Windows VM hangs after live-migration
Bug #1657722 reported by
Oleksandr Liemieshko
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Invalid
|
High
|
MOS Nova | ||
6.0.x |
Won't Fix
|
High
|
MOS Nova | ||
6.1.x |
Won't Fix
|
High
|
MOS Nova | ||
7.0.x |
Won't Fix
|
High
|
MOS Nova | ||
8.0.x |
Won't Fix
|
High
|
MOS Nova |
Bug Description
After a live-migration of an instance(Windows) from one compute node to another it becomes unresponsive - ping status works however RDP does not work and VNC console in horizon just shows a black screen. The problem does not reproduce continuously.
The only way to get instance up and running is reboot
Steps to reproduce:
- 1 controller or HA, 2 computes
- VLAN
- Ceph for all
- image for Windows
Scenario:
1. Create new VW from Windows image
2. Try to migrate "Live Migrate Instance". It may require a large number of attempts
Changed in mos: | |
importance: | Undecided → High |
tags: | added: customer-found |
tags: | added: support |
description: | updated |
Changed in mos: | |
assignee: | nobody → MOS Nova (mos-nova) |
Changed in mos: | |
milestone: | 9.x-updates → 9.2-mu-1 |
To post a comment you must log in.
Alexander,
If possible please add the Windows version you run inside the image.
Additionally, please check the CPU model on the compute node hosts you use to run and migrate Windows instance.
And finally, please check power saving settings in the compute host BIOS.
The root cause might be Windows does not recognize all modern powersaving and frequency management CPU features due to lack of ACPI/ATK or other driver.