Skip to content

Extract explicit EC storage selection and celld - #1537

Open
ChristianPavilonis wants to merge 7 commits into
baseline/sourcepoint-20261110from
baseline/ec-storage-20261110
Open

ChristianPavilonis wants to merge 7 commits into
baseline/sourcepoint-20261110from
baseline/ec-storage-20261110

Conversation

@ChristianPavilonis

@ChristianPavilonis ChristianPavilonis commented Nov 11, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

  • Original typed EC provider selection, read/write/expiry/generation bridge and celld integration.
  • Reorganize the existing working demo using EdgeZero83d7a629, not current-main forwarding compatibility.

Draft baseline review unit. Not a current-main merge candidate, deployment approval or completed current-head CI claim.

Baseline and sequence

Head: f14eefa672e83517dc57ff0c0d87383447e07807. Base: baseline/sourcepoint-20261110.

Historical main anchor: 2dd968aeebeb2cc0d5110115f64e0ad58e2e1d13. Original demo: TS2423a2ed plus separately preserved dirty inputs. Matching EdgeZero: draft431, exact revision 83d7a6296d285e135c4ea70021b19e6c8dca7e9f.

The user selected compatible-baseline cleanup. Current-main integration with EdgeZero462 and the authenticated-forwarder incompatibility are deferred. Preserve current-main PR1525/1526 ready and1527/1528 draft, with their existing bases. No excluded1499 or independent1506 implementation, relay/privacy experiment, new transport fallback, group/codec/phase design or operational deployment is included.

Changes

Area Change
This unit Original typed EC provider selection, read/write/expiry/generation bridge and celld integration.
Unchanged work Original checkouts, dirty Prebid inputs, recovery references and existing PRs remain preserved.

Closes

Refs #1530. This unit completes one part of the baseline extraction and does not close the overall tracker on its own.

Spec

Not a new feature. This is extraction of existing demo behavior and bounded build/caller corrections. Original native and EC storage design/review records remain in the baseline where applicable; they are not presented as approval of later combined source. No new forwarding, consent, timer or provider policy is designed here.

Test plan

Core3016 plus5 storage configuration cases and native all-feature suites pass. Three explicitly selected real local celld cases pass for lifecycle, prepared startup and native HTTP issuance/withdrawal/restart/outage. Local fixture proof is not replicated/fleet durability.

  • Required current-head hosted checks complete before ready-state transition.
  • Final baseline combined-check results recorded with exact input revisions and limitations.
  • Focused failures and corrections retained rather than assertions relaxed.

The b67ffc6 baseline checkpoint has passing documented local gates: Fastly/default/reuse/native/Cloudflare/Spin suites, all adapter strict lint lanes, 17 cross-adapter parity cases, 1252 JS cases, CLI suites, formatting, clock/digest/proposal checks and changed-lines domain lint. Passing checks at121ebcc36 are reused only for unchanged relevant inputs; the final follow-up covers the CLI test-import and documentation-only corrections. These results are not a substitute for this PR's current-head hosted checks.

Checklist

  • Existing behavior extracted into a scoped review unit.
  • Public examples use fictional inputs and no licensed Geo data.
  • No deployment, main merge, old-RC closure or recovery deletion.
  • No new worker launched or paused experiment resumed.
  • Review readiness confirmed independently from later units.

Operational Terraform remains in trusted-server-deployer. Local checks and prototype tests do not qualify cloud access, fleet durability or a live publisher.

Hosted docs-build correction

The first hosted format-docs job caught an additional site-build gate that the original local formatter checkpoint had not run: a relative link from the native EC guide to a Terraform example outside the site. Native EC correction8b11dbad3 replaces it with the plain repository path; no dead-link check is disabled. That docs-only correction is carried through the existing upper PR identities. Native EC and final integration8985b111 both pass docs ESLint and VitePress build; the final docs formatter also passes. Runtime, dependency, image-build-context and test inputs are unchanged from b67ffc6, so their prior exact-input proofs are reused. Hosted current-head checks still control readiness.

Startup negative-probe correction

A hosted native EC case failed the no-listener probe while its child logged the expected startup validation failure. The old log has no socket endpoints, so the failure's exact subtype is not claimed confirmed. A separate controlled Linux reproduction demonstrates a successful TCP self-connection with no listening server. Native correction8490c691d distinguishes equal local/peer endpoints from a real listener; actual socket controls prove the self-connection is excluded and a real listener is still detected. Nonzero exit, expected diagnostic, secret-redaction and the original startup deadline remain required. This changes only the acceptance fixture, not production admission, transport or trust.

The earliest native unit and final9356c557 default/all-feature acceptance checks and strict lint pass, including both controls and the original failure case. Ignores remain explicit; earlier separately executed celld proof is not relabeled as a new ignored-case run. Correction is merged upward through existing PR identities without force-pushing or retargeting. Current-head hosted checks must pass; the old failure is retained rather than rerun-only waived. Runtime/API/dependency inputs remain unchanged. A fresh exact9356 producer build and real runtime smoke now pass: image sha256:65a735d38e9dbc0573d0f6068a8b261bd202c43aea05085c2b2d6bccc91ebdbd, linux/amd64 UID/GID10001, including startup rejection and in-flight SIGTERM drain. Prior b67 image receipts remain separate; no image was pushed. All local parent runners are complete.

