Wrong OS_AUTH_URL makes keystone operations fail
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Invalid
|
High
|
Alexey Deryugin | ||
8.0.x |
Invalid
|
High
|
Alexey Deryugin | ||
9.x |
Invalid
|
High
|
Alexey Deryugin |
Bug Description
Every keystone command fails with error: Authorization Failed: The resource could not be found. (HTTP 404)
Example:
root@node-1:~# keystone tenant-create --name demo
/usr/lib/
'python-
/usr/lib/
'the 2.0.0 release.', DeprecationWarning)
/usr/lib/
super(Client, self)._
/usr/lib/
return f(*args, **kwargs)
/usr/lib/
'the 2.0.0 release.', DeprecationWarning)
Authorization Failed: The resource could not be found. (HTTP 404)
Steps to reproduce:
1. Deploy MOS cluster.
2. SSH to Controller node.
3. source openrc.
4. type "keystone tenant-create --name demo", or "keystone user-list" etc
Contents of openrc:
root@node-1:~# cat openrc
#!/bin/sh
export LC_ALL=C
export OS_NO_CACHE='true'
export OS_TENANT_
export OS_PROJECT_
export OS_USERNAME='admin'
export OS_PASSWORD='admin'
export OS_AUTH_URL='http://
export OS_DEFAULT_
export OS_AUTH_
export OS_REGION_
export CINDER_
export GLANCE_
export KEYSTONE_
export NOVA_ENDPOINT_
export NEUTRON_
export OS_ENDPOINT_
export MURANO_REPO_URL='http://
See that export OS_AUTH_URL='http://
here suffix v2.0 is lost.
To fix this it should be:
export OS_AUTH_URL='http://
Cluster configuration:
Controller+CephOSD
Controller+CephOSD
Compute
Mongo
Ironic
Neutron VLAN network, Ceph for all storages.
VERSION:
feature_groups:
- mirantis
production: "docker"
release: "8.0"
api: "1.0"
build_number: "529"
build_id: "529"
fuel-nailgun_sha: "baec8643ca624e
python-
fuel-agent_sha: "658be72c4b42d3
fuel-
astute_sha: "b81577a5b7857c
fuel-library_sha: "e2d79330d5d708
fuel-ostf_sha: "3bc76a63a9e7d1
fuel-mirror_sha: "fb45b80d7bee58
fuelmenu_sha: "e071216cb214e3
shotgun_sha: "63645dea384a37
network-
fuel-upgrade_sha: "616a7490ec7199
fuelmain_sha: "a365f05b903368
description: | updated |
Changed in mos: | |
status: | New → Confirmed |
tags: |
added: area-keystone removed: keystone |
tags: | added: release-notes |
Changed in mos: | |
status: | Confirmed → Won't Fix |
tags: |
added: 8.0 release-notes-done removed: release-notes |
tags: | added: keep-in-9.0 |
Hi team,
we are going to investigate why this issue was not detected by our automated tests, it was fund during manual acceptance of RC1.
It looks like it is regression issue which will affect Keystone CLI (other services work fine, WebUI works fine as well).
And we have the workaround for the issue (manually fix openrc file)
Workaround:
Execute on all controller nodes:
sed -i 's/5000\ //5000\ /v2.0\/ /' /root/openrc
source /root/openrc