[undertow-dev] FormParserFactory getting stuck

[Hicks, Matt]

Well, it appears I fixed the problem.  After switching to the following
code everything is working fine now:

if (exchange.isInIoThread()) {
  exchange.dispatch(this);
} else {
  exchange.startBlocking();
  formParserBuilder.build().createParser(exchange).parseBlocking();
  ...
}

It's unfortunately that the non-blocking approach seems to be unreliable.

On Fri, Feb 24, 2017 at 7:18 PM Hicks, Matt <matt at matthicks.com> wrote:

> I gave it several minutes to run and finally just forcibly reloaded the
> page and it spun for a few seconds like it normally does, but I waited and
> eventually it threw this exception:
>
> appJVM[ERROR] Feb 24, 2017 7:15:45 PM org.xnio.ChannelListeners
> invokeChannelListener
> appJVM[ERROR] ERROR: XNIO001007: A channel event listener threw an
> exception
> appJVM[ERROR] java.lang.IllegalStateException
> appJVM[ERROR]   at
> io.undertow.server.protocol.framed.AbstractFramedStreamSinkChannel.getBuffer(AbstractFramedStreamSinkChannel.java:574)
> appJVM[ERROR]   at
> io.undertow.server.protocol.framed.AbstractFramedChannel.flushSenders(AbstractFramedChannel.java:629)
> appJVM[ERROR]   at
> io.undertow.server.protocol.framed.AbstractFramedChannel$FrameWriteListener.handleEvent(AbstractFramedChannel.java:951)
> appJVM[ERROR]   at
> io.undertow.server.protocol.framed.AbstractFramedChannel$FrameWriteListener.handleEvent(AbstractFramedChannel.java:948)
> appJVM[ERROR]   at
> org.xnio.ChannelListeners.invokeChannelListener(ChannelListeners.java:92)
> appJVM[ERROR]   at
> org.xnio.conduits.WriteReadyHandler$ChannelListenerHandler.writeReady(WriteReadyHandler.java:65)
> appJVM[ERROR]   at
> io.undertow.protocols.ssl.SslConduit$SslWriteReadyHandler.writeReady(SslConduit.java:1225)
> appJVM[ERROR]   at
> io.undertow.protocols.ssl.SslConduit$3.run(SslConduit.java:275)
> appJVM[ERROR]   at org.xnio.nio.WorkerThread.safeRun(WorkerThread.java:580)
> appJVM[ERROR]   at org.xnio.nio.WorkerThread.run(WorkerThread.java:464)
>
> I don't know if it's relevant to the problem or not, but at this point I'm
> just grasping for any straw.
>
> On Fri, Feb 24, 2017 at 5:18 PM Hicks, Matt <matt at matthicks.com> wrote:
>
> 4.) I have three threads "runnable":
>
> "XNIO-1 Accept at 6446" prio=5 tid=0x21 nid=NA runnable
>   java.lang.Thread.State: RUNNABLE
>  at sun.nio.ch.EPollArrayWrapper.epollWait(EPollArrayWrapper.java:-1)
>  at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:269)
>  at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:93)
>  at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:86)
>  - locked <merged>(a sun.nio.ch.EPollSelectorImpl)
>  - locked <merged>(a java.util.Collections$UnmodifiableSet)
>  - locked <merged>(a sun.nio.ch.Util$3)
>  at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:97)
>  at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:101)
>  at org.xnio.nio.WorkerThread.run(WorkerThread.java:509)
>
> "XNIO-1 I/O-7 at 6448" prio=5 tid=0x1f nid=NA runnable
>   java.lang.Thread.State: RUNNABLE
>  at sun.nio.ch.EPollArrayWrapper.epollWait(EPollArrayWrapper.java:-1)
>  at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:269)
>  at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:93)
>  at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:86)
>  - locked <merged>(a sun.nio.ch.EPollSelectorImpl)
>  - locked <merged>(a java.util.Collections$UnmodifiableSet)
>  - locked <merged>(a sun.nio.ch.Util$3)
>  at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:97)
>  at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:101)
>  at org.xnio.nio.WorkerThread.run(WorkerThread.java:509)
>
> "XNIO-1 I/O-3 at 6452" prio=5 tid=0x1b nid=NA runnable
>   java.lang.Thread.State: RUNNABLE
>  at sun.nio.ch.EPollArrayWrapper.epollWait(EPollArrayWrapper.java:-1)
>  at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:269)
>  at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:93)
>  at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:86)
>  - locked <0x1953> (a sun.nio.ch.EPollSelectorImpl)
>  - locked <0x1954> (a java.util.Collections$UnmodifiableSet)
>  - locked <0x1955> (a sun.nio.ch.Util$3)
>  at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:97)
>  at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:101)
>  at org.xnio.nio.WorkerThread.run(WorkerThread.java:509)
>
>
>
> On Fri, Feb 24, 2017 at 5:15 PM Hicks, Matt <matt at matthicks.com> wrote:
>
> 1.) I was running 1.4.10.Final, but I upgraded to 1.4.11.Final hoping it
> was a bug that was fixed, but it didn't make any difference.
> 2.) Yes
> 3.) No, of the 8 virtual CPUs none of them are maxed out.
> 4.) I'll have to get back to you...
>
> On Fri, Feb 24, 2017 at 5:10 PM Stuart Douglas <sdouglas at redhat.com>
> wrote:
>
> - What version of Undertow?
> - Is SSL in use?
> - Does the server enter some kind of spin loop (i.e. 100% CPU usage)?
> - What does the stack trace look like when this happens?
>
> Stuart
>
> On Sat, Feb 25, 2017 at 9:14 AM, Hicks, Matt <matt at matthicks.com> wrote:
> > I'm uploading files as "multipart/form-data" and using the FormParser as
> > follows:
> >
> >
> FormParserFactory.builder().build().createParser(exchange).parse(nextHandler)
> >
> > Sometimes this works fine, but very often my client shows progress up to
> a
> > certain percentage complete (monitoring the AJAX request) and then just
> > stops and never kicks to `nextHandler`, leaves the connection just
> sitting
> > there forever, and the server seems to stop accepting any future
> > connections.
> >
> > Any idea what might be causing this?
> >
> > _______________________________________________
> > undertow-dev mailing list
> > undertow-dev at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/undertow-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/undertow-dev/attachments/20170225/86b361d0/attachment-0001.html