Router VMs that are slow to delete cause a tight loop in the state machine
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
akanda |
New
|
Medium
|
Unassigned |
Bug Description
We've noticed in situations where Nova is slow to delete router VMs that rug workers can get caught in a tight loop, preventing other events from being processed, e.g.,
2015-05-18 18:47:07:
2015-05-18 18:47:07:
2015-05-18 18:47:07:
2015-05-18 18:47:07:
2015-05-18 18:47:08:
2015-05-18 18:47:09:
2015-05-18 18:47:09:
2015-05-18 18:47:09:
2015-05-18 18:47:09:
2015-05-18 18:47:11:
2015-05-18 18:47:12:
2015-05-18 18:47:12:
2015-05-18 18:47:12:
2015-05-18 18:47:12:
2015-05-18 18:47:12:
2015-05-18 18:47:13:
2015-05-18 18:47:13:
2015-05-18 18:47:13:
2015-05-18 18:47:13:
2015-05-18 18:47:14:
It might be nice for this process to be asynchronous in the same way as `CreateVM -> CheckBoot`.
Changed in akanda: | |
importance: | Undecided → Medium |
tags: | added: akanda-rug |