Book Image

Troubleshooting OpenStack

By : Tony Campbell
Book Image

Troubleshooting OpenStack

By: Tony Campbell

Overview of this book

OpenStack is a collection of software projects that work together to provide a cloud fabric. OpenStack is one of the fastest growing open source projects in history that unlocks cloud computing for everyone. With OpenStack, you are able to create public or private clouds on your own hardware. The flexibility and control afforded by OpenStack puts the cloud within reach of anyone willing to learn this technology. Starting with an introduction to OpenStack troubleshooting tools, we’ll walk through each OpenStack service and how you can quickly diagnose, troubleshoot, and correct problems in your OpenStack. Understanding the various projects and how they interact is essential for anyone attempting to troubleshoot an OpenStack cloud. We will start by explaining each of the major components and the dependencies between them, and move on to show you how to identify and utilize an effective set of OpenStack troubleshooting tools and fix common Keystone problems. Next, we will expose you to common errors and problems you may encounter when using the OpenStack Block Storage service (Cinder). We will then examine Heat, the OpenStack Orchestration Service, where you will learn how to trace errors, determine their root cause, and effectively correct the issue. Finally, you will get to know the best practices to architect your OpenStack cloud in order to achieve optimal performance, availability, and reliability.
Table of Contents (17 chapters)
Troubleshooting OpenStack
About the Author
About the Reviewer

Neutron services and agents

A properly running Neutron service will launch several Linux processes. When troubleshooting, you will want to make sure that all services are configured correctly and are running actively. If you run pgrep –l neutron, you should see a list of Neutron services, as shown in the following screenshot:

Depending on how you have deployed Neutron, the services listed in the preceding screenshot may be spread across multiple servers. For example, your neutron-server processes may be running on your controller node, all of the agents may be on your Network node, and your Compute nodes may also be running some of the agents.

To confirm that all the agents are running as expected, you can execute the neutron agent-list command. The output of this will look similar to what is shown in the following screenshot:

When the agent is up and running, the alive column will have :-), and when the agent is not alive, the value will be xxx. If an agent is not alive and you are expecting...