hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Doug Cutting <cutt...@apache.org>
Subject Re: is it closed nutch-14 bug? (minimized efects)
Date Wed, 14 Feb 2007 21:34:09 GMT
Alvaro Cabrerizo wrote:
> We can see that all of the handlers (I launch the server using 5 handlers)
> are waiting for new queued calls, we can watch it in the next  piece of 
> code
> from the run method that belongs to class Handler:
> 
> while (running && callQueue.size()==0) { // wait for a call
>              callQueue.wait(timeout);    //line 471
>            }
> 
> In the other hand the Listener is waiting because it thinks queue is full.
> This is shown in the next piece of code from the processData method that
> belongs to class Connection:
> 
> while (running && callQueue.size() >= maxQueuedCalls) {
>          synchronized (callDequeued) {         // queue is full1
>          callDequeued.wait(timeout);         // wait for a dequeue for 60
> seconds (line 433)
>        }
> }
> 
> So Listener thinks queue is full and Handlers that it is empty, somebody
> must be wrong.

In trunk Server.java has since changed significantly, now using async 
i/o.  The callDequeued object is no longer used.  So I'm not sure that 
your concerns are still relevant.  (Nor am I certain they are not, but 
at least they need to be re-evaluated for the current code.)

Doug

Mime
View raw message