article/awx-hop-nodes/hopnodes/

[giscus[bot]]

article/awx-hop-nodes/hopnodes/

Malfunctioning Eddie's Ramblings

https://malfunceddie.github.io/article/awx-hop-nodes/hopnodes/

[MalfuncEddie]

I think

https://github.com/ansible/awx/blob/cb63d92bbf5d8e10834264f0eb8142c4eb5c9161/awx/main/scheduler/task_manager.py#L484-486

is throwing me under the bus
if settings.IS_K8S and task.capacity_type == 'execution':

looking further I see

sign_work = False if settings.IS_K8S else True
result = receptor_ctl.submit_work(worktype='ansible-runner', node=node, signwork=sign_work, **kwargs)

So I think I need work singing keys ;)

[MalfuncEddie]

After adding work keys (with some help from Seth Foster on the mailing list) I still get errors:

I think I am running into this:

ansible/awx@f850f8d

This means that when IS_K8S=True, any Job Template associated with an Instance Group will never actually go from pending -> running (because there's no capacity - all playbooks must run through Container Groups). For an improved ux, our intention is to introduce logic into the operator install process such that the default group that's created at install time is a Container Group that's configured to point at the K8S cluster where awx itself is deployed.

It looks like this global flag completely blocks the intention of running execution/hop/hybrid nodes on AWX since "only container groups are intended to be used".

Whem I edit the awx-awx-configmap and set the setting setting IS_K8S=False

I get the folowing result.

The the control node is now a hybrid node? But my job is trying to run on the execution node :O

I still got an error (no output form job) but:

WARNING 2022/05/10 08:31:36 Received unreachable message from axw-ee
WARNING 2022/05/10 08:31:36 Received unreachable message from axw-ee
ERROR 2022/05/10 08:31:36 Error locating unit: VECXTy1J
ERROR 2022/05/10 08:31:36 unknown work unit VECXTy1J

So there is progress and since now there is the intention to run on the execution node.

[guyomog78]

Hello,

With recent developpement on the awx version, did it solved the goal?

[MalfuncEddie]

I think that at the current state (or when I was testing) it was not efficient. If the is_k8s setting is set it expects podman to run stuff.

So I would have to rebuild images to include podman. So you would have a k8s cluster that runs awx that runs a container that needs podman to run something else. Bit of virtualization inception.

I did spin up a docker-compose dev build to see it that would work. But not enough time and issue after issue. ;(

[EDIT]
Good news
Apparently receptor/awx team is implementing this feature

ansible/receptor#615

[guyomog78]

Hello,

Well, now it is include with a bundle

[MalfuncEddie]

Hello,

Well, now it is include with a bundle

idd in the latest version it is included. also the nice receptor collecton is quite nice