Rework inventory generation #2341
Conversation
|
Skipping CI for Draft Pull Request. |
First we move the inventory files generation to a far earlier stage.
While it doesn't change anything on its content, having the inventory
generated earlier helps debugging it without the need to create any
actual workload (hacky approach I used to validate the generated
inventories).
Then, the data used to generate the inventories is updated, in order to
ensure we're still puting the nodes under the right group.
This is especially important for the OCP cluster, when we have workers
in the picture.
Before this change, the workers were nested under the "ocps" group,
leaving the "ocp_workers" empty.
This change corrects the situation, leading the generated inventories to
look like this:
```yaml
ocps:
hosts:
ocp-master-0:
ansible_host: master-0.utility
ansible_user: core
ansible_ssh_common_args: '-o StrictHostKeyChecking=no'
ansible_ssh_private_key_file: ~/.ssh/devscripts_key
ocp_workers:
hosts:
ocp-worker-0:
ansible_host: worker-0.utility
ansible_user: core
ansible_ssh_common_args: '-o StrictHostKeyChecking=no'
ansible_ssh_private_key_file: ~/.ssh/devscripts_key
computes:
hosts:
compute-xio2mlri-0:
ansible_host: compute-xio2mlri-0.utility
ansible_user: zuul
ansible_ssh_common_args: '-o StrictHostKeyChecking=no'
ansible_ssh_private_key_file: ~/.ssh/id_cifw
controllers:
hosts:
controller-0:
ansible_host: controller-0.utility
ansible_user: zuul
ansible_ssh_common_args: '-o StrictHostKeyChecking=no'
ansible_ssh_private_key_file: ~/.ssh/id_cifw
```
cjeanner
force-pushed
the
libvirt_manager/ocp-inventory
branch
from
258b3b3
to
5f822b6
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/7510933306b348fea31e7d11e9d4baee ✔️ openstack-k8s-operators-content-provider SUCCESS in 1h 36m 25s |
|
recheck |
|
/approve |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: raukadah The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
First we move the inventory files generation to a far earlier stage.
While it doesn't change anything on its content, having the inventory
generated earlier helps debugging it without the need to create any
actual workload (hacky approach I used to validate the generated
inventories).
Then, the data used to generate the inventories is updated, in order to
ensure we're still puting the nodes under the right group.
This is especially important for the OCP cluster, when we have workers
in the picture.
Before this change, the workers were nested under the "ocps" group,
leaving the "ocp_workers" empty.
This change corrects the situation, leading the generated inventories to
look like this: