Use stack-update to attempt recovery of failed create

Registered by Steve Baker

Currently if a stack fails on create or update then the only recovery option is to delete and recreate the stack.

It should be possible to recover a failed stack by performing a stack update. Any failed resources can be replaced and an attempt can be made to bring the entire stack into a completed state.

After some discussion, it was determined that this approach can only work properly for a failed CREATE. For a failed update, we will need to keep track of the progress of the update, so that we do not repeat actions it has already completed. That work is under the scope of this blueprint:

https://blueprints.launchpad.net/heat/+spec/update-failure-recovery

Blueprint information

Status:
Complete
Approver:
Steven Hardy
Priority:
Undefined
Drafter:
Steve Baker
Direction:
Approved
Assignee:
Zane Bitter
Definition:
Superseded
Series goal:
None
Implementation:
Good progress
Milestone target:
None
Started by
Clint Byrum
Completed by
Zane Bitter

Related branches

Sprints

Whiteboard

Gerrit topic: https://review.openstack.org/#q,topic:bp/retry-failed-update,n,z

Addressed by: https://review.openstack.org/62936
    Retry failed actions on update

(?)

Work Items