[fuel-octane] Octane doesn't handle upgrade_levels in nova.conf
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Committed
|
Medium
|
Oleg S. Gelbukh | ||
8.0.x |
Fix Committed
|
Medium
|
Unassigned | ||
Newton |
Fix Committed
|
Medium
|
Oleg S. Gelbukh |
Bug Description
Detailed bug description:
Octane doesn't handle upgrade_levels in nova.conf when source release version is less then 6.1
Steps to reproduce:
Upgrade existing installation until the step for workload should be live-migrated to another hypervisor
nova host-evacuate-live node-2
Expected results:
A virtual machine, which undergo live migration, is migrated to the other hypervisor.
Actual result:
A virtual machine, which undergo live migration, ends up in migrating state.
In /var/log/
UnsupportedVer
Reproducibility:
Always
Workaround:
(on an upgraded controller)
echo -e "[upgrade_
for i in $(initctl list|grep nova|grep run|awk '{print $1}'); do restart $i; done
Impact:
Live migration doesn't work
Description of the environment:
Source release: 6.0
Destination release: 8.0
summary: |
- [fuel-octane] Octane doesn't handle upgrade_level in nova.conf + [fuel-octane] Octane doesn't handle upgrade_levels in nova.conf |
tags: | added: area-python |
Changed in fuel: | |
milestone: | none → 10.0 |
assignee: | nobody → Fuel Octane (fuel-octane-team) |
importance: | Undecided → Medium |
status: | New → Confirmed |
I've found out that it doesn't work for 6.1 also (not only for 6.0).
Function below assumes that [upgrade_levels] already in place, which is not true. upgrade_ levels( node, version): file(sftp, '/etc/nova/ nova.conf' ) as (old, new):
new. write(line) ("[upgrade_ levels] "):
new.write( "compute= {0}\n". format( version) )
=============
def add_compute_
sftp = ssh.sftp(node)
with ssh.update_
for line in old:
if line.startswith
=============