Browser CI host compatibility

Preserved older browser failures show Playwright1.58.2 rejecting ubuntu-latest's new Ubuntu26.04 image before browser installation; later missing-executable failures are consequences, not passing browser tests. Nativeb1216763a pins only the browser job to ubuntu-24.04, leaving dependency versions, tests and assertions unchanged. Direct Actionlint1.7.12 passes at the owning native and final integration points; the first shim-only attempt could not run because no global version was configured, and is not counted as lint proof. No global tooling was changed. The only9356 ->ee7 delta is this workflow host/comment; runtime, Cargo, docs, tests and Docker-context inputs are unchanged. Actual browser execution on the supported hosted runner remains pending and must pass before readiness. No waiver or browser skip.

@prk-Jr prk-Jr left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Summary

The typed [ec.storage] model, the Fastly legacy compatibility and the Cloudflare/Spin rejection paths look sound. Two things should be addressed before this unit is ready: the shared example config now fails at startup on every adapter except Fastly, and the new cargo test (axum native) job fails at this head.

Blocking

🔧 wrench

  • The example config fails at startup off Fastly. See the inline comment on trusted-server.example.toml.

  • CI: cargo test (axum native) fails at f14eefa (not a required check on this base). active_legacy_ec_environment_cannot_disable_or_override_published_storage panics at crates/trusted-server-adapter-axum/tests/startup.rs:165 with "should not listen while required startup inputs are invalid". The same test passed on the earlier cdad6df31 run, so this looks like a flake rather than a production defect. The likely cause is the Linux self-connection already fixed in native_acceptance.rs (connected_to_listener, merged in 83f58c5e3). ServerFixture::assert_failed_before_listen in startup.rs uses the same reserve-release-reconnect probe without that guard, and this PR adds three new callers. The CI log has no socket endpoints, so the cause is inferred rather than confirmed. Suggested change (outside the diff, so not a one-click suggestion; not compiled):

    if let Ok(stream) = TcpStream::connect_timeout(&self.addr, Duration::from_millis(20)) {
        // Linux can self-connect when an ephemeral source port equals
        // the target; only distinct endpoints prove a real listener.
        assert_eq!(
            stream.local_addr().expect("should read probe source"),
            stream.peer_addr().expect("should read probe peer"),
            "should not listen while required startup inputs are invalid"
        );
    }

Non-blocking

🤔 thinking

  • ts config push --adapter <x> already knows its target runtime, but publishing only runs validate_storage. Resolving against the target adapter at push time would surface the example-config failure before the server starts. This is separate from the "do not resolve according to the CLI host" rule.
  • Celld preparation errors: see the inline comment on celld_ec.rs.

📝 note

  • The [ec] field table in docs/guide/configuration.md (around line 645) still lists only ec_store and does not document storage.

CI Status

  • 16 checks pass.
  • FAIL: cargo test (axum native), which is not in the base branch's required set.
  • Locally, the core ec_storage_config tests pass (5/5). I also ran a throwaway probe that resolves the shipped example config for each runtime. The full adapter suites, the real-celld ignored tests and the image script were not re-run for this review.

Comment on lines +88 to +89
[ec.storage]
store = "ec_identity_store"

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🔧 wrench: This makes the shared starter config fail at startup on Axum, Cloudflare and Spin.

ts config init writes this file, and the Axum quickstart (docs/guide/getting-started.md, Option B) copies it unchanged. Previously, ec_store = "ec_identity_store" was inactive off Fastly, so one starter worked on every adapter. With the provider omitted, resolve_storage returns ec.storage.provider is required outside Fastly for Native, Cloudflare and Spin. Axum fails in build_state_with_services and initialize_native, while ts config push --adapter axum still succeeds because publication only checks shape.

I reproduced this by parsing this exact file (with resolved test secrets) and resolving it for each runtime: Fastly returns Ok(Some), and the other three return that error. The integration fixtures in this PR already remove ec.storage for Axum and Cloudflare to work around it (tests/common/config.rs). This conflicts with the spec's goal to preserve the behavior of deployments without configured persistence.

Options:

  • (a), preferred: keep ec_store = "ec_identity_store" in the example. It is still supported with no removal deadline, so one starter keeps working everywhere. The cost is a one-line spec change ("new examples use [ec.storage]").
  • (b): keep this section and add a "remove [ec.storage]" step to the Axum, Cloudflare and Spin setup docs. ts config init would then still produce a file that only works on Fastly.

records: CelldRecordStore,
}

fn preparation_error() -> EdgeError {

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🤔 thinking: Every preparation failure (bad or rejected token, unsupported protocol or record version, unreachable service) collapses into the same native EC celld preparation failed, and store_error does the same for request-time errors. Hiding input values is intended, but an operator can't tell a credential problem from an outage. Could you attach a value-free error category from the celld error kind (for example credential, protocol or transport)?

@ChristianPavilonis
ChristianPavilonis marked this pull request as ready for review November 11, 2026 13:50

This branch has not been deployed

No deployments
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