Question

OS: Linux ( Debian )
Java: Java(TM) SE Runtime Environment (build 1.7.0_05-b06)
Java: Java HotSpot(TM) 64-Bit Server VM (build 23.1-b03, mixed mode)
Apache http client: v4.2.3 ( latest )

I want to create one PoolingConnectionManager with several clients, each with its unique local address (ConnRoutePNames.LOCAL_ADDRESS).
Those clients will be used by several workers ( each worker selects random client and executes request ).

The problem is: When i set local address for clients, after some time (e.g. 1 min) i always get java.net.BindException with message “Address already in use”.

Question: Is it bug?

java.net.BindException: Address already in use
at java.net.PlainSocketImpl.socketBind(Native Method) ~[na:1.7.0_05]
at java.net.AbstractPlainSocketImpl.bind(AbstractPlainSocketImpl.java:376) ~[na:1.7.0_05]
at java.net.Socket.bind(Socket.java:627) ~[na:1.7.0_05]
at org.apache.http.conn.scheme.PlainSocketFactory.connectSocket(PlainSocketFactory.java:120) ~[httpclient-4.2.3.jar:4.2.3]
at org.apache.http.impl.conn.DefaultClientConnectionOperator.openConnection(DefaultClientConnectionOperator.java:180) ~[httpclient-4.2.3.jar:4.2.3]
at org.apache.http.impl.conn.ManagedClientConnectionImpl.open(ManagedClientConnectionImpl.java:294) ~[httpclient-4.2.3.jar:4.2.3]
at org.apache.http.impl.client.DefaultRequestDirector.tryConnect(DefaultRequestDirector.java:645) ~[httpclient-4.2.3.jar:4.2.3]
at org.apache.http.impl.client.DefaultRequestDirector.execute(DefaultRequestDirector.java:480) ~[httpclient-4.2.3.jar:4.2.3]
at org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:906) ~[httpclient-4.2.3.jar:4.2.3]
at org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:805) ~[httpclient-4.2.3.jar:4.2.3]
at org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:784) ~[httpclient-4.2.3.jar:4.2.3]
at controllers.Test$Worker.run(Test.java:67) ~[test_2.9.1-1.0-SNAPSHOT.jar:1.0-SNAPSHOT]
at java.lang.Thread.run(Thread.java:722) [na:1.7.0_05]

Code ( simplified ):

public static void main(String[] args) throws UnknownHostException
{
    ClientConnectionManager connManager = buildConnectionManager(30);

    List<HttpClient> clients = new ArrayList<>();
    for(int i=0; i<8; ++i)
    {
        HttpClient client = buildClient(connManager, InetAddress.getByName("111.111.111." + (50 + i));
        clients.add(client);
    }

    for(int i=0; i<30; ++i)
    {
        new Thread(new Worker(clients)).start();
    }
}

public static ClientConnectionManager buildConnectionManager(Integer parallelism)
{
    PoolingClientConnectionManager connectionManager = new PoolingClientConnectionManager();
    connectionManager.setMaxTotal(parallelism);
    connectionManager.setDefaultMaxPerRoute(parallelism);

    return connectionManager;
}

public static HttpClient buildClient(ClientConnectionManager connectionManager, InetAddress localAddress)
{
    DefaultHttpClient httpClient = new DefaultHttpClient(connectionManager);
    HttpParams params = httpClient.getParams();
    params.setParameter(ConnRoutePNames.LOCAL_ADDRESS, localAddress);

    return httpClient;
}

private static class Worker implements Runnable
{
    private List<HttpClient> clients = null;

    public Worker(List<HttpClient> clients)
    {
        this.clients = clients;
    }

    public void run()
    {

        do
        {
            HttpGet httpGet = new HttpGet("http://google.com/robots.txt");
            HttpClient client = this.clients.get(new Random().nextInt(this.clients.size()));

            try
            {
                HttpResponse httpResponse = client.execute(httpGet);
                EntityUtils.consume(httpResponse.getEntity());

                logger.debug("Success request");
            }
            catch(IOException e)
            {
                httpGet.abort();
                logger.info("IO error", e);
            }
        }
        while(true);
    }
}

Answer 1

Thanks for your answers!
Yes, the problem was with tcp/ip stack, but i was unable to get why 🙂

So, the reason explained below.

I created connection manager ( e.g. size=10 ), several clients ( each with its unique local address ), and several workers.
Each worker takes random client and executes the request.
So, at one time its possible that one client can be executed by up to 10 workers in parallel.
At another time another client can be executed by up to 10 workers in parallel.
Thats why, all 10 connections in manager can not be keep-alived, they are closing/creating all the time.

Sorry for bothering you!