Deployment - there is no single server

By Ulrich Habel (‎rhaen‎) from Augsburg.pm
Date: Saturday, 24 November 2012 11:20
Duration: 50 minutes
Target audience: Any
Language: English
Tags: automation dependency deployment tracking

The system environments which are running in the data centers today do have all different kinds of development servers, staging servers, production serers, qa servers and probably servers for customer tests. While the development guys are using continous integration build environments and unit tests to check if everything is working, the operations departments are often unable to deploy the software that has been successfully built.

The reasons may vary, however, some problems have to be dealt with all the time: automatic deployment and tracking dependencies. If you need to deploy Perl software across different environments, you want to make sure that everything runs smoothly without interaction. The talks shows some strategies how to deal with those kind of problems. It will show some development concepts which have been evolved in praxis. Setting up a development toolchain based on puppet/rpm can be one solution for dependeny problems. For those who are in the BSD worlds, the talk will explain how to use the NetBSD pkgsrc tree for software deployment.