One of the biggest questions for the operational teams supporting our website at the University of Edinburgh following our move of the central University CMS to Drupal is how we update the site with no downtime. For us, the answer is Continuous Integration, a cornerstone of DevOps. In this session we will build on our talk at DrupalCamp Scotland 2014 on automated deployment, diving into the detail of our deployment process to demonstrate how we reconcile the business demand of 100% site availability with the need to deploy regular updates to our Drupal CMS.

We will talk about how our existing enterprise CI and automated deployment tools can help us to achieve the 'no downtime' goal, covering:
-- how we automate our build with SVN, Bamboo, Ant and drush;
-- how we replicate our production environment into a staging environment during deployment;
-- how our automated test suite, run against the replicated production environment, gives confidence that it is safe to deploy code into production;
-- how our infrastructure supports staged deployments into production, ensuring that the University website always remains available;
-- how we handle database changes without affecting public traffic to the University website;
-- how we harness varnish and caching to support our production deployment.

Our session will show that with a robust automated deployment process and creative use of CI tools, change needn't mean page unavailable.