fix(db): repair fresh migration chain
CI and Deploy / test (pull_request) Successful in 5m19s
CI and Deploy / deploy (pull_request) Has been skipped

This commit is contained in:
cesnimda
2026-08-15 20:47:36 +02:00
parent 191de69c48
commit 74a1e0d845
15 changed files with 297 additions and 35 deletions
+2 -1
View File
@@ -1,6 +1,6 @@
# Technical debt
Last reconciled: 2026-07-31
Last reconciled: 2026-08-15
This ledger contains verified engineering debt only. Product ideas belong in the roadmaps and
operator/external dependencies belong in `BLOCKERS.md`.
@@ -40,6 +40,7 @@ operator/external dependencies belong in `BLOCKERS.md`.
| Priority | Debt | Current decision / trigger |
|---|---|---|
| P1 | EF migrations and startup reconciliation still share historical schema ownership. Blank SQLite EF migration, populated upgrade, and fresh/restarted MariaDB now pass, but the two mechanisms remain tightly coupled. | Keep the compatibility bootstraps and provider-specific ordering covered by `MigrationChainTests`. Consolidate ownership only through an expand/verify/contract migration after a production restore rehearsal; do not rewrite applied migration history. |
| P1 | `JobApplication` still duplicates opportunity data now owned by `Job`. | Startup now backfills missing `Job` rows and both create paths dual-write. Keep compatibility reads until the production report and restore rehearsal pass; observe one release, then remove the legacy columns. |
| P1 | Background workers assume one API instance. Restart recovery is durable, but there is no row lease for concurrent workers. | Add database leasing only before deploying more than one backend replica. |
| P2 | Production log aggregation is still deployment-owned; Compose now bounds each container's local logs to 3 × 10 MB. | Add an OTLP/Seq sink only before multi-host operation or when incident-response needs exceed `docker logs`. |
+1
View File
@@ -217,3 +217,4 @@ Output was reduced to filenames and commit counts. The token artifact appears un
| V-183 | Owner-filtered job-choice API test; correspondence Jest; frontend production build | Repository root / `job-tracker-ui` | Remove the email compose/thread-move selectors' false 100-job ceiling | PASS — backend search finds the oldest target among 130 owned rows and excludes another tenant; correspondence 20/20 proves debounced server search, compose selection and thread-move selection; TypeScript/production build passes | Synthetic rows/JSDOM only; no provider, email or production action | MAIL-001 exhaustive job selection gap closed |
| V-184 | Shared synchronous AI provider decorator, durable/workspace suppression scopes, quota exception handler and full backend | Repository root | Make numeric Free/Pro AI limits universal without double-counting already-reserved work | PASS — focused shared-provider/accounting suite 25/25 and backend 674/674; success finalizes measured characters, Free/exhausted requests stop before provider I/O, workspace/operation scopes create no second row, and quota failures return stable 429 details | Fake in-process provider and SQLite only; no model, Stripe, MariaDB or production call | POL-001 repository accounting gap closed; Stripe lifecycle and production smoke remain |
| V-185 | Stripe gateway seam, mocked checkout/webhook lifecycle, entitlement tests and full backend | Repository root | Prove checkout identity and downgrade safety without using external Stripe | PASS — entitlement/billing 33/33 and backend 677/677; configured `price_` and stable user metadata reach Checkout, active grants Pro, `past_due` revokes it, canceled replay remains revoked without duplicate role mutation, non-AI profile data survives, and `prod_` in the price setting fails closed | In-process fake only; no Stripe network, customer, secret mutation, MariaDB or production call | Local POL-001 Stripe lifecycle gap closed; configured Stripe account journey remains blocked |
| V-186 | Migration-chain regression tests, EF model parity/scripts, direct EF SQLite, real application startup and disposable MariaDB 11.8 fresh/restart | Repository root / disposable local databases | Close the historical blank-chain defect without changing applied production state or losing populated rows | PASS — migration tests 3/3 and backend 680/680; blank SQLite reaches all 29 migrations twice, an older populated checkpoint preserves job title/date/owner/summary, EF-only SQLite subsequently serves `/health`, and fresh/restarted MariaDB serves `/health` with 29 migrations, 49 tables and provider-correct sampled ID/owner/decimal/timestamp types. MariaDB script constrains identifiers to 64 characters | Synthetic disposable databases only; no production migration, downgrade, backup restore or private row. Migration/reconciler dual ownership remains JT-019 architectural debt | Blank-chain blocker closed; production restore/rollout remains gated |
+32 -9
View File
@@ -1,6 +1,6 @@
# Database ownership and startup order
> 2026-07-19. Which component creates which table, in what order, and why a clean MariaDB install used
> Updated 2026-08-15. Which component creates which table, in what order, and why a clean MariaDB install used
> to fail. Read this before adding a table or touching `StartupInitializationExtensions`.
