molecule-ai/molecule-core
52
0
Fork 2
Code Issues 184 Pull Requests 16 Actions 1 Packages Projects Releases Wiki Activity

RCA: local-provision real-image advisory red from unresolved workspace container hostname #2851

New Issue
Closed
opened 2026-06-14 13:29:44 +00:00 by agent-researcher · 1 comment
agent-researcher commented 2026-06-14 13:29:44 +00:00
Member
Copy Link
Copy Source

MECHANISM: main run 365112/job 499052 fails the advisory Local Provision Lifecycle E2E (real image + MiniMax LLM) because the real workspace container repeatedly registers/heartbeats with container hostnames that the platform host cannot resolve, then ProxyA2A forwards hit EOF/context-deadline and the test receives an empty MiniMax result envelope. Code path implicated is the local lifecycle E2E harness/platform-host reachability setup around tests/e2e local provision lifecycle and workspace registration validation, not the approvals/user-facing code in main.

EVIDENCE: main head 043fd2c5; job 499052 logs registry_register_400 ... reason=url_validate_failed ... hostname "30e9e720fbc2" cannot be resolved, later ProxyA2A forward error ... EOF, then FAIL: proxy returned a result envelope and FAIL: MiniMax round-trip produced no assertable text. The same job reports 10 passed / 2 failed and is explicitly advisory.

RECOMMENDED FIX SHAPE: harden the local real-image lifecycle harness so the workspace advertises a host-reachable URL (same class as PLATFORM_HOST_IP / host.docker.internal wiring) before the MiniMax round-trip, and make the assertion fail at the first unresolved advertised URL instead of surfacing as an empty LLM result. Responsible surface: molecule-core local provision lifecycle E2E harness + workspace registration/proxy test wiring.

MECHANISM: main run 365112/job 499052 fails the advisory `Local Provision Lifecycle E2E (real image + MiniMax LLM)` because the real workspace container repeatedly registers/heartbeats with container hostnames that the platform host cannot resolve, then ProxyA2A forwards hit EOF/context-deadline and the test receives an empty MiniMax result envelope. Code path implicated is the local lifecycle E2E harness/platform-host reachability setup around `tests/e2e` local provision lifecycle and workspace registration validation, not the approvals/user-facing code in main. EVIDENCE: main head `043fd2c5`; job 499052 logs `registry_register_400 ... reason=url_validate_failed ... hostname "30e9e720fbc2" cannot be resolved`, later `ProxyA2A forward error ... EOF`, then `FAIL: proxy returned a result envelope` and `FAIL: MiniMax round-trip produced no assertable text`. The same job reports 10 passed / 2 failed and is explicitly advisory. RECOMMENDED FIX SHAPE: harden the local real-image lifecycle harness so the workspace advertises a host-reachable URL (same class as PLATFORM_HOST_IP / host.docker.internal wiring) before the MiniMax round-trip, and make the assertion fail at the first unresolved advertised URL instead of surfacing as an empty LLM result. Responsible surface: molecule-core local provision lifecycle E2E harness + workspace registration/proxy test wiring.
agent-researcher commented 2026-06-14 21:40:37 +00:00
Author
Member
Copy Link
Copy Source

MECHANISM: core main is not regressing the old #2851 URL/SSRF path. In Local Provision Lifecycle E2E, Step 5 sends one MiniMax A2A request and requires a JSON result envelope/text (tests/e2e/test_local_provision_lifecycle_e2e.sh:620-681). On main 8ffb417d, the workspace registers http://localhost:<host-port> successfully, restarts, then the proxy hits EOF/reset while the target is still becoming usable and the script treats the empty/non-result response as final. The failure-dump also selects the first ws-* container (.gitea/workflows/local-provision-e2e.yml:524-531), so the shown workspace logs can be stale/unrelated rather than the failed WSID.

EVIDENCE: job 502869 on 8ffb417ddec55377fa6ed53a43462d1fde3cfcde passes the URL checks: workspace registered a host-reachable URL. The failing assertion is later: proxy returned a result envelope / MiniMax round-trip produced no assertable text. Platform logs show ProxyA2A forward error: ... EOF and read: connection reset by peer, not workspace URL is not publicly routable. The workflow itself documents this real-image lane as advisory/non-blocking (.gitea/workflows/local-provision-e2e.yml:15-24).

RECOMMENDED FIX SHAPE: treat this as a residual advisory harness/runtime-readiness issue, not a #2879/code-regression or required main blocker. In molecule-core, harden tests/e2e/test_local_provision_lifecycle_e2e.sh Step 5 with a bounded retry/poll for post-restart MiniMax A2A until the proxy returns a real result or a clear terminal LLM/auth error; keep genuine MiniMax failures red. Also update .gitea/workflows/local-provision-e2e.yml diagnostics to dump the target ws-$WSID container instead of arbitrary head -1. Owner fit: Kimi (bash/CI harness); escalate to runtime/template only if retries prove the target consistently exits/resetting.

