OpenStack Architecture and Troubleshooting

• Basic Linux administration skills
• Basic networking knowledge
• Basic knowledge of cloud computing paradigm
• Basic practical and theoretical knowledge of OpenStack

The aim of this course is twofold - gain solid knowledge about OpenStack architecture and acquire profound troubleshooting skills. In the architecture part you will learn how to design the hardware for the public and private cloud, prepare high-availability control plane, choose proper size of the flavours, and safely overcommit cloud resources, among other things. Getting knowledge about the internals of Neutron (Networking Service) is another essential part in OpenStack architecture. Next the participant will be introduced into methods of troubleshooting issues in OpenStack clusters starting from general tips, through network troubleshooting up to caveats of main OpenStack services. Around 75% of the course comprises of hands-on workshops in the real OpenStack training environment.

• General architecture of OpenStack 
• OpenStack lab (VM) for this course
• Message queue and database
• Hardware considerations and capacity planning
  • Compute hardware
  • Network design
  • Storage design
  • Flavour sizing
  • Resource overcommitment
• Highly Available control plane
  • HA in OpenStack services
  • HA database
  • HA message queue
• Cloud partitioning and scheduler filters
  • Why and how implement cloud partitions (host-aggregates)
  • Nova scheduler filters
• Workload migration
  • Cold and live migration
  • Live migration tweaking
  • Watcher project
• In-depth OpenStack networking (SDN) (2-3h)
  • Types of network (local, flat, vlan, vxlan, gre)
  • Neutron plugins
    • Linux Bridge
    • Open vSwitch
  • Distributed Virtual Routers
  • LBaaS + Octavia project
  • VPNaaS
• Troubleshooting
  • General way of  diagnosing OpenStack issues
  • Troubleshooting network problems
  • Analyzing log files
  • Centralized logging
  • Debugging OpenStack client queries
  • Managing OpenStack database
  • Backing up OpenStack
  • Analyzing compute node status
  • Analyzing instance status
  • Analyzing AMQP broker (RabbitMQ)
  • Metadata services
  • Troubleshooting network performance
  • Instance backup and recovery
• OpenStack monitoring and telemetry
  • Ceilometer service
  • External monitoring
• Advances cloud/hypervisor features
  • CPU pinning / NUMA architecture
  • SR-IOV
• Cloud-init and image customization
  • Metadata Service
• Block storage backends
  • LVM
  • Ceph RBD
  • Physical appliances
  • Storage network considerations
• Upgrading OpenStack
  • Upgrade strategies and procedures
  • Zero-downtime upgrade
• Bare-metal provisioning with OpenStack
  • Ironic module
  • Undercloud and overcloud concepts
• Future of OpenStack