## The problem this document exists to prevent
@@ -22,28 +22,35 @@ provider.
## Startup order
`InitializeJobTrackerAsync` runs exactly this sequence:
`InitializeJobTrackerAsync` runs this provider-aware sequence:
```
1. Connect
2. ReconcileSchema() ← pass 1: repair existing schema, create reconciler-owned tables
3. Database.Migrate() ← create every migration-owned table
4. ReconcileSchema() ← pass 2: everything pass 1 had to skip
2. ReconcileSchema() ← repair legacy schema/create prerequisites
3a. SQLite: apply one migration, reconcile, repeat
3b. MariaDB: apply the complete migration chain
4. ReconcileSchema() ← create/repair everything skipped before migrations
5. Seed admin, start services
```
### Why the reconciler runs twice
### Why migration sequencing differs by provider
Neither position alone works:
Neither a single reconciliation position nor one shared provider sequence works:
- **Pass 1 must come first.** A legacy database has hand-added columns and Identity tables that
predate the migrations; without repairing them (and stamping the legacy migration id into
`__EFMigrationsHistory`) `Migrate()` collides with them. `AddCareerProfileRelationalChildren` also
adds children that reference `CareerProfiles`, a **reconciler-owned** table — so it must exist
before migrations run.
- **Pass 2 must come after.** On a brand-new database the migration-owned tables do not exist during
- **The final pass must come after.** On a brand-new database the migration-owned tables do not exist during
pass 1, so every reconciler table that references one (FK into `JobApplications`) is skipped, as
are the index and `AUTO_INCREMENT` repairs.
- **SQLite reconciles between migrations.** Historical SQLite table rebuilds read the current model
shape, including columns that were originally supplied by reconciliation. The per-migration pass
establishes that shape before a later rebuild reads it.
- **MariaDB does not reconcile between migrations.** Its ALTER operations do not use SQLite table
rebuilds, and an intermediate pass could create a later migration's column early and cause a
duplicate-column failure. It applies the chain first and uses the shared final repair pass.
Every statement in `ReconcileSchema` is existence-guarded, so the second pass is a no-op scan on an
already-correct database. Two consequences worth knowing:
@@ -59,7 +66,9 @@ already-correct database. Two consequences worth knowing:
Created by EF migrations, never by the reconciler:
`Companies`, `JobApplications`, `Jobs`, `Correspondences`, `Attachments`, `JobEvents`,
`RuleSettings`, and the ASP.NET Identity tables.
`RuleSettings`, and the ASP.NET Identity tables. Two compatibility migrations use guarded
`CREATE TABLE IF NOT EXISTS` bootstraps for `AspNetUsers` and `AiInteractions` so standalone EF
tooling can traverse the historical chain; normal application startup makes those statements no-ops.
The reconciler may **repair** these (add a missing column, add an index, fix a non-`AUTO_INCREMENT`
primary key) and may seed the default `RuleSettings` row — but it must never `CREATE TABLE` them.
