Rongsong Shen
570e464dda
Task 6.3's own "before deciding" gate (re-run the flaky primary-push E2E 20-25x standalone with real stderr capture) was already satisfied by bug-1485's 2026-08-01 follow-up: 40/40 clean runs, zero captured crash evidence, after replacing the DEVNULL stderr redirect with real per-node log capture. Decision reached there (mergeable as-is, bug-1485 downgraded to low-priority/environment-flake) was never reflected back into tasks.md/spec.json. Marks 6.3 and its parent task-group 6 [x], and reconciles spec.json's stale "tasks-generated" phase (21/21 tasks were already complete) to "implementation-complete" with a validate_impl block recording the known limitations (bug-1485 stays open/low-priority, single-node ackedCount==0 exit-1 behavior, no CI live-E2E run yet). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
44 lines
3.6 KiB
JSON
44 lines
3.6 KiB
JSON
{
|
|
"feature_name": "admin-config-tool",
|
|
"created_at": "2026-07-06T01:30:00Z",
|
|
"updated_at": "2026-08-16T00:00:00Z",
|
|
"language": "en",
|
|
"phase": "implementation-complete",
|
|
"approvals": {
|
|
"requirements": {
|
|
"generated": true,
|
|
"approved": true
|
|
},
|
|
"design": {
|
|
"generated": true,
|
|
"approved": true
|
|
},
|
|
"tasks": {
|
|
"generated": true,
|
|
"approved": true
|
|
}
|
|
},
|
|
"ready_for_implementation": true,
|
|
"blocked_on": [
|
|
"RESOLVED 2026-07-25 (bug-1468, commit 959b0cda): controller OP_upgradeConfig dispatch defect (cluster.c) -- myreq_handler's dispatch guard was inverted (introduced by commit fd93997f, 2026-05-31), rejecting the only role that could ever legitimately reach it and making handle_upgrade_config() permanently unreachable from any network client. Fixed by config-upgrade-snapshot task 8.6 (which independently hit the same wall building a live OP_upgradeConfig sender). The controller-side path this blocker described has since been exercised live and successfully many times via config-upgrade-snapshot's --config-upgrade soak. push_config() itself (xceadmin's TLS/redirect/framing code) remains mock-tested only -- a live `xceadmin config push` run against a deployed cluster would give first-hand confirmation of the CLI's own wire code, but the blocking condition (dispatch unreachable) no longer exists.",
|
|
"RESOLVED 2026-07-22 (bug-1461, commit e53c086): controller.capnp's struct Data shadowed the capnp builtin blob type, miscompiling UpgradeConfigRequest.sha256/.bytes — fixed at the schema root (renamed to XceData), unblocking push_config()'s request construction. The dispatch defect above (bug-1468) is a separate issue and is now ALSO resolved (see entry above) — push_config() remains mock-tested only, no live xceadmin-specific E2E run yet, but nothing blocks one."
|
|
],
|
|
"gitea": {
|
|
"milestone_id": 1,
|
|
"issues": {
|
|
"requirements_design": 22,
|
|
"implementation": 23,
|
|
"validation": 24
|
|
}
|
|
},
|
|
"validate_impl": {
|
|
"decision": "GO",
|
|
"validated_at": "2026-08-16T00:00:00Z",
|
|
"note": "Corrected stale phase metadata (was 'tasks-generated' though 20/21 tasks were already [x] complete). Task 6.3's remaining blocker -- bug-1485's ~24/25 flaky primary-push failure -- was already followed up (2026-08-01): DEVNULL stderr redirect replaced with real per-node log capture, test re-run 40x standalone with zero reproductions and zero captured crash evidence. Decision reached: mergeable as-is, bug-1485 downgraded to low-priority/environment-flake rather than a blocker. Task 6.3 now marked [x] complete.",
|
|
"known_limitations": [
|
|
"bug-1485: primary-push E2E test (test_e2e_live_ha_push.py) had a ~24/25 flaky failure rate in one investigation round (elected primary observed <defunct> shortly after promotion) that could not be reproduced in 40 clean standalone runs; left OPEN as a low-priority environment-flake, not closed, since absence of reproduction here is not proof of absence. Real stderr capture is now permanently in place to diagnose any recurrence.",
|
|
"6.2's own finding stands: a genuinely single-node config push always exits 1 (ackedCount == 0, no standbys to ACK) even on full server-side success -- spec-consistent per cli.py/design.md's exit-0-iff-acked_count>0 postcondition, not a regression, but a future task may want a distinct single-node success signal.",
|
|
"push_config()'s TLS/redirect/framing code has had live E2E coverage only against the single-node (6.2) and 3-node (6.3) topologies exercised in this spec's own tests -- no production/CI live run yet."
|
|
]
|
|
}
|
|
}
|