OpenStack Summit November 2014 Paris

Designed to deploy composite cloud applications, the OpenStack Heat orchestration engine is a convenient and powerful automation tool.  Unfortunately, when a failure happens during the provisioning and configuration of resources, diagnosing the cause of the problem can be a difficult proposition for the end user. In this talk, we first describe a number of ad hoc troubleshooting techniques which Heat users have been resorting to that have proved effective as well as the new enhancements delivered with the Juno release that could make this process easier.

In addition, we will describe recent proposals made by the community aiming to mitigate the troubleshooting pain, including a template debugger inspired by debuggers for traditional programming languages and making the OpenStack logs available to the end users for debugging.  Finally, we will propose a new approach for troubleshooting a wide range of failures by analyzing (1) the OpenStack logs as they are made available to the end users, (2) snapshots of the cloud state, and (3) the HOT template itself and demonstrate end-to-end scenarios helping end users diagnose the error, fix the template, and successfully redeploy.