@@ -133,6 +142,12 @@ dotnet run --project JobTrackerApi/JobTrackerApi.csproj
Create the empty schema/database itself (`CREATE DATABASE jobtracker;`); the application builds
everything inside it.
Standalone EF tooling is also supported for a blank SQLite database. The historical initial
migration now supplies the stable JobApplication columns required by later SQLite rebuilds, and
guarded compatibility bootstraps provide the reconciler-owned source tables used by later additive
migrations. Application startup may subsequently reconcile the remaining Identity and auxiliary
tables without losing rows.
## Production upgrade
Deploy and restart. The reconciler is idempotent and additive:
@@ -159,3 +174,11 @@ All four scenarios, 2026-07-19, against MariaDB 11 and SQLite:
Column types on MariaDB spot-checked: `int AUTO_INCREMENT` primary keys, `varchar(255)` owner keys,
`datetime(6)` timestamps, `tinyint(1)` booleans, and every composite index inside the key limit.
On 2026-08-15 the current 29-migration chain was additionally verified against a blank standalone
SQLite database, an older populated SQLite checkpoint, and a disposable MariaDB 11.8 database.
Standalone SQLite migration and retry both reached the latest migration; populated title/date and
reconciler-owned owner/summary data survived. Starting the application over that EF-only database
served `/health` successfully. Fresh MariaDB startup and restart both served `/health` with 29
migrations and 49 tables; provider-sensitive ID, owner, decimal and timestamp column types were
spot-checked. No production database was changed.
@@ -45,6 +45,6 @@ The exact isolated API process was stopped and port 5303 was confirmed closed. L
## Limitations
- Browser localhost access remains denied by the in-app browser administrator policy; no browser claim is made.
- No MariaDB server was available. Pomelo SQL generation passed, but execution awaits a disposable or production-safe MariaDB smoke.
- Direct `dotnet ef database update` against a completely blank SQLite file still fails in the pre-existing reconciler-owned schema gap at `AddJobEntityAndProspectStages`. The documented application startup path succeeds because the reconciler establishes those columns before migrations. This is JT-019 schema-ownership debt, not the JT-003 query defect, and historical migrations were not changed.
- Disposable MariaDB 11.8 execution now passes for a fresh application start and restart with all 29 migrations and 49 tables. Production MariaDB execution remains unverified.
- Direct `dotnet ef database update` against a blank SQLite file now reaches the latest migration and is idempotent. A populated older checkpoint preserves its job data through the same chain, and real application startup over the EF-only database serves `/health` (V-186). The broader migration/reconciler dual-ownership architecture remains JT-019 debt.
- Execution policy denied deletion of the exact disposable nested data directory; it is stopped and recorded in the session handoff.
+2 -2
View File
@@ -1,6 +1,6 @@
# MAIL-001 consolidated job-email hub
Updated: 2026-08-10
Updated: 2026-08-15
Status: `IN PROGRESS`. Canonical hub routing, provider-neutral reads, explicit-confirmed send API, persisted reply/new-message UI, interrupted-send recovery, legacy SMTP retirement, send-attempt/draft export coverage, shared application context and tenant-owned draft persistence/API are implemented and locally verified; remaining provider mailbox actions and full account-deletion lifecycle remain.
@@ -142,7 +142,7 @@ Status: `IN PROGRESS`. Canonical hub routing, provider-neutral reads, explicit-c
- The owning job has a cascade relationship; a real-SQLite two-owner test proves User A sees only User A's draft and deleting User A's job removes only that draft while preserving User B's data.
- The additive migration has explicit SQLite and MariaDB types plus reversible down SQL. EF reports the model current; backend passes 625/625 and both provider scripts generate successfully.
- No route, UI, provider call, token or content log was added. Export coverage and the complete SEC-009 deletion lifecycle remain prerequisites before private draft content becomes reachable.
- A disposable full migration-chain SQLite rehearsal is blocked in the older `AddJobEntityAndProspectStages` migration because it references `LastReminderEmailSentAt` before any migration creates it. The failure occurs before `AddEmailDrafts` and remains tracked as JT-019 schema-chain debt.
- The repaired historical chain now reaches this migration from a blank standalone SQLite database and remains idempotent. A populated older checkpoint also preserves job data through the chain (V-186); production migration remains gated.
## Implemented readable draft export coverage
+2 -2
View File
@@ -12,10 +12,10 @@ Updated: 2026-08-15
- **Deferred work:** None. Conditional multi-replica coordination, model deletion, realtime operation delivery and unrelated production changes remain outside current packages.
- **Immediate order:** the eight-item immediate queue is complete locally: admin version (`a6cffe0`), Career persistence (`f0b9b22`), CV contrast (`3b86ea2`), JOBS-002 (`deed948`), accessibility (`a7c2549`), PRODUCT-001 (`a25c31b`), VER-001 and tracking reconciliation. SEC-009 is complete at `842e793`; PROD-001 read-only evidence and the PROD-003 plan-only harness are complete pending this documentation commit. No further independent implementation remains.
- **Status counts:** 8 `VERIFIED LOCALLY`; 25 `IMPLEMENTED — NOT VERIFIED`; 0 `IN PROGRESS`; 1 `NOT STARTED`; 5 `BLOCKED`; 0 `DONE`; 0 `DEFERRED`.
- **Test status:** backend 657/657; frontend 58/58 suites and 237/237 tests; AI sidecar 22/22; Ollama benchmark harness 4/4 plus safe dry-run; optimized production build/TypeScript; EF model parity and MariaDB migration-script generation; Docker Compose config; safe-failure deployment preflight; and Playwright 9/9. Five real-SQLite deletion tests cover lockout, isolation, quarantine failure and restored-backup replay. npm audit 0 evidence remains current because the lockfile did not change. Historical JT-019 and Jest force-exit/open-handle behavior remain recorded.
- **Test status:** backend 680/680; frontend 58/58 suites and 237/237 tests; AI sidecar 23/23; Ollama benchmark harness 5/5 plus safe dry-run; optimized production build/TypeScript; EF model parity; SQLite/MariaDB migration scripts; blank/idempotent/populated SQLite migration-chain tests; disposable fresh/restarted MariaDB 11.8 application startup; Docker Compose config; safe-failure deployment preflight; and Playwright 9/9. npm audit 0 evidence remains current because the lockfile did not change. Jest force-exit/open-handle behavior remains recorded.
- **Deployment status:** Gitea pull-request run 609 passed the complete CI job in 4m20s. Deploy was intentionally skipped because the workflow deploys only a `push` to `main`; live remains unchanged. No merge/deployment was performed directly, no production migrations were run and the AI operation worker remains disabled by default.
- **Production status:** State unchanged. Sanitized read-only SSH inventory was performed; no logs, prompts, private rows/content or secret values were read, and no provider/model call, model pull, service restart, file/config change, backup, restore, migration or deployment occurred. It confirmed all-interface Ollama/frontend listeners and stale database-only backups as rollout blockers.
- **Known regressions:** None found by automated/local browser checks. Jest still needs `--forceExit` and reports its existing open-handle notice. Email-provider/send tests are fake/local only; real delivery is not claimed. Current MAIL browser evidence is 1280×720 only because the browser surface could not resize or perform native Tab traversal. Interrupted attempts are aged after 15 minutes and notified without retry; the five-minute scan is unmeasured on a large ledger. Direct clean EF-only SQLite migration still hits the pre-existing historical blank-chain defect before later migrations; normal startup owns reconciliation. Cross-feature monthly AI usage accounting remains a rollout gap.
- **Known regressions:** None found by automated/local browser checks. Jest still needs `--forceExit` and reports its existing open-handle notice. Email-provider/send tests are fake/local only; real delivery is not claimed. Current MAIL browser evidence is 1280×720 only because the browser surface could not resize or perform native Tab traversal. Interrupted attempts are aged after 15 minutes and notified without retry; the five-minute scan is unmeasured on a large ledger. The direct clean EF-only SQLite defect and synchronous AI accounting gap are closed; migration/reconciler dual ownership remains architectural debt.
- **Outstanding security findings:** JT-001 repository ownership remains High deployment risk until migration/inventory/provider checks; production portion of JT-002; JT-006 and SEC-009 production retention/restore plus JT-011/JT-012/JT-022 prerequisites. JT-005 foundations are implemented; AI worker activation awaits controlled rollout. Production still exposes ports contrary to the release-branch contract, and JT-007/JT-008/JT-010 lack provider/production verification.
## Current evidence
+14 -8
View File
@@ -62,6 +62,12 @@ This queue records the highest-value work that can proceed without production cr
| 6 | Honest Free/Pro homepage and upgrade surfaces | PRODUCT-001 | Locally complete. One catalogue drives exactly Free/Pro; retired tier/price/Free-AI/unlimited claims are gone; configured billing state controls the real upgrade action; contextual notices are reusable/dismissible. Frontend 232/232, policy/billing 30/30, build and Chromium 8/8 pass. |
| 7 | Complete application action matrix and full regression | VER-001 | Locally complete. Backend 647/647, frontend 232/232, sidecar 22/22, build, Compose configuration, safe-failure preflight and Chromium 9/9 pass; external provider/native-AT/production cells remain explicitly unverified. |
| 8 | Tracking and blocker reconciliation | All | Complete for this checkpoint. The plan, progress, handoff, verification log, action matrix and `BLOCKERS.md` distinguish repository work from external gates; continue updating them with each later package. |
| 9 | Complete live-account deletion cache/tombstone safety | SEC-009 | Locally complete. Sidecar-cache failure is retryable and fail-closed; tombstones use separate persistent storage; activation remains disabled pending retention and restore decisions. |
| 10 | Prove worker clocks and restart idempotency | BG-001 | Locally complete with injected clocks, exact-threshold tests, fresh worker instances and reminder/export deduplication. Production canary remains disabled. |
| 11 | Universal synchronous AI usage admission | POL-001 | Locally complete. Shared generation paths reject Free/exhausted users before provider I/O and avoid double-counting durable/workspace operations. |
| 12 | Prove email-token and Stripe downgrade lifecycles | SEC-005B, POL-001 | Locally complete with real Identity-token replay/expiry/custom-username tests and fake-gateway active/past-due/canceled Stripe transitions. External SMTP/Stripe journeys remain blocked. |
| 13 | Remove the Job email 100-application selector ceiling | MAIL-001 | Locally complete through bounded owner-filtered server search and tenant/UI regressions. |
| 14 | Repair the full historical migration chain | CORE-001, JT-019 | Locally complete for blank/idempotent/populated SQLite, EF-only-to-application startup, provider scripts and disposable MariaDB 11.8 fresh/restart. Production restore/rollout remains blocked. |
## Requirement coverage index
@@ -287,10 +293,10 @@ This queue records the highest-value work that can proceed without production cr
- **Required browser verification:** SQLite Career/Application workspace after API tests.
- **Required production verification:** MariaDB smoke after deployment.
- **Status:** `VERIFIED LOCALLY`.
- **Blocker:** production MariaDB execution and browser checks remain unavailable; direct blank-file EF-only migration is separate JT-019 schema-ownership debt while fresh application startup passes.
- **Evidence:** audit runtime reproduction JT-003; `docs/verification/core-001-sqlite-provider-parity.md`; 3/3 real-provider tests; 509/509 backend regression; isolated fresh-SQLite owner/empty/non-owner HTTP matrix.
- **Blocker:** production MariaDB execution and production browser checks remain unavailable. The direct blank-file EF-only defect is closed; historical dual schema ownership remains JT-019 architectural debt.
