Change Cron Trigger life cycle
There's a problem in HEAT with healing functionality because:
* Heat recreates object if it has been deleted not via Heat but Cron Trigger gets deleted by Mistral if its "remaining_
Proposed solution:
* Never delete a cron trigger (even if it's remaining executions is 0)
* Remove "unique" constraint for cron trigger "name" field because otherwise usability will suffer (it's inconvenient to invent new names for cron triggers all the time)
* Have two fields for remaining executions: one means initial value and is immutable, another one is actual value meaning how many executions are left
Blueprint information
- Status:
- Not started
- Approver:
- Renat Akhmerov
- Priority:
- Undefined
- Drafter:
- Renat Akhmerov
- Direction:
- Needs approval
- Assignee:
- None
- Definition:
- New
- Series goal:
- None
- Implementation:
-
Unknown
- Milestone target:
- None
- Started by
- Completed by
Related branches
Related bugs
Sprints
Whiteboard
While I am generally in favor of this - AFAIK this does not cause the described issue in Heat.
Heat will only recreate the resource after a stack-check was performed, the stack resource was changed to CHECK_FAILED and a stack-update is performed.
The Mistral Cron Trigger resource type does not implement the "handle_check" method which means that the logic of checking if the cron trigger exists is not there.
This means that the cron trigger stack resource will never go to CHECK_FAILED and will always be in CHECK_SUCCESS even if the cron trigger does not exist.
The fact that Heat does not create the cron trigger if it is missing can be considered as a bug and if that is fixed - this BP should be implemented.
