296856101f
Currently, when an audit with strategy zone_migration has added at least one volume_migration action, it will not process the instances migrations according to the definition of the `compute_nodes` parameter. This behavior is unexpected according to the documentation of the strategy. This patch is fixing that behavior and making sure that not duplicated actions are added to the solution, to handle the case where instances migration actions are created when analyzing the volumes if the `with_attached_volume` parameter is enabled. The patch is also removing the method `instances_no_attached` which is not longer used. Finally, it's adding some unit tests for the new method and fixing the ones to cover the mixed instances and volumes migration situation. Closes-Bug: #2109722 Change-Id: Ief7386ab448c2711d0d8a94a77fa9ba189c8b7d2 Signed-off-by: jgilaber <jgilaber@redhat.com>
14 lines
537 B
YAML
14 lines
537 B
YAML
---
|
|
fixes:
|
|
- |
|
|
Previously, when an audit was created with zone_migration strategy and
|
|
both storage_pools and compute_nodes parameters are passed, the audit
|
|
did not created the required instances migration actions if any volume
|
|
migration action was created.
|
|
|
|
Now, in that situation the audit will create both instance and volume
|
|
migrations according to the expected behavior and the limits defined
|
|
by the parallelization parameters.
|
|
|
|
For more information: https://bugs.launchpad.net/watcher/+bug/2109722
|