2015-08-16 35 views
0

我正在用HttpAsyncClient和Jsoup为特定网站编写Web刮取器。我正在使用一个HttpAsyncClient实例,它执行来自不同线程的多个请求。经过几次成功的请求,客户端似乎挂起并不会执行更多请求。调用execute方法通常会返回FutureCallback对象,但其方法(completedfailedcancelled)都不会执行。Apache的HttpAsyncClient在执行后永远不会返回

通过查看日志看来,如果请求永远不会发送:

33572 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.client.MainClientExec - [exchange: 5] start execution 
33573 [I/O dispatcher 1] DEBUG org.apache.http.client.protocol.RequestAddCookies - CookieSpec selected: default 
33575 [I/O dispatcher 1] DEBUG org.apache.http.client.protocol.RequestAddCookies - Cookie [version: 0][name: .ASPXAUTH][value: removed][domain: www.predictit.org][path: /][expiry: null] match [(secure)www.predictit.org:443/Home/SingleOption] 
33576 [I/O dispatcher 1] DEBUG org.apache.http.client.protocol.RequestAddCookies - Cookie [version: 0][name: ARRAffinity][value: removed][domain: www.predictit.org][path: /][expiry: null] match [(secure)www.predictit.org:443/Home/SingleOption] 
33577 [I/O dispatcher 1] DEBUG org.apache.http.client.protocol.RequestAddCookies - Cookie [version: 0][name: ASP.NET_SessionId][value: removed][domain: www.predictit.org][path: /][expiry: null] match [(secure)www.predictit.org:443/Home/SingleOption] 
33579 [I/O dispatcher 1] DEBUG org.apache.http.client.protocol.RequestAddCookies - Cookie [version: 0][name: __RequestVerificationToken][value: removed][domain: www.predictit.org][path: /][expiry: null] match [(secure)www.predictit.org:443/Home/SingleOption] 
33579 [I/O dispatcher 1] DEBUG org.apache.http.client.protocol.RequestAuthCache - Auth cache not set in the context 
33580 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.client.InternalHttpAsyncClient - [exchange: 5] Request connection for {s}->https://www.predictit.org:443 
33580 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.conn.PoolingNHttpClientConnectionManager - Connection request: [route: {s}->https://www.predictit.org:443][total kept alive: 1; route allocated: 1 of 2; total allocated: 1 of 20] 
33581 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.conn.ManagedNHttpClientConnectionImpl - http-outgoing-0 192.168.1.6:38559<->23.96.96.142:443[ACTIVE][r:r][ACTIVE][r][NOT_HANDSHAKING][0][0][0][0]: Set timeout 0 
33581 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.conn.PoolingNHttpClientConnectionManager - Connection leased: [id: http-outgoing-0][route: {s}->https://www.predictit.org:443][total kept alive: 0; route allocated: 1 of 2; total allocated: 1 of 20] 
33582 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.client.InternalHttpAsyncClient - [exchange: 5] Connection allocated: CPoolProxy{http-outgoing-0 [ACTIVE]} 
33582 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.conn.ManagedNHttpClientConnectionImpl - http-outgoing-0 192.168.1.6:38559<->23.96.96.142:443[ACTIVE][r:r][ACTIVE][r][NOT_HANDSHAKING][0][0][0][0]: Set attribute http.nio.exchange-handler 
33583 [I/O dispatcher 1] DEBUG org.apache.http.impl.nio.conn.ManagedNHttpClientConnectionImpl - http-outgoing-0 192.168.1.6:38559<->23.96.96.142:443[ACTIVE][rw:r][ACTIVE][rw][NOT_HANDSHAKING][0][0][0][0]: Event set [w] 

它只是挂在那儿,在最后一行并没有任何反应。它似乎是某种死锁,但CloseableHttpAsyncClient应该是线程安全的。

您可以看到代码here,如果您使用these依赖关系创建maven项目,您可以自己运行它(它有一个主要方法)。

我感谢您的帮助。

回答

0

看起来我是通过在同一个HttpAsyncClient返回的FutureCallback的completed方法内调用execute方法导致死锁的。我更改了代码,以便completed方法返回方法完成后处理的链接列表,现在所有工作都可以使用。

+1

此处记录问题和解决方法:https://issues.apache.org/jira/browse/HTTPCLIENT-1805 –

相关问题