Re: [jetty-users] Jetty hanging after upgrade from 6.1.14 to 7.1.15

[Greg Wilkins <gregw@xxxxxxxxxxx>]

Nick,

I had a quick look, but cannot follow what you have said/sent.
I cannot see any SelectChannelConnectors in your thread dump.  All I
see is a LocalConnector:

    [junit] "qtp1177106648-212 - Acceptor0 LocalConnector@0.0.0.0:0"
prio=5 tid=17970d000 nid=0x16cf9b000 waiting on condition [16cf9a000]
    [junit]    java.lang.Thread.State: WAITING (parking)
    [junit] 	at sun.misc.Unsafe.park(Native Method)
    [junit] 	- parking to wait for  <10fb31178> (a
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
    [junit] 	at
java.util.concurrent.locks.LockSupport.park(LockSupport.java:158)
    [junit] 	at
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:1925)
    [junit] 	at
java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:399)
    [junit] 	at
org.eclipse.jetty.server.LocalConnector.accept(LocalConnector.java:59)
    [junit] 	at
org.eclipse.jetty.server.AbstractConnector$Acceptor.run(AbstractConnector.java:793)
    [junit] 	at
org.eclipse.jetty.util.thread.QueuedThreadPool$2.run(QueuedThreadPool.java:436)
    [junit] 	at java.lang.Thread.run(Thread.java:637)


can you clarify what your configuration was, and perhaps raise an
issue in the jetty JIRA

Note that 7.1.5 did have a few problems (though nothing obviously
related), so try 7.1.6 now.

regards









On 14 July 2010 10:11, Nick Pellow <npellow@xxxxxxxxxxxxx> wrote:
> Hi,
> We start and stop Jetty during the setUp() and tearDown() methods in some of
> our tests.
> We recently tried upgrading from Jetty 6.1.14, to Jetty 7.1.15 - however had
> to revert this for the time being because some of our tests started randomly
> hanging.
>
> After Jetty starts, our tests use HttpClient to make requests to FishEye
> running inside Jetty.
> It is one of these client requests that hang, indefinitely. From what I can
> tell, this is more easily reproducable on machines that run quite slow, and
> have few (<3) cores. (e.g a mac-mini, but not a 8-core MacPro)
>
> It appears that the SelectChannelConnector we configure Jetty with to handle
> remote requests is no longer running (or was never started?) when the hang
> occurs.
>
> There are three DEBUG logs, including thread dumps, attached
> to http://jira.atlassian.com/browse/CRUC-3721, showing:
>
> a build that hangs using a SelectChannelConnector
> a build that doesn't hang using a SelectChannelConnector (only a single test
> is run)
> a built that hangs using a normal (non-nio) SocketConnector.
>
> Has anyone seen similar behaviour using Jetty 7 before, or does anything in
> the logs standout as a candidate for causing Jetty to hang?
>
> Is there anything else we can enable to try and detect what could be going
> on?
>
> Cheers,
>
> Nick
>
> _______________________________________________
> jetty-users mailing list
> jetty-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/jetty-users
>
>