In this workshop, we will focus on analyzing some very common network issues that administrators encounter when managing an OpenStack cloud. The workshop will cover several popular network configuration options such as Open vSwitch, Designate, and Octavia. We will follow packets as they flow deep underneath the surface travelling in and out of running instances. We also will investigate how instances are automatically configured with resolvable DNS names. In addition, we will discover what happens during network load balancing. We will use a set of standard opensource tools to read log files, capture packets and verify network flow. The attendees will each be given their own instance and will work through each of the scenarios. They will gain an understanding of how each of these operations works. They will learn what to investigate when they are called upon to resolve issues found in their OpenStack clouds.
All attendees should bring a laptop to participate in this workshop.
Attendees will learn how to troubleshoot in detail several OpenStack workflows.
This workshop will leverage the following tools:
- tcpdump (packet capture)
- Linux network namespaces (ip netns)
- OVS
- lnav (logfile viewer)
The workflows covered will be:
- Analyzing and capturing packet flow in and out of nova instances.
- Tracing dns name generation for nova instances.
- Deploying and verifying Octavia Load Balancers.
NOTE!!! This is a more advanced edition of the session given at Vancouver. Because of the feedback we received, we are offering a more advanced session. This is intended for users that want a deeper understanding of common workflows in OpenStack.
We also submitted the Vancouver session again as it was very well attended and we imagine a different geographical audience in Berlin. The title of that submission is "Leveraging Your OpenStack Troubleshooting Tool Box - Learning the Basics." If both sessions are accepted, please schedule this one after the beginner session.
