> Are you deliberately compiling the adaptor from scratch? If nothing else I > guess that means you don't have to determine what architecture you're on. > I'm just thinking you could compile it once for each, store those somewhere > and download them with that script, then you could avoid installing gcc.
yeah, i know what you mean. at first i was just downloading the binary versions from mdimension but i thought building it each time will avoid any issues (or maybe even highlight issues) each time you boot a new instance, because each time you boot an amazon linux box you get the very latest version. i envisaged a situation where a minor upgrade to apache for example might rock the boat if we were using binaries, whereas building natively stands more chance of navigating around such issues. i don't have a lot of experience in this area - maybe someone with more knowledge of the woadaptor could recommend which is the best approach ? ultimately my vision for ec2 automation is that we have some uber-script that runs first to determine what sub-scripts you need to run based on the ami you are booting. we're going to publish some more scripts (hopefully this week) that are examples of how to auto-deploy your apps right the way from hudson to your freshly created ec2 instances... simon _______________________________________________ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