- **Evidence:** audit runtime reproduction JT-003; `docs/verification/core-001-sqlite-provider-parity.md`; V-186; 3/3 real-provider query tests; 3/3 migration-chain tests; 680/680 backend regression; isolated fresh-SQLite owner/empty/non-owner HTTP matrix; disposable MariaDB 11.8 fresh/restart smoke.
- **Commit:** none.
- **Remaining work:** browser Career/Application workspace verification and executable MariaDB smoke after safe provider/deployment access; address EF-only blank-chain drift under JT-019 rather than editing already-applied historical migrations here.
- **Remaining work:** production Career/Application browser verification and production MariaDB restore/rollout smoke. Keep the compatibility migration/reconciler contract covered until a later expand/verify/contract release can consolidate ownership safely.
### CORE-002 — Remove ambiguous application-workspace routes
@@ -341,10 +347,10 @@ This queue records the highest-value work that can proceed without production cr
- **Required browser verification:** not applicable until OPS-001C exposes owner APIs.
- **Required production verification:** executable MariaDB upgrade/down rehearsal and monitored schema rollout.
- **Status:** `IMPLEMENTED — NOT VERIFIED`.
- **Blocker:** no disposable MariaDB or production environment; application consumers deliberately not migrated yet.
- **Blocker:** disposable fresh/restart MariaDB now passes; production schema rollout and monitored consumer canary remain unavailable.
- **Evidence:** `docs/verification/ops-001a-durable-operations.md`; 7/7 focused and 539/539 full backend tests; SQLite upgrade/down/up and fresh startup; dual-provider scripts.
- **Commit:** none.
- **Remaining work:** MariaDB execution/production rollout; task-specific producers must validate references/policies and use OPS-001B/C rather than storing private payloads.
- **Remaining work:** production rollout; task-specific producers must validate references/policies and use OPS-001B/C rather than storing private payloads.
### OPS-001B — Persistent operation notifications and terminal outbox
@@ -359,10 +365,10 @@ This queue records the highest-value work that can proceed without production cr
- **Required browser verification:** deferred to OPS-001C.
- **Required production verification:** schema rollout and synthetic notification canary only.
- **Status:** `IMPLEMENTED — NOT VERIFIED`.
- **Blocker:** MariaDB execution unavailable; repository implementation can continue.
- **Blocker:** disposable fresh/restart MariaDB now passes; production migration/canary remains unavailable.
- **Evidence:** `docs/verification/ops-001b-notifications.md`; 9/9 focused and 541/541 full backend tests; forced transaction rollback; SQLite upgrade/down/up; current model snapshot; generated SQLite/MariaDB up/down SQL.
- **Commit:** none.
- **Remaining work:** execute the migration on MariaDB; expose owner APIs/UI in OPS-001C; complete browser and production canaries. No email delivery is part of this package.
- **Remaining work:** complete production schema/notification canaries. Owner APIs/UI are implemented in OPS-001C; no email delivery is part of this package.
### OPS-001C — Owner operation/notification APIs and frontend queue client
@@ -668,7 +674,7 @@ This queue records the highest-value work that can proceed without production cr
- **Blocker:** real provider/re-consent, full SEC-009 deletion, MariaDB, production and required 375/768/1440/theme/keyboard browser gates are unavailable or require new authority.
- **Evidence:** `docs/verification/mail-001-job-email-hub.md`; V-126V-153. Draft/new-message UI 13/13, API/idempotency/rotation 10/10, persistence 1/1 with dual-provider reversible SQL and readable export 4/4; Free send policy 7/7; provider states 9/9; hub unlink 8/8 UI and 2/2 API; shared application context focused 10/10; prior send export/cascade focused 16/16; recovery/send focused 10/10; legacy follow-up/worker 10/10; delivery/capability 18/18; provider/correspondence 5/5; hub detail 5/5; backend 630/630; frontend 50/50 suites and 198/198 tests plus build/audit; local empty/disconnected and compatibility-route browser smoke at 1280×720.
