Skip to content

chart 0.4.2: ATOM discovery fixes from the live cluster, site extension points - #5

Merged
gspark-etri merged 4 commits into
masterfrom
chart-extension-points
Oct 19, 2026
Merged

gspark-etri merged 4 commits into
masterfrom
chart-extension-points

Conversation

@gspark-etri

Copy link
Copy Markdown
Collaborator

Found while running the daemon on the live cluster for the ai-rm end-to-end run.

  • Writable /tmp: rbln-stat is a PyInstaller binary; under readOnlyRootFilesystem it failed ("Could not create temporary directory") and no ATOM was published.
  • hostDev (default on): mounts the host's /dev live. A privileged container sees /dev as it was at start; after a pod was allocated and released an ATOM, the driver re-created the device and it dropped out of discovery (profile removed after 3 rounds). Verified: allocate/release an ATOM → all 4 ATOM stay published.
  • Site extension points (empty by default): command, extraEnv, initContainers, extraVolumes, extraVolumeMounts — the cluster runs the daemon from python:3.11-slim with an init container installing rm-abstract v0.4.1 until the image is published.

Live result: etri-001 publishes 3 RNGD + 4 ATOM, etri-esc8000-e11 2 RNGD; node labels drive rngd-pd placement. Defaults render unchanged apart from the two volumes.

🤖 Generated with Claude Code

gspark-etri and others added 4 commits October 19, 2026 14:04
…aVolumes, extraVolumeMounts)

Lets a site run the daemon from a stock Python image until the rm-abstract-node
image is published. All empty by default; chart 0.4.2.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…TMPDIR)

Without it rbln-stat fails with 'Could not create temporary directory' and the
node publishes no ATOM profiles.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A privileged container sees /dev as it was at start. On etri-001 an ATOM that the
device plugin allocated and released was re-created by the driver, dropped out of
rbln-stat inside the daemon, and its profile was removed after 3 rounds.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@gspark-etri
gspark-etri merged commit c72f356 into master Oct 19, 2026
8 checks passed
@gspark-etri
gspark-etri deleted the chart-extension-points branch October 19, 2026 05:44
gspark-etri added a commit that referenced this pull request Oct 20, 2026
chart 0.4.2: ATOM discovery fixes from the live cluster, site extension points
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.

1 participant