #6994 daemon prio=5 os_prio=31 tid=0x00007ffcab9ce800 nid=0x42d07 runnable [0x0000700109ffd000]
java.lang.Thread.State: RUNNABLE
at java.net.SocketInputStream.socketRead0(Native Method)
Its very clear that we are getting socket exception from external http request. I have also set readtimeout in jersey client. All the threads are hanged there and it throws outofmemory error.
java.lang.OutOfMemoryError: unable to create new native thread
at java.lang.Thread.start0(Native Method)
at java.lang.Thread.start(Thread.java:714)
at java.util.concurrent.ThreadPoolExecutor.addWorker(ThreadPoolExecutor.java:950)
at java.util.concurrent.ThreadPoolExecutor.ensurePrestart(ThreadPoolExecutor.java:1587)
at java.util.concurrent.ScheduledThreadPoolExecutor.delayedExecute(ScheduledThreadPoolExecutor.java:334)
at java.util.concurrent.ScheduledThreadPoolExecutor.schedule(ScheduledThreadPoolExecutor.java:549)
at java.util.concurrent.ScheduledThreadPoolExecutor.submit(ScheduledThreadPoolExecutor.java:648)
at io.reactivex.internal.schedulers.NewThreadWorker.scheduleActual(NewThreadWorker.java:146)
at io.reactivex.internal.schedulers.IoScheduler$EventLoopWorker.schedule(IoScheduler.java:230)
at io.reactivex.Scheduler$Worker.schedule(Scheduler.java:273)
at io.reactivex.internal.operators.flowable.FlowableSubscribeOn.subscribeActual(FlowableSubscribeOn.java:48)
at io.reactivex.Flowable.subscribe(Flowable.java:13234)
at io.reactivex.Flowable.subscribe(Flowable.java:13170)
at io.reactivex.Flowable.subscribe(Flowable.java:13091)
Can someone help me on this?
Related
I'm trying to develop Netty using TCP. I am using the IBM MQ client to connect to the MQ broker, and the idea is I need to develop a TCP server that receives a message passes it to MQ and if the server responds send it to the client that sent the request. Therefore, I need to implement a JMS listener for async message. The problem is that the JMS listener is outside of the Netty channel and I'm trying to figure out how to read the message add it to a Netty channel and send it immediately to the client connected to TCP socket. I can send messages perfectly. The problem is when the server responds. I receive the message, get the context/channel from the clientConnectionProvider and I writeAndFlush, but I don't see the message arrive at the client.
I create the listener in the main class.
public class Main {
private final Integer port;
private final Destination sendDestination;
private final JMSContext jmsSendContext;
private final JMSConsumer consumer;
private final JMSContext jmsRecieveContext;
private final Destination consumerDestination;
private final ClientConnectionProvider clientConnectionProvider;
public Main(Properties properties)
throws JMSException {
if (properties.containsKey(ConfigurationEnum.SERVER_PORT) {
this.port = properties.getProperty(ConfigurationEnum.SERVER_PORT)
} else {
log.error("server.port not defined in properties"
throw new ConfigException(
String.format("server.port not defined in properties");
}
JmsFactoryFactory ff = JmsFactoryFactory.getInstance(JmsConstants.WMQ_PROVIDER);
JmsConnectionFactory cf = ff.createConnectionFactory();
// Set the properties
cf.setStringProperty(CommonConstants.WMQ_HOST_NAME,
properties.getProperty(ConfigurationEnum.IBM_MQ_HOST.getValue()));
cf.setIntProperty(CommonConstants.WMQ_PORT,
Integer.parseInt(properties.getProperty(ConfigurationEnum.IBM_MQ_PORT.getValue())));
cf.setStringProperty(CommonConstants.WMQ_CHANNEL,
properties.getProperty(ConfigurationEnum.IBM_MQ_CHANNEL.getValue()));
cf.setIntProperty(CommonConstants.WMQ_CONNECTION_MODE, WMQConstants.WMQ_CM_CLIENT);
cf.setStringProperty(CommonConstants.WMQ_QUEUE_MANAGER,
properties.getProperty(ConfigurationEnum.IBM_QUEUE_MANAGER.getValue()));
