Skip to content

feat: test that a network reaches the cell a workload runs in - #228

Merged
scotwells merged 2 commits into
mainfrom
feat/e2e-nso-crds-on-hub
Aug 14, 2026
Merged

feat: test that a network reaches the cell a workload runs in#228
scotwells merged 2 commits into
mainfrom
feat/e2e-nso-crds-on-hub

Conversation

@scotwells

Copy link
Copy Markdown
Contributor

A workload gets an address from the network layer at the location it runs in. Nothing in our tests proves that actually happens end to end, because the test environment cannot carry a network's presence from the federation layer down to a cell.

This makes that path exist locally, so the address workloads depend on can be tested rather than assumed.

What changes

Only the test environment and the permission compute needs to ask for a network in a location. No product behaviour, and nothing switched on.

Related

scotwells and others added 2 commits August 14, 2026 08:26
The harness installed NSO's CRDs on the control plane and both POP cells but
not on the Karmada hub, and installed no NSO propagation policy at all. A
NetworkContext is written on the hub and propagated to the cells serving its
location, so there was no local path by which one could reach a cell and no
way to test the claim path end to end.

Install the NSO CRDs on the hub too, apply NSO's own federation propagation
policy there, and label the member clusters infra.datum.net/gateways=enabled
so that policy actually selects them — it matches nothing without the label,
and does so silently.

Also grant the hub identity networkbindings on networking.datumapis.com.
Compute asks for a network in a location by writing a binding on the hub, and
the hub role granted nothing in that group at all.
…ronment

Running the local e2e environment required Python plus the PyYAML module, an
interpreter dependency unrelated to anything the environment actually tests.
Every use is expressible with tools the Taskfile already depends on, so a
contributor now needs only go, kubectl, docker, and a POSIX shell.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@scotwells
scotwells marked this pull request as ready for review August 14, 2026 14:38
@scotwells
scotwells requested review from ecv and privateip August 14, 2026 14:38
@scotwells
scotwells enabled auto-merge August 14, 2026 14:38
@scotwells
scotwells merged commit 0283a7c into main Aug 14, 2026
10 checks passed
@scotwells
scotwells deleted the feat/e2e-nso-crds-on-hub branch August 14, 2026 14:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants