[work-c3n] ee: drop the receptor binary the image never used #27
Loading…
Reference in a new issue
No description provided.
Delete branch "work-c3n-receptor-binary-is-unpinned-devel-unused"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
The
COPY --from=quay.io/ansible/receptor:develline was the only thing tying this build to a tag that is republished daily — run 3404 died onsha256:f186e516…because that digest 404s now, and PR #22 re-pinned it, which buys until the next devel push.It is also unused. AWX container-group jobs run the pod with
args = ['ansible-runner', 'worker', '--private-data-dir=/runner'](awx/main/utils/execution_environments.py), so receptor is never the entry process. The layer is 73.4 MB and the binary in it was 1.6.8+git58f9101 against a controller shipping 1.4.8+d7fe592 — a version skew that only matters in a topology this image is not used in.Removed: the
COPY, the/var/run/receptordirectory that existed for the daemon, and the two README claims about it.receptorctlstays inrequirements.txt: it is a separate CLI, and it talks to a receptor node over the network rather than to a local binary.Verification. No container daemon on fractal, so I cannot build here. Grep of the whole build context finds no remaining reference: only
receptorctlinrequirements.txtand the README sentence explaining the removal. The smoke test inbuild.ymlbuilds the image and runstests/playbook.ymlthroughansible-runnerin it, so the main build answers whether the image still works without the layer. I will watch that run and revert if it fails.Small, contained, reversible: no data, no credentials, no access path. If it turns out some job template does call
receptoras an execution node in this EE, the revert is one line.