cf.setStringProperty(CommonConstants.WMQ_APPLICATIONNAME, "FIX Orchestra Gateway");
cf.setBooleanProperty(JmsConstants.USER_AUTHENTICATION_MQCSP, true);
cf.setStringProperty(JmsConstants.USERID, properties.getProperty(ConfigurationEnum.IBM_APP_USER.getValue()));
cf.setStringProperty(JmsConstants.PASSWORD, properties.getProperty(ConfigurationEnum.IBM_APP_PASS.getValue()));
clientConnectionProvider = new ClientConnectionProvider();
jmsRecieveContext = cf.createContext();
consumerDestination = jmsRecieveContext
.createQueue(properties.getProperty(ConfigurationEnum.IBM_QUEUE_CONSUMER.getValue()));
consumer = jmsRecieveContext.createConsumer(consumerDestination);
consumer.setMessageListener(new JMSMessageListener(clientConnectionProvider));
jmsRecieveContext.start();
jmsSendContext = cf.createContext();
sendDestination = jmsSendContext
.createQueue(properties.getProperty(ConfigurationEnum.IBM_QUEUE_TRANSACTION.getValue()));
}
public void start() throws InterruptedException {
EventLoopGroup bossGroup = new NioEventLoopGroup(1);
EventLoopGroup workerGroup = new NioEventLoopGroup(10);
try {
ServerBootstrap serverBootstrap = new ServerBootstrap();
serverBootstrap.group(bossGroup, workerGroup).channel(NioServerSocketChannel.class)
.option(ChannelOption.SO_BACKLOG, 100).option(ChannelOption.SO_KEEPALIVE, Boolean.TRUE)
.handler(new LoggingHandler(LogLevel.INFO))
.childHandler(new DefaultChannelInitializer());
// Start the server.
ChannelFuture f = serverBootstrap.bind(port).sync();
// Wait until the server socket is closed.
f.channel().closeFuture().sync();
} finally {
bossGroup.shutdownGracefully();
workerGroup.shutdownGracefully();
jmsRecieveContext.stop();
jmsRecieveContext.close();
jmsSendContext.close();
}
}
public static void main(String[] args) throws InterruptedException {
Properties properties = new Properties();
try (InputStream inputStream = new FileInputStream(args[0])) {
properties.load(inputStream);
new Main(properties).start();
} catch (FileNotFoundException e) {
log.error("Properties file specified in path {} was not found.", args[0], e);
} catch (IOException e) {
log.error("There was an IO error.", e);
} catch (JMSException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (ConfigException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
}
The listener is a simple class.
#AllArgsConstructor
public class JMSMessageListener implements MessageListener {
private final ClientConnectionProvider clientConnectionProvider;
#Override
public void onMessage(Message message) {
try {
String messageString = message.getBody(String.class);
if (clientConnectionProvider.contains(ClientID.get(messageString))) {
ClientConnection cc = clientConnectionProvider.getConnection(ClientID.get(messageString));
if (cc.getCtx() == null) {
// TODO: Need to save message when client reconects
} else {
cc.getCtx().channel().write(messageString);
}
}
} catch (JMSException e) {
e.printStackTrace();
}
}
}
You should call writeAndFlush(...) and attach a ChannelFutureListener to the ChannelFuture returned to it. In the listener you can check if the write did succeed or fail (and if so print the exception). In your current code you only call write(...) which only put the message in the outboundbuffer of the Channel but not actually flush it to the socket.
I have a BizTalk Custom Pipeline Component that writes an SFTP file (using SSH.net), triggered by an SFTP (WinSCP) receive location.
The code within the Retry occasionally (around half the time) does not hit either the "Success" nor the logging catch block and no further processing occurs within the Pipeline. I assume that means the thread has been destroyed.
I added the Retry code later to make it try a few times but with the thread being destroyed I don't always get a success or 3 failures.
