You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hc.apache.org by "Raul (JIRA)" <ji...@apache.org> on 2012/10/15 09:12:03 UTC
[jira] [Commented] (HTTPCLIENT-1251) PoolingClientConnectionManager
and multithreading problem
[ https://issues.apache.org/jira/browse/HTTPCLIENT-1251?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13475999#comment-13475999 ]
Raul commented on HTTPCLIENT-1251:
----------------------------------
Ok, It's your opinion. I wrote my code to put the people in context. My conclusion Israel that if i reuse the threads to lancha the repuestos, http clientes doesn't Works as spected
> PoolingClientConnectionManager and multithreading problem
> ----------------------------------------------------------
>
> Key: HTTPCLIENT-1251
> URL: https://issues.apache.org/jira/browse/HTTPCLIENT-1251
> Project: HttpComponents HttpClient
> Issue Type: Bug
> Components: HttpConn
> Affects Versions: 4.2.1
> Reporter: Raul
>
> I have developed a miltithread application for crawling a remote web page.
> Client reused in all threads is initialized in this way
> PoolingClientConnectionManager cm = new PoolingClientConnectionManager();
> int maxConnections = crawlerParameters.getIntegerParameter(CrawlerParams.CLIENT_MAX_CONNECTIONS_PARAM,200);
> cm.setMaxTotal(maxConnections);
> cm.setMaxPerRoute(new HttpRoute(new HttpHost(baseUriBuilder.getHost())), maxConnections);
> cm.setDefaultMaxPerRoute(maxConnections);
> httpclient = new DefaultHttpClient(cm);
> The request to execute are stored in a queue (thread safe). Then a certain number of threads (configuration parameter) are launched. The overall execution finish when all threads finish.
> ConcurrentLinkedQueue<QueueElement> queue = CrawlerUtils.buildQueue(baseUrl, crawlerParameters, page, replacements);
> int threadsPerPage = crawlerParameters.getIntegerParameter(CrawlerParams.CLIENT_THREADS_PER_PAGE_PARAM, 4);
> FetcherThread[] threads = new FetcherThread[threadsPerPage];
> final CountDownLatch countDownLatch = new CountDownLatch(threadsPerPage);
> for (int i = 0 ; i<threadsPerPage; i++){
> threads[i] = new FetcherThread(String.valueOf(i),type, httpclient, crawlerParameters, queue, page, xmlMarshaller,crawlBaseDir,countDownLatch);
> }
>
> // start the threads
> for (int j = 0; j < threads.length; j++) {
> threads[j].start();
> }
> //Wait until all threads finish
> countDownLatch.await();
> This is the thread detail. HttpClient, request queue and the countdown latch are passed as arguments in the constructor....
> public class FetcherThread extends Thread {
> private transient Logger _log = LoggerFactory.getLogger(FetcherThread.class);
> private ConcurrentLinkedQueue<QueueElement> queue;
> private CrawlPage page;
> private XStreamMarshaller xmlMarshaller;
> private String id;
> private DefaultHttpClient httpclient;
> private String crawlBaseDir;
> private String type;
> private CountDownLatch countDownLatch;
> private CrawlerParams crawlerParameters;
> public FetcherThread(final String id, String type, final DefaultHttpClient httpclient, final CrawlerParams crawlParameters, final ConcurrentLinkedQueue<QueueElement> queue, final CrawlPage page, final XStreamMarshaller xmlMarshaller, final String crawlBaseDir, final CountDownLatch countDownLatch) {
> this.queue = queue;
> this.page = page;
> this.xmlMarshaller = xmlMarshaller;
> this.id = id;
> this.httpclient = httpclient;
> this.crawlBaseDir = crawlBaseDir;
> this.type = type;
> this.countDownLatch = countDownLatch;
> this.crawlerParameters = crawlParameters;
> super.setName(this.page.getName()+"_"+id);
> }
>
> /**
> * Executes the GetMethod and prints some status information.
> */
> @Override
> public void run() {
> try{
> if (_log.isInfoEnabled()){
> _log.info("Thread request {} started.", this.getName());
> }
> while (true) {
> QueueElement element = queue.poll();
> if (element == null) break;
> processElement(element);
> }
> }finally{
> countDownLatch.countDown();
> }
> }
>
>
> private void processElement(QueueElement element){
> try{
> //Process
> if (_log.isDebugEnabled()){
> _log.debug("Request executor thread {}, Starting request {}",this.getName(),element.getRequest().getURI().toString());
> }
> StopWatch totalwatch = new StopWatch();
> StopWatch watch = new StopWatch();
> totalwatch.start();
> watch.start();
> processRequest(element, new BasicHttpContext(),crawlerParameters);
> watch.stop();
> long duration = watch.getLastTaskTimeMillis();
> Integer minDurationStr = crawlerParameters.getIntegerParameter(CrawlerParams.CLIENT_MIN_REQUEST_PROCESS_DURATION_PARAM);
> if (minDurationStr!=null){
> Long minDuration = Long.valueOf(minDurationStr);
> if (duration<minDuration){
> FetcherThread.sleep(minDuration-duration);
> }
> }
> totalwatch.stop();
> if (_log.isDebugEnabled()){
> _log.debug("Request executor thread {}, request {} --> performed in "+totalwatch.getLastTaskTimeMillis()+" millis", id, element.getId());
> }
> }catch (Exception e){
> _log.error("Exception: {}",e);
> element.getRequest().abort();
> }finally{
> element.getRequest().releaseConnection();
> }
> }
> The process method only performs the request execution and reads the content....
> HttpResponse response = httpclient.execute(element.getRequest(),httpContext);
> String content = readEntity(response);
> The problem is that for one request i receive the response from other request....
> The correct way should be
> request A ---> response A
> request B ---> response B
> But when i configure more than one thread polling request from the queue....
> request A --> response B
> request B ---> ressponse A
> If i configure only one fetcher thread, all works well...
> I tried with one context per thread, and one context per request, but with the same wrong result....
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org