fix(stack): refresh x402 runtime image pins#557
Open
bussyjd wants to merge 1 commit into
Open
Conversation
f5d94fc to
7af6ca6
Compare
OisinKyne
approved these changes
May 27, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes the rc6 runtime-image mismatch behind two reported upgrade findings:
/skill.mdcould fail under the new restricted Pod Security admission because the releasedserviceoffer-controllerimage was still the rc0-erab13254ebuild, which did not render restricted security contexts for controller-owned httpd workloads.This PR pins the controller and buyer sidecar to branch-built multi-arch images and adds tests that prevent these refs from drifting back to stale runtime builds.
Root Cause
The rc6 source at
98fa406contains the intended controller-render fixes, but the embedded production manifest still deployed a stale controller image:flowchart TD A[v0.10.0-rc6 source tag<br/>98fa406] --> B[Restricted PSS render code present] A --> C[Agent pod fsGroupChangePolicy present] D[embedded x402.yaml] --> E[serviceoffer-controller:b13254e] E --> F[Controller runs old renderer] F --> G[obol-skill-md missing restricted fields] F --> H[per-agent Hermes misses later pod spec fixes]The corrected deployment path is:
flowchart TD A[Fix branch source<br/>f5d94fc] --> B[Published multi-arch images<br/>f5d94fc@sha256] B --> C[embedded templates] C --> D[serviceoffer-controller renders restricted httpd workloads] C --> E[serviceoffer-controller renders per-agent fsGroup fields] D --> F[PodSecurity restricted admission succeeds] E --> G[per-agent Hermes PVC ownership path matches fixed source]Changes
serviceoffer-controllertof5d94fc@sha256:c6aa..., the branch-built image containing the controller-side hardening.x402-buyertof5d94fc@sha256:0c431...so the buyer sidecar stays on the same x402 branch build./data/.hermes/logsto both default Hermes and controller-rendered per-agent Hermes init paths, with test assertions. This is a narrow hardening for the exact path reported in the crash loop.Validation
Image publication:
Live sanity check against the existing smoke cluster, without teardown:
sequenceDiagram participant CLI as obol agent new --create-wallet participant Controller as serviceoffer-controller latest local image participant K8s as k3d cluster participant Hermes as per-agent Hermes CLI->>K8s: Apply Agent rc6-pvc-probe Controller->>K8s: Render namespace, PVC, Deployment, remote-signer K8s->>Hermes: Start pod with fsGroupChangePolicy Hermes-->>K8s: Running 1/1Observed: per-agent Hermes and remote-signer reached
Runningwith no manual PVC chown on the live smoke cluster.Remaining smoke gate
Full flow smoke is running from a combined branch with forced local dev images, but it is currently blocked on the local sudo prompt. The LLM-gated dual-stack flows are also blocked because
silvermesh.v1337.lan:8081is reachable on the LAN but refusing TCP connections.