What could cause this behaviour in BizTalk 2016?
public void Archive(byte[] content,
string archivePath,
string userName,
string password,
string serverAddress,
string sshHostKeyFingerprint)
{
Retry(3, () =>
{
try
{
using (var sftpClient = new SftpClient(serverAddress, userName, password))
{
if (!string.IsNullOrWhiteSpace(sshHostKeyFingerprint))
{
sshHostKeyFingerprint = sshHostKeyFingerprint.Split(' ').Last();
sftpClient.HostKeyReceived += delegate (object sender, HostKeyEventArgs e)
{
if (e.FingerPrint.SequenceEqual(ConvertFingerprintToByteArray(sshHostKeyFingerprint)))
e.CanTrust = true;
else
e.CanTrust = false;
};
}
sftpClient.Connect();
sftpClient.WriteAllBytes(archivePath, content);
sftpClient.Disconnect();
LogInfo($"Success");
}
}
catch (Exception exception)
{
// show the bad path for "No such file" errors
throw new InvalidOperationException($"Failed to create file '{archivePath}'", exception);
}
});
}
private void Retry(int maxAttempts, Action action)
{
int attempt = 1;
while (attempt <= maxAttempts)
{
try
{
action();
break; // success
}
catch (Exception exception)
{
LogWarning($"Attempt {attempt} Error: {exception.ToString()}");
if (attempt == maxAttempts)
throw; // final attempt exception propagated
}
finally
{
attempt++;
}
}
}
I'm using Jersey Multipart for uploading file to the server via Rest API. In the resource method, I accessed the file content via InputStream. I want to return the uploaded file size to the client with EventOutput using SSE so the client easily get the uploaded file size directly from upload resource method.
I'm using Jersey as JAX-RS implementation in java with Grizzly Http server. Here is my code:
#POST
#Path("upload")
#Produces(SseFeature.SERVER_SENT_EVENTS)
#Consumes("multipart/form-data;charset=utf-8")
public EventOutput upload(#FormDataParam("file") InputStream file,
#FormDataParam("file") FormDataContentDisposition fileDisposition) {
final EventOutput eventOutput = new EventOutput();
try {
new Thread(new Runnable() {
#Override
public void run() {
try {
int read = -1;
byte[] buffer = new byte[1024];
OutboundEvent.Builder eventBuilder
= new OutboundEvent.Builder();
OutboundEvent event = null;
long totalRead = 0, lastReadMB = 0;
while ((read = file.read(buffer)) != -1) {
totalRead += read;
if (lastReadMB != (totalRead / (1024 * 1024))) {
lastReadMB = totalRead / (1024 * 1024);
event = eventBuilder.name("uploaded").data(Long.class, totalRead).build();
eventOutput.write(event);
}
}
event = eventBuilder.name("uploaded").data(Long.class, totalRead).build();
eventOutput.write(event);
} catch (Exception e) {
throw new RuntimeException(
"Error when writing the event.", e);
} finally {
try {
eventOutput.close();
} catch (Exception ioClose) {
throw new RuntimeException(
"Error when closing the event output.", ioClose);
}
}
}
}).start();
return eventOutput;
} catch (Exception e) {
logger.error(e.toString(), e);
}
throw new WebApplicationException(Response.status(Response.Status.INTERNAL_SERVER_ERROR).
entity("something happened").build());
}
The problem is when my resource method return EventOutput as a response and request processing thread back to the I/O container, the InputStream closed and the processing thread can't access to the uploaded file. Here is the exception:
Exception in thread "Thread-1" java.lang.RuntimeException: Error when writing the event.
at com.WebService.ContentService$1.run(ContentService.java:192)
at java.lang.Thread.run(Thread.java:745)
Caused by: org.jvnet.mimepull.MIMEParsingException: java.io.IOException: Stream Closed
at org.jvnet.mimepull.WeakDataFile.read(WeakDataFile.java:115)
at org.jvnet.mimepull.DataFile.read(DataFile.java:77)
at org.jvnet.mimepull.FileData.read(FileData.java:69)
at org.jvnet.mimepull.DataHead$ReadMultiStream.fetch(DataHead.java:265)
at org.jvnet.mimepull.DataHead$ReadMultiStream.read(DataHead.java:219)
at java.io.InputStream.read(InputStream.java:101)
at com.WebService.ContentService$1.run(ContentService.java:181)
... 1 more
Caused by: java.io.IOException: Stream Closed
at java.io.RandomAccessFile.seek0(Native Method)
at java.io.RandomAccessFile.seek(RandomAccessFile.java:557)
at org.jvnet.mimepull.WeakDataFile.read(WeakDataFile.java:112)
... 7 more
1- What's the problem in the code? Why InputStream is closed in the middle of the file transfer?
