Deployed Ceph fails when using custom network name
During deployment of Ceph the following error I seen:
2021-11-08 15:15:30.838746 | 244201c6-c8f6-fff8-c2cd-00000000001c | TASK | Ensure storage networks are in the inventory
2021-11-08 15:15:30.906178 | 244201c6-c8f6-fff8-c2cd-00000000001c | FATAL | Ensure storage networks are in the inventory | oc0-ceph-0 | error={
"changed": false,
"msg": "Both Storage and StorageMgmt should be in ['StorageCloud0', 'StorageMgmtCloud0', 'ctlplane'] but are not. If desired, override the network name variables, e.g. -e storage_network_name=ctlplane -e storage_mgmt_network_name=ctlplane\n"
}
In roles_data.yaml my ceph nodes had the following entry:
networks:
Storage:
subnet: storage_cloud_0_subnet
StorageMgmt:
subnet: storage_mgmt_cloud_0_subnet
Which differs from the default:
networks:
Storage:
subnet: storage_subnet
StorageMgmt:
subnet: storage_mgmt_subnet
In the config-download directory my tripleo-ansible-inventory has the following:
vars:
networks_all:
- StorageCloud0
- StorageMgmtCloud0
networks_lower:
StorageCloud0: storage_cloud_0
StorageMgmtCloud0: storage_mgmt_cloud_0
ctlplane: ctlplane
role_networks:
- StorageCloud0
- ctlplane
- StorageMgmtCloud0
Later in the inventory each host has its own variable map:
CephAll:
hosts:
...
oc0-ceph-2:
...
storage_cloud_0_hostname: oc0-ceph-2.storagecloud0.mydomain.tld
storage_cloud_0_ip: 172.16.11.96
storage_mgmt_cloud_0_hostname: oc0-ceph-2.storagemgmtcloud0.mydomain.tld
storage_mgmt_cloud_0_ip: 172.16.12.134
As it's valid for a user to customize these parameters and the cusotmized
roles_data.yaml is passed when the command is run with --roles-data. The
playbook should be updated handle this and use the customized network name.
https://github.com/openstack/tripleo-ansible/blob/master/tripleo_ansible/playbooks/cli-deployed-ceph.yaml#L117-L141
Related fix proposed to branch: master
Review: https:/