Maxim wrote: > Hmm, it seems like we solved problem just partially. > With > > MonoMaxActiveRequests 10 > MonoMaxWaitingRequests 100 > > settings there are a lot messages like "Maximum number (10) of > concurrent mod_mono requests to > /tmp/mod_mono_dashboard_nike-outdoor.ru_2.lock reached. Droping request." > and some visitors get "Service Temporary Unavailable" error.
Well that's exactly what's going to happen when you limit the number of connections --- the hope is that with a reasonably large MonoMaxWaitingRequests it doesn't happen often. > Is option "MonoMaxWaitingRequest" implemented? Yes. (I did it myself...) If you access a page during high load and see your browser waiting with no response for a while, that's MonoMaxWaitingRequest in action holding onto the request until it gets a chance to be processed. > Is the only way to solve this situation in upgrading server CPU? Anything that will make your site process requests faster (which might involve improving mono like finding out why more concurrency causes problems, etc.). -- - Josh Tauberer http://razor.occams.info "Yields falsehood when preceded by its quotation! Yields falsehood when preceded by its quotation!" Achilles to Tortoise (in "Godel, Escher, Bach" by Douglas Hofstadter) > > Best wishes, > Max Karavaev > >> Hello! >> >> It seems like we solved this problem. >> Joshua is right. >> >> Settings >> MonoMaxActiveRequests 10 >> MonoMaxWaitingRequests 100 >> fixed exponentially grow of connections by reducing concurrent processes. >> On other server with two dual core xeons this problem never appear, but >> for other with single core duo 2 these settings are very important. >> >> And settings >> MonoAutoRestartMode Requests >> MonoAutoRestartRequests 10000 >> reduce memory consumption by restarting hight loaded websites. >> May be with compacting GC this option become obsolete. I've never seen >> such a situation on IIS+.Net. >> >> It would be great to write some tips about these situations on mod_mono >> FAQ page. >> >> Thanks a lot for help! >> >> All the best, >> Max Karavaev >> >> >>> Hello Maxim, >>> >>> We are unaware of this problem, do you think you could provide a >>> test case that illustrates this problem so we can look into it? >>> >>> >> _______________________________________________ >> Mono-list maillist - [email protected] >> http://lists.ximian.com/mailman/listinfo/mono-list >> >> _______________________________________________ Mono-list maillist - [email protected] http://lists.ximian.com/mailman/listinfo/mono-list
