sön 2009-02-08 klockan 01:18 +0100 skrev Steinar H. Gunderson: > I was asked for the same traces with client_persistent_connections and > server_persistent_connections off. Here they are -- the server ran much > shorter before I Ctrl-C-ed it, though, so the numbers will be smaller.
To get an even better trace the best is to build Squid with valgrind support (--with-valgrind-debug), let it run for a while under valgrind and then view the memory statistics page. squidclient mgr:mem This will trigger a runtime leak report while Squid is still running. Can also be repeated to look for changes after certain request patterns. The report is saved in cache.log or stderr if in foreground mode. But it's quite likely the leak can be identified from the traces you already sent. Regards Henrik
