Mitigation/Recovery from an OpenStack Clustering Outage Replacing a Controller

Solution Verified - Updated -

Issue

  • A controller went down and is not recoverable.
  • The OS is corrupt on a controller.
  • A controller needs to be replaced in the environment, but time is needed to schedule a maintenance window and prepare for the activity.
  • How to restore service to an HA environment that has lost one controller?
  • Galera cluster will not sync after the loss of a controller and the remaining two nodes created a split cluster.

Environment

  • Red Hat OpenStack Platform 8
  • Red Hat OpenStack Platform 10
  • Red Hat OpenStack Platform 13

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In