Darrell Esau <darrell.e...@gmail.com> wrote: >Hi all, > >I'm running into a problem with tomcat, which I believe is >self-inflicted. > Mostly I just want to confirm my suspicions. > >I'm doing some fairly high-rate polling (using >JavaScript/XmlHttpRequest/DWR). I'm polling, which is just calling the >DWR >servlet once every second (using setTimeout on the response callback). > >Now.. So far nothing should seem out of the ordinary. And indeed many >of >these polling cycles do run and complete successfully. However after a >number of cycles (somewhere around 60), I start to get >NullPointerExceptions from the apache tomcat code. Looks like from the >Response.recycle method, etc. Full stack trace below. > >Now.. I don't think there's anything wrong with tomcat here.. I've been >using tomcat for years, and it's pretty rock solid. I suspect the >problem >is in my lack of understanding of the life cycle of the >request/response >objects. Here's the details in what I'm doing, and what I suspect I'm >doing wrong: > >This poll has the job of batching up "data requests" from the client >JavaScript code and passing them off to the server. On the server, >each >data request is spawned off to a new thread using a thread pool. After >the >threads are posted to the thread pool, the originating servlet request >returns. However, before the response is committed, when the thread is >created for the thread pool, I'm assigning the HttpServletRequest and >Response (and the HttpSession) as local thread variables. Once the >thread >is finished, I remove the references by assigning the local variables >to >null. > >So.. When I'm trying to preserve the request, response and session >objects >in a thread, knowing that the response will be committed before the >thread >dies, am I breaking a cardinal rule?
Yes. > Should I expect that tomcat will >throw NullPointerExceptions because of this? Yes. However, there is some good news. If you use the async API, the app can stay pretty much as is. The short version is: - use startAsync and pass the async context to your threads - have those threads write to the response (make sure you don't write concurrently) - call complete once all the threads have completed Mark > >Full stack trace:Aug 30, 2013 9:50:51 PM >org.apache.coyote.http11.Http11Processor process >SEVERE: Error finishing response >java.lang.NullPointerException >at >org.apache.tomcat.util.http.MimeHeaders.setValue(MimeHeaders.java:302) >at >org.apache.coyote.http11.Http11Processor.prepareResponse(Http11Processor.java:1558) >at >org.apache.coyote.http11.Http11Processor.action(Http11Processor.java:958) >at org.apache.coyote.Response.action(Response.java:181) >at >org.apache.coyote.http11.InternalOutputBuffer.endRequest(InternalOutputBuffer.java:403) >at >org.apache.coyote.http11.Http11Processor.process(Http11Processor.java:903) >at >org.apache.coyote.http11.Http11Protocol$Http11ConnectionHandler.process(Http11Protocol.java:606) >at >org.apache.tomcat.util.net.JIoEndpoint$Worker.run(JIoEndpoint.java:489) >at java.lang.Thread.run(Thread.java:680) >Aug 30, 2013 9:50:51 PM >org.apache.coyote.http11.Http11Protocol$Http11ConnectionHandler process >SEVERE: Error reading request, ignored >java.lang.NullPointerException >at org.apache.tomcat.util.http.MimeHeaders.clear(MimeHeaders.java:152) >at org.apache.coyote.Response.recycle(Response.java:579) >at >org.apache.coyote.http11.InternalOutputBuffer.nextRequest(InternalOutputBuffer.java:373) >at >org.apache.coyote.http11.Http11Processor.process(Http11Processor.java:925) >at >org.apache.coyote.http11.Http11Protocol$Http11ConnectionHandler.process(Http11Protocol.java:606) >at >org.apache.tomcat.util.net.JIoEndpoint$Worker.run(JIoEndpoint.java:489) >at java.lang.Thread.run(Thread.java:680) --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