2- Is there any alternative way to return the uploaded file size to the client in server side? (REQUIREMENT: the upload resource method must handle upload file asynchronously in different thread)
My app crashes whenever a 401 error occurs in the app. I created a custom error handler but i'm not sure how to get the app to use this error handler so the app stops gracefully, without flashing the user "Unforunately, Application has stopped."
Here's my code:
... new RestAdapter.Builder()...setErrorHandler(getErrorHandler()).build();
public ErrorHandler getErrorHandler() {
return new ErrorHandler() {
#Override
public Throwable handleError(RetrofitError cause) {
Response r = cause.getResponse();
if (r != null && r.getStatus() == 401) {
Log.e(TAG, "user not authorized:" + cause.toString());
} else {
Log.e(TAG, "regular exception thrown for CloudManager");
}
return cause;
}
};
As you can check with javadoc ErrorHandler is used when you want to wrap RetrofitError into CustomError exception. It doesn't catch exception for you.
class MyErrorHandler implements ErrorHandler {
#Override public Throwable handleError(RetrofitError cause) {
Response r = cause.getResponse();
if (r != null && r.getStatus() == 401) {
return new UnauthorizedException(cause);
}
return cause;
}
}
To prevent app from crashing you have to catch error earlier.
#GET("/users/{user}/repos")
List<Repo> listRepos(#Path("user") String user);
try {
List<Repo> repos = retrofit.listRepos("octocat");
} catch (RetrofitError error) {
// TODO: handle error
}
I am trying to reuse a Jersey2(Jersey 2.16) Client for async invocation. However after 2 requests, I see that the threads going into a waiting state, waiting on a lock. Since client creation is an expensive operation, I am trying to reuse the client in the async calls. The issue occurs only with ApacheConnectorProvider as the connector class. I want to use ApacheConnectorProvider, as I need to use a proxy and set SSL properties and I want to use PoolingHttpClientConnectionManager.
The sample code is given below:
public class Example {
Integer eventId = 0;
private ClientConfig getClientConfig()
{
ClientConfig clientConfig = new ClientConfig();
ApacheConnectorProvider provider = new ApacheConnectorProvider();
clientConfig.property(ClientProperties.REQUEST_ENTITY_PROCESSING,RequestEntityProcessing.BUFFERED);
clientConfig.connectorProvider(provider);
return clientConfig;
}
private Client createClient()
{
Client client = ClientBuilder.newClient(getClientConfig());
return client;
}
public void testAsyncCall()
{
Client client = createClient();
System.out.println("Testing a new Async call on thread " + Thread.currentThread().getId());
JSONObject jsonObject = new JSONObject();
jsonObject.put("value", eventId++);
invoker(client, "http://requestb.in/nn0sffnn" , jsonObject);
invoker(client, "http://requestb.in/nn0sffnn" , jsonObject);
invoker(client, "http://requestb.in/nn0sffnn" , jsonObject);
client.close();
}
private void invoker(Client client, String URI, JSONObject jsonObject)
{
final Future<Response> responseFuture = client.target(URI)
.request()
.async()
.post(Entity.entity(jsonObject.toJSONString(), MediaType.TEXT_PLAIN));
try {
Response r = responseFuture.get();
System.out.println("Response is on URI " + URI + " : " + r.getStatus());
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (ExecutionException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
public static void main(String args[])
{
Example client1 = new Example();
client1.testAsyncCall();
return;
}
}
The response I see is:
Testing a new Async call on thread 1
Response is on URI http://requestb.in/nn0sffnn : 200
Response is on URI http://requestb.in/nn0sffnn : 200
On looking at the thread stack, I see the following trace:
"jersey-client-async-executor-0" prio=6 tid=0x043a4c00 nid=0x56f0 waiting on condition [0x03e5f000]
java.lang.Thread.State: WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x238ee148> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java:186)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2043)
at org.apache.http.pool.PoolEntryFuture.await(PoolEntryFuture.java:133)
at org.apache.http.pool.AbstractConnPool.getPoolEntryBlocking(AbstractConnPool.java:282)
at org.apache.http.pool.AbstractConnPool.access$000(AbstractConnPool.java:64)
at org.apache.http.pool.AbstractConnPool$2.getPoolEntry(AbstractConnPool.java:177)
at org.apache.http.pool.AbstractConnPool$2.getPoolEntry(AbstractConnPool.java:170)
Can someone give me a suggestion as to how to reuse Client objects for async requests and may be how to get over this issue as well.