Directional network performance issue
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Won't Fix
|
Medium
|
Andrey Epifanov | ||
6.0.x |
Won't Fix
|
Medium
|
Andrey Epifanov | ||
6.1.x |
Won't Fix
|
Medium
|
Andrey Epifanov | ||
7.0.x |
Won't Fix
|
Medium
|
Andrey Epifanov |
Bug Description
I was able to reproduce https:/
Version:
{"build_id": "2014-09-
We saw huge network performance degradation, almost the same as described in http://
Lab schema included.
Env configuration:
1x controller - Multi-node, no HA, GRE.
1x compute - KVM
I was able to fix that with the commands
# ethtool --offload eth0 gro off
# ethtool --offload eth1 gro off
executed on the controller node only, as proposed in https:/
Changed in mos: | |
assignee: | nobody → MOS Neutron (mos-neutron) |
tags: | added: neutron |
Changed in mos: | |
assignee: | MOS Neutron (mos-neutron) → Andrey Epifanov (aepifanov) |
Changed in mos: | |
milestone: | 6.0 → 6.0.1 |
Changed in mos: | |
status: | Confirmed → Won't Fix |
It has a workaround in Icehouse. See item 5 in the following doc docs.openstack. org/icehouse/ install- guide/install/ apt/content/ neutron- network- node.html
http://
It describes the same way Dmitry fixed this issue by the following command "ethtool --offload eth0 gro off"
Not sure if need to add this command to manifests and trigger it during the deployment stage.