Rally install fails on existing deployment creation
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack-Ansible |
Fix Released
|
Undecided
|
shubham |
Bug Description
After deploying my new OpenStack Cloud with OpenStack-Ansible tag 17.0.6, I wanted to know if everything worked.
So I thought about Rally and try to install it.
Reading the doc here https:/
I saw nothing special to configure.
When I launch the playbook os-rally-
TASK [os_rally : Setup the Rally deployment] *******
Wednesday 18 July 2018 16:26:39 +0200 (0:00:01.626) 0:00:34.712 ********
fatal: [fidcllabs-
bin/activate\n /openstack/
-07-18 14:26:41.096022", "failed": true, "msg": "non-zero return code", "rc": 231, "start": "2018-07-18 14:26:39.922005", "stderr": "", "std
err_lines": [], "stdout": "2018-07-18 14:26:40.977 15491 WARNING rally.common.
ckage (see https:/
nEnv manager got invalid spec: \n[\"There is no Platform plugin with name: 'existing@
77 15491 WARNING rally.common.
ack). In-tree OpenStack plugins will be removed from the Rally main package soon.\u001b[00m", "Env manager got invalid spec: ", "[\"There is
no Platform plugin with name: 'existing@
No other errors before.
Did I miss something ?
Changed in openstack-ansible: | |
status: | New → Confirmed |
Ah, I can add some specificities of my deployment :
- Multi domain, with one backed by Active Directory
- Different internal and public endpoints, managed by haproxy
- openrc points to the Default domain admin account, with internal endpoints
- flat Netwok for external shared network
- VXLAN for private networks with Linux Bridges
- Ceph for glance/nova/cinder with erasure coding pools