On Jun 2, 2005, at 9:43 AM, Dain Sundstrom wrote:
How do we set it up so that if the pool is undersized then we get
an error rather than just have the transport hang?
I'm looking into this one. The default for the current thread pool
is to have the requesting thread run the task, which bad for our
case since the requesting thread is the socket acceptor. I think
we should instead set abortWhenBlocked on the thread pool so we get
an exception. It would be nice if there were a wait 10sec option
to deal with floods, but I don't see any options for that.
The problem was in the ThreadPool class we were using a work queue in
addition to the thread pooling, so when we ran out of threads the
work was just queued forever. Normally a small fixed size buffer is
desireable, since it can even out connection bursts, but since we had
a total of one threads available any code that needs two threads to
work (like a test using a timer) would lock up. For now I have
turned off the work queue and set the pool to throw an exception if a
thread can not be immediately acquired. This will help us drive out
other bugs in the system. Of course later on when tuning the server,
the thread pool will be an obvious place to start.
-dain
