Hello list,

I've been busy investigating how to best introduce HA to Ironic Inspector
and I got referred to TaskFlow.

With Ironic Inspector, besides handling other issues, each Node being
introspected undergoes couple of steps with side-effects.
This can be described using a finite automaton[1].
However, there are these constrains with that:
- User can Cancel the Introspection in the Waiting state
- Automaton advances from Waiting to Error if it reaches a Timeout
- Introspection can be restarted from either Error or Finished State (would
induce cyclic graph, not DAG)

Is there some best-practise how to address those?

Thanks,
milan


[1] Inspection State Transition System, Line #64,
http://www.fpaste.org/294451/
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to