Hi Chris, On 28.09.2013 01:49, Christopher Schultz wrote: > All, > > I know Mark and Rainer are working on trying to avoid corruption of what > looks like pollset counts or whatever, but I wanted to know if anyone > can help me decode the cause of the null-dereference that's occurring. > > Obviously bandaging the symptom is not the best solution, but keeping > the JVM up and running is a nice feature to have and, since I have no > understanding of APR polling, I can at least modify tcnative to throw an > exception instead of bombing the JVM. > > I can't seem to reproduce the problem using an APR connector on Linux: > reports seems to indicate that trivially accessing Tomcat via an HTTP > APR connector will cause a crash, and I was able to run the following > command without bringing down my instance: > > $ ab -n 10000 -c 50 http://localhost:28215/my-webapp/index.html
Did not yet try, but without "-k" ab doesn't use HTTP Keep-Alive, and I think the reports mentioned that connectionTimeout influenced things. So all test scenarios should probably include HTTP Keep-Alive. Try with "ab -k ..."? Still the whole poll stuff is platform dependent, so it might be you will not be able to reproduce the Windows crash even when the test scenario is exactly the same as there. > ... where localhost:28215 is bound to an > org.apache.coyote.http11.Http11AprProtocol connector. > > So, I need help decoding the tcnative-1.dll+0x7e23 pointer, which is > obviously a code pointer, but I have no idea how far into tcnative.dll > 0x7e23 points. It's obviously in ..._jni_Poll_poll() but I'm not sure > which pointer is likely to be null. > > Can anyone point me in the right direction? I currently can not really investigate in Windows, sorry. Regards, Rainer --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
