After rolling back to old Nio all clients works very well as expected.
I'm sorry, I don't know how to reproduce this kind of issue on a test case, I have tried to write some simple multi-thread tests but I cannot break Tomcat Nio2 connector, it happens only on my QA environment. At this point I'm not able to submit any issue to your bug tracker. Thank you 2017-01-13 0:43 GMT+01:00 Mark Thomas <ma...@apache.org>: > On 12/01/2017 21:09, Enrico Olivelli wrote: > > Thank you Chris > > > > Il gio 12 gen 2017, 18:04 Christopher Schultz < > ch...@christopherschultz.net> > > ha scritto: > > > > Enrico, > > > > On 1/12/17 8:53 AM, Enrico Olivelli wrote: > >>>> I'm upgrading from Tomcat 8.0.33. I see that after a period of work > >>>> requests remains "pending", for instance I get all clients > >>>> remaining waiting for a response (parsing HTTP Response header) and > >>>> no active thread on my Tomcat. > >>>> > >>>> This happens in my QA environment where I start several > >>>> WebDriver/Unit tests againts my web application. I have many cases, > >>>> the most simple I this is the following: > >>>> > >>>> on the client side (a JAX-WS client): > >>>> > >>>> "main" #1 prio=5 os_prio=0 tid=0x00007f7cf0009000 nid=0x7ddc > >>>> runnable [0x00007f7cf6e3f000] > >>>> > >>>> java.lang.Thread.State: RUNNABLE > >>>> > >>>> at java.net.SocketInputStream.socketRead0(Native Method) > >>>> > >>>> at > >>>> java.net.SocketInputStream.socketRead(SocketInputStream.java:116) > >>>> > >>>> at java.net.SocketInputStream.read(SocketInputStream.java:170) > >>>> > >>>> at java.net.SocketInputStream.read(SocketInputStream.java:141) > >>>> > >>>> at java.io.BufferedInputStream.fill(BufferedInputStream.java:246) > >>>> > >>>> at java.io.BufferedInputStream.read1(BufferedInputStream.java:286) > >>>> > >>>> at java.io.BufferedInputStream.read(BufferedInputStream.java:345) > >>>> > >>>> - locked <0x00000000fd159a18> (a java.io.BufferedInputStream) > >>>> > >>>> at > >>>> sun.net.www.http.HttpClient.parseHTTPHeader(HttpClient.java:704) > > > > Okay, client is waiting. > > > > > >> From server side application logs I see that the webapp has processed > the > >> request. In this case I am using Metro JAX-WS. > >> It seems that the reponse is waiting in some buffer. > >> Is there any way to inspect tomcat buffers? Network buffers? I am using > >> linux. > > > >> The strange thing is that with Tomcat 8.0.33 there is no problem. > >> The same problem happens even with other servlets of different > frameworks > >> and other webapps. > > > >> On 8.0.x I was using Nio connector, on 8.5 I am using Nio2. > >> The problem is both on http and jsse https. > > Try NIO. > > Mark > > > > > >> Is any new relevant default value changed from 8.0 and 8.5? I can't find > >> any idea on changelogs or documentation > > > > > > > > > > > >>>> in the server side: Full thread dump Java HotSpot(TM) 64-Bit Server > >>>> VM (25.92-b14 mixed mode): > >>>> > >>>> > >>>> "anInnocuousThread" #861 daemon prio=5 os_prio=0 > >>>> tid=0x00007ff23c0de800 nid=0xdb26 runnable [0x00007ff1ad379000] > >>>> java.lang.Thread.State: RUNNABLE at > >>>> sun.nio.ch.EPoll.epollWait(Native Method) at > >>>> sun.nio.ch.EPollPort$EventHandlerTask.poll(EPollPort.java:194) at > >>>> sun.nio.ch.EPollPort$EventHandlerTask.run(EPollPort.java:268) at > >>>> java.lang.Thread.run(Thread.java:745) at > >>>> sun.misc.InnocuousThread.run(InnocuousThread.java:74) > >>>> > >>>> (lots of this kind....) (HTTPS Connector I think) > > > > Without the rest of the stack trace, it's hard to tell if those are okay > > > > > >> I have cut all the other threads which had the same stacktrace. > >> I have stripped out GC threads, JMX... > >> My application is not present in any thread but I have several clients > >> stuck at waiting for a response. > > > >> As soon as I can I will attach the full dump > > > >> Thank you very much > >> I appreciate your help > > > >> Enrico > > > >> . > > > >>>> "https-jsse-nio2-10.168.10.55-8443-exec-130" #572 daemon prio=5 > >>>> os_prio=0 tid=0x00007ff23c0a0000 nid=0xd4a2 waiting on condition > >>>> [0x00007ff1b2bbf000] java.lang.Thread.State: WAITING (parking) at > >>>> sun.misc.Unsafe.park(Native Method) - parking to wait for > >>>> <0x0000000080f1d420> (a > >>>> java.util.concurrent.locks.AbstractQueuedSynchronizer$ > ConditionObject) > > > > These > >>>> > > are request-processing threads, waiting for work. They are idle. > > > >>>> (lots of this kind....) (HTTP Connector I think) > >>>> > >>>> "http-nio2-10.168.10.55-8080-exec-21" #281 daemon prio=5 os_prio=0 > >>>> tid=0x00007ff2340e7000 nid=0xd262 waiting on condition > >>>> [0x00007ff1af995000] java.lang.Thread.State: WAITING (parking) at > >>>> sun.misc.Unsafe.park(Native Method) - parking to wait for > >>>> <0x0000000080f0cdb0> (a > >>>> java.util.concurrent.locks.AbstractQueuedSynchronizer$ > ConditionObject) > > > > Same > >>>> > > here. > > > > The others look okay to me. > > > > -chris > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > >> For additional commands, e-mail: users-h...@tomcat.apache.org > >> > >> -- > > > > > > -- Enrico Olivelli > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > For additional commands, e-mail: users-h...@tomcat.apache.org > >