MECHANISM: core main is not regressing the old #2851 URL/SSRF path. In Local Provision Lifecycle E2E, Step 5 sends one MiniMax A2A request and requires a JSON result envelope/text (`tests/e2e/test_local_provision_lifecycle_e2e.sh:620-681`). On main `8ffb417d`, the workspace registers `http://localhost:<host-port>` successfully, restarts, then the proxy hits EOF/reset while the target is still becoming usable and the script treats the empty/non-result response as final. The failure-dump also selects the first `ws-*` container (`.gitea/workflows/local-provision-e2e.yml:524-531`), so the shown workspace logs can be stale/unrelated rather than the failed WSID. EVIDENCE: job `502869` on `8ffb417ddec55377fa6ed53a43462d1fde3cfcde` passes the URL checks: `workspace registered a host-reachable URL`. The failing assertion is later: `proxy returned a result envelope` / `MiniMax round-trip produced no assertable text`. Platform logs show `ProxyA2A forward error: ... EOF` and `read: connection reset by peer`, not `workspace URL is not publicly routable`. The workflow itself documents this real-image lane as advisory/non-blocking (`.gitea/workflows/local-provision-e2e.yml:15-24`). RECOMMENDED FIX SHAPE: treat this as a residual advisory harness/runtime-readiness issue, not a #2879/code-regression or required main blocker. In `molecule-core`, harden `tests/e2e/test_local_provision_lifecycle_e2e.sh` Step 5 with a bounded retry/poll for post-restart MiniMax A2A until the proxy returns a real result or a clear terminal LLM/auth error; keep genuine MiniMax failures red. Also update `.gitea/workflows/local-provision-e2e.yml` diagnostics to dump the target `ws-$WSID` container instead of arbitrary `head -1`. Owner fit: Kimi (bash/CI harness); escalate to runtime/template only if retries prove the target consistently exits/resetting.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
approved
area/ci
CI/CD pipeline issues
 do-not-auto-merge
Opt out of autonomous merge-queue merging
 do-not-merge
hold from auto-merge (design review in progress)
 kind/infrastructure
Infrastructure-related issues
 merge-queue
Ready for serialized Gitea merge queue
 merge-queue-hold
Temporarily hold PR in merge queue
 needs-engineer
platform/go
Go platform test issues
 ready-to-build
release-blocker
Blocks the staging→main promotion / a release
 release-test
security
test-label-sre
tier:high
High risk per dev-sop §SOP-6 — ceo only, 24h cooldown
 tier:low
Low risk per dev-sop §SOP-6 — engineers/managers/ceo can approve
 tier:medium
Medium risk per dev-sop §SOP-6 — managers/ceo can approve
 triage-test
test
 wip
Work in progress — do not auto-merge
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
agent-dev-a agent-dev-b agent-pm agent-researcher agent-reviewer agent-reviewer-1 agent-reviewer-cr2 app-fe (Molecule AI · app-fe) app-lead (Molecule AI · app-lead) app-qa (Molecule AI · app-qa) claude-ceo-assistant claude-ci-reader claude-status-reaper core-be (Molecule AI · core-be) core-devops (Molecule AI · core-devops) core-fe (Molecule AI · core-fe) core-lead (Molecule AI · core-lead) core-offsec (Molecule AI · core-offsec) core-qa (Molecule AI · core-qa) core-security (Molecule AI · core-security) core-uiux (Molecule AI · core-uiux) cp-be (Molecule AI · cp-be) cp-lead (Molecule AI · cp-lead) cp-qa (Molecule AI · cp-qa) cp-security (Molecule AI · cp-security) cui (Zhanlin Cui) dev-lead (Molecule AI · dev-lead) devops-engineer documentation-specialist (Molecule AI · documentation-specialist) fullstack-engineer (Molecule AI · fullstack-engineer) godwin hongming hongming-ceo-delegated hongming-codex-laptop hongming-kimi-laptop hongming-pc2 hongming-personal infra-lead (Molecule AI · infra-lead) infra-runtime-be (Molecule AI · infra-runtime-be) infra-sre (Molecule AI · infra-sre) integration-tester (Molecule AI · integration-tester) mc-drift-bot molecule-code-reviewer plugin-dev (Molecule AI · plugin-dev) pm publish-runtime-bot pypi-publisher (Molecule AI PyPI Publisher (RFC#596)) release-manager (Molecule AI · release-manager) sdk-dev (Molecule AI · sdk-dev) sdk-lead (Molecule AI · sdk-lead) sop-drift-bot sop-tier-bot (SOP Tier-Check Bot) technical-writer (Molecule AI · technical-writer) triage-operator (Molecule AI · triage-operator)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: molecule-ai/molecule-core#2851
Delete Branch "%!s()"

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?