Cinder

Live migration of server with attached volume in state backing-up

Registered by Niklas Schwarz

This blueprint has been superseded. See the newer blueprint "Move task status from volume_status to a dedicated field" for updated plans.

Currently a server cannot be live migrated if a attached volume is in state 'backing-up'. This is due to some drive limitations, but some backup drivers working with snapshots should be able to live migrate a server with a volume in status 'backing-up'.

The request to live migrate a server is handled via the nova api but the the state of the volume(s) is checked via cinder
and throws following exception:

ERROR nova.compute.manager [-] [instance: <instance-id>] Pre live migration failed at <node>:
nova.exception_Remote.InvalidInput_Remote: Invalid input received: Invalid volume: Volume <volume-id> status must be
available or downloading to reserve, but the current status is backing-up.

The suggestion is to retrieve additional valid states for volumes for different methods from the configured backup-driver to enable more functionality of the backup drivers.

Blueprint information

Status:
Complete
Approver:
None
Priority:
Undefined
Drafter:
Niklas Schwarz
Direction:
Needs approval
Assignee:
None
Definition:
Superseded
Series goal:
None
Implementation:
Unknown
Milestone target:
None
Completed by
Brian Rosmaita

Related branches

Sprints

Whiteboard

(?)

Work Items

This blueprint contains Public information 
Everyone can see this information.