Dave Shuck said the following on 08/19/2010 11:29 PM: > Thanks for the reply Peter. I think that this is definitely something > that needs some attention if we are going to consider GAE as a truly > viable solution for even modest apps. We're not the only ones with the problem. Look at this ugly thread for Grails on GAE (which is even supported by Google in their list):
http://jira.codehaus.org/browse/GRAILSPLUGINS-1736 If you have can get it running, one solution is to have a ping (using the GAE scheduler) to ping itself every 30 seconds to keep it running. That was one idea on the Grails thread. I have a few questions: * How many event-handlers do you have? This is one place we can more easily defer loading in Mach-II and of course any gains you might get on GAE depends on the number of event-handlers. I'm just trying to figure out a place to start any sort of performance changes in the terms of deferring in Mach-II for GAE by getting a better idea of your application. * Are you defining your views or are you using the view loader? The reason I ask is that "file" operations (and I quote that since it's all VFS) on GAE can be slow even if the files are concrete files in the deployment package. Maybe try straight up <page-view> definitions? HTH, .Peter -- Open BlueDragon Public Mailing List http://www.openbluedragon.org/ http://twitter.com/OpenBlueDragon official manual: http://www.openbluedragon.org/manual/ Ready2Run CFML http://www.openbluedragon.org/openbdjam/ mailing list - http://groups.google.com/group/openbd?hl=en
