Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

Swift and Cinder high availability


Do Swift or Cinder provide high availability for all components in the storage system? What happens to a running virtual machine if any of the following happens?

  • a hard drive where the image is stored fails
  • a switch that connects the compute node to the storage nodes is powered off
  • a storage node kernel panics and is taken offline

Is the redundancy and high availability in Swift or Cinder able to seamlessly handle these failures, so that the running VM is not interrupted?

Based on what I've read about Swift and Cinder (including this page, it sounds like they protect against single points of failure, but can you please elaborate on this?