Reschedule during cold migration fails if compute RPC is pinned to 5.0
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
New
|
Undecided
|
Unassigned |
Bug Description
Steps to reproduce
==================
1) Set the [upgrade_
2) Boot a simple instance
3) Prepare your env that the first destination host for the next migration will fail during _prep_resize (for example the resize_claim fails)
4) Do the migration
Expected result
===============
The migration is re-scheduled to an alternative host where it succeeds.
Actual result
=============
The migration fails, the instance goes to ERROR state, and the log contains a nice stack trace:
2019-09-20 08:15:26,980 ERROR [nova.compute.
Traceback (most recent call last):
File "nova/compute/
File "nova/conductor
File "nova/conductor
return cctxt.call(context, 'migrate_server', **kw)
File "/home/
File "/home/
File "/home/
File "/home/
raise failure
AttributeError: 'dict' object has no attribute 'flavor'
Reproduction in functional test env coming soon.
Environment
===========
nova commit: fb3ba4443854683
tags: | added: rpc |
I believe I already have a fix for this: https:/ /review. opendev. org/#/c/ 680762/