https://bz.apache.org/bugzilla/show_bug.cgi?id=63916
--- Comment #13 from Michael Osipov <micha...@apache.org> --- (In reply to Mark Thomas from comment #11) > Simpler test case that reproduces the issue > > https://github.com/markt-asf/tomcat-bugs/blob/master/src/java/org/apache/ > tomcat/Bug63916NioPoller.java > > Run the test case and connect with telnet to localhost 8080. > Uncomment the line that sets the send buffer to 8k and watch the time taken > to write the 10MB jump from ~400ms to ~50s. > > Tomorrow's task is to experiment to see if I can figure out the root cause > of the time difference. Ran the test on the same FreeBSD server: Without explicit buffer size: > Default send buffer size is [49032] > Writing 10MB took [2601] milliseconds With explicit buffer size of 8 KiB: > Writing 10MB took [135095] milliseconds That's insane! On HP-UX with Java 8 for comparison: Without explicit buffer size: > Default send buffer size is [32768] > Writing 10MB took [2851] milliseconds With explicit buffer size of 8 KiB: > Writing 10MB took [89457] milliseconds and on RHEL 7, Java 8u232: Without explicit buffer size: > Default send buffer size is [1313280] > Writing 10MB took [661] milliseconds With explicit buffer size of 8 KiB: Writing 10MB took [50696] milliseconds -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
