This appears to fix the problem. We are pushing to our integration server, and hope to update our supported technologies page soon.
Thanks all, Matt On Thu, Dec 8, 2016 at 1:09 AM, Mark Thomas <[email protected]> wrote: > On 08/12/2016 07:32, Violeta Georgieva wrote: > > 2016-12-08 3:48 GMT+02:00 Matthew Bellew <[email protected]>: > >> > >> I have narrowed this down quite a lot. This bug is caused by the same > >> Http11Processor being pushed on to the recycledProcessors stack twice. > I > >> discovered this by add a duplicates check in recycledProcessors.push() > > Kudos for tracking that down. That sort of bug can be difficult to find. > > <snip/> > > > May be you are facing the issue reported here [1] > > > > Can you test with 8.5.9 that we vote currently [2]? > > +1. I was going to suggest exactly the same thing. > > One of the strack traces I trimmed included one of the cancelledKey() > calls we removed as part of the fix for [1] so I am 99% sure that it is > the root cause of this problem. > > A quick test would be to disable sendfile with useSendfile="false" on > the <Connector ... />. However, you do really need to test with 8.5.9 to > be certain. > > Mark > > > > > Regards, > > Violeta > > > > [1] https://bz.apache.org/bugzilla/show_bug.cgi?id=60409 > > [2] http://marc.info/?l=tomcat-dev&m=148097070419476&w=2 > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
