{var%20f='http://v.t.sina.com.cn/share/share.php?appkey=1515056452',u=z||d.location,p=['&url=',e(u),'&title=',e(t||d.title),'&source=',e(r),'&sourceUrl=',e(l),'&content=',c||'gb2312','&pic=',e(p||'')].join('');function%20a(){if(!window.open([f,p].join(''),'mb',['toolbar=0,status=0,resizable=1,width=440,height=430,left=',(s.width-440)/2,',top=',(s.height-430)/2].join('')))u.href=[f,p].join('');};if(/Firefox/.test(navigator.userAgent))setTimeout(a,0);else%20a();})(screen,document,encodeURIComponent,'','','https://www.manongdao.com/data/attach/logo/logo.png', '推荐 做个烂人 的问题《apr_socket_recv: An established connection was abo》','https://www.manongdao.com/q-848934.html','页面编码gb2312|utf-8默认gb2312'));){kind=link}
I'm creating a small server using java.nio, but when trying to stress test it I keep getting messages about the connection being reset on the server side, or more specifically:
apr_socket_recv: An established connection was aborted by the software in your host machine
I've tried to narrow it down to the most simple of loops, but still no luck. I can get the error after a hundred or so connections, or maybe just after 1 or 2.
Here's the server loop:
byte[] response = ("HTTP/1.1 200 OK\r\n"
+ "Server: TestServer\r\n"
+ "Content-Type: text/html\r\n"
+ "\r\n"
+ "<html><b>Hello</b></html>").getBytes();
SocketChannel newChannel = null;
while (active) {
try {
//get a new connection and delegate it.
System.out.print("Waiting for connection..");
newChannel = serverSocketChannel.accept();
System.out.println("ok");
newChannel.configureBlocking(true);
newChannel.write(ByteBuffer.wrap(response));
}
catch (IOException e) {
e.printStackTrace();
}
finally {
try {
newChannel.close();
} catch (IOException ex) {
Logger.getLogger(Server.class.getName()).log(Level.SEVERE, null, ex);
}
}
}
I've tried checking if the write didn't write all requested byte, but it seemingly does. Interestingly enough, calling System.gc() after each newChannel.close() makes the problem disappear (but in return, it's horribly slow). So either I'm not releasing all resources I should release, or the application just needs a pause..
I'm losing all of my best years on this. Oh, and by the way.. if I ignore writing to the channel and just close after I accept the connection, the problem still doesn't go away.
From the docs for SocketChannel#Write (emphasis mine):
It's up to you to check the return value from the write call (which you're not doing presently), and issue successive write calls until the whole of the buffer has been sent. Something like this, I guess:
You'll obviously get aborts if you don't write all of your response data and then just close the socket.
Well I found it out, so I might as well share it.
My app needed a pause. It was simply going too fast, and closing the connection before the client had written all of its request data. The fix would be to keep on reading until the entire HTTP request had been received. D'oh.. lesson learned.