https://issues.apache.org/bugzilla/show_bug.cgi?id=56733
--- Comment #14 from Andy Wang <do...@moonteeth.com> --- Sorry I didn't mention this in my original bug report, but we didn't simply just see memory growth. I wouldn't have been worried if it was just that and it stabilized. The issue is that in a 32-bit environment, httpd actually crashes with memory allocation problems once the 32-bit address space is exhausted. So I reconfigured my workers.properties to a minimal config that allows me to reproduce this: worker.list=tomcat1 worker.maintain=60 worker.tomcat1.type=ajp13 worker.tomcat1.host=localhost worker.tomcat1.port=8010 worker.tomcat1.connection_pool_timeout=900 I'm using the default ThreadsPerChild for mpm_winnt (docs say 64) JkOptions +FlushPackets is enabled and MaxMemFree 2048 is set I did the following: 120000 HEAD requests to spool up all the connectors. I ran this in two 60000 batches. After each run the committed heap via vmmap.exe was consistently at 7580kb I ran ab -c 10 -n 10 to download my 600mb file and after that committed heap is at 29616kb after a second run it's at 51724kb I'm now running ab -c 10 -n 1000 And memory is still continually growing. -- 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