- **Commit:** `6008b4a` (hub), `536d403` (neutral reads), `a20775c` (safe detail), `653f011` (ledger), `e9937ac` (Gmail/Graph delivery adapters and consent), `123fc55` (explicit-confirmed send API), `449faeb` (confirmed reply composer), `ee5ef7e` (interrupted-send recovery), `8fe3903` (legacy SMTP retirement), `aff34cc` (content-free export and cascade evidence), `ff547df` (shared application context), `1dabbeb` (confirmed hub unlink), `f9e641c` (honest provider states), `7f41cb2` (Free email policy regression), `14b396a` (inert tenant draft persistence), `2fa4e38` (owner-isolated readable draft export), `a9bb22e` (tenant-safe revisioned draft API), `80b5532` (persisted draft send identity), `d3d2b67` (saved reply recovery/conflicts), `29de263` (definitive-failure identity rotation), `b735963` (new-message job/provider drafting).
- **Remaining work:** full account deletion remains SEC-009; provider mailbox category capabilities require separately authorized scopes/re-consent and remain absent; browser/production verification. Server-side owner-filtered job choice search now reaches applications beyond the old first-page ceiling (V-183). Existing connections need explicit re-consent; IMAP remains read-only. No real email; uncertain sends need manual reconciliation. JT-019 blocks a clean full-chain SQLite rehearsal before the new draft migration.
- **Remaining work:** SEC-009 repository deletion coverage is complete but production activation remains gated; provider mailbox category capabilities require separately authorized scopes/re-consent and remain absent; browser/production verification remains. Server-side owner-filtered job choice search now reaches applications beyond the old first-page ceiling (V-183). Existing connections need explicit re-consent; IMAP remains read-only. No real email; uncertain sends need manual reconciliation. The clean full-chain SQLite rehearsal now passes (V-186).
### JOBS-001 — Job-search source and assessment redesign
+2 -2
View File
@@ -3,7 +3,7 @@
Updated: 2026-08-15
- **Exact current task:** no independent implementation remains. SEC-009, PROD-001 read-only inventory/reporting and the PROD-003 plan-only benchmark harness are complete; continue only after a recorded blocker is authorized/resolved.
- **Last completed step:** measured production hardware/runtime/backups read-only, documented rollout stops, then added a synthetic-only benchmark harness that defaults to no network execution.
- **Last completed step:** repaired and regression-tested the full historical migration chain for standalone SQLite tooling and provider-aware application startup, including a disposable MariaDB 11.8 fresh/restart rehearsal.
- **Files currently modified:** master progress/work-plan/handoff/decisions/blockers/evidence; SEC-009 verification; production hardware/rollout/benchmark reports; production backup checkpoint; Ollama benchmark script/tests.
- **Commands already run:** SEC-009 focused backend 21/21, backend 657/657, frontend focused 8/8 and full 237/237, optimized build, EF parity, MariaDB script generation and Chromium 9/9; benchmark harness 4/4 plus plan-only dry run; sanitized read-only SSH inventory and gzip integrity across 21 existing dumps.
- **Test results:** all repository gates pass. PROD-001 is PASS/PARTIAL because measured all-interface ports and incomplete/stale backup/restore evidence fail its safety acceptance. No model inference was run. Jest retains the documented open-handle notice.
@@ -12,7 +12,7 @@ Updated: 2026-08-15
- **Production changes currently active:** none. Read-only SSH observed metadata/health/selected non-secret settings and backup integrity only. No log/private-row/content/secret read, deployment, migration, provider call, inference, model pull, restart, backup, restore or production file/config change occurred.
- **Rollback status:** SEC-009 is additive migration `20260815164027_AddAccountDeletionLifecycle`. Keep deletion disabled, reconcile any durable request, and retain tombstones before downgrade. `842e793` is pushed. Production still runs `de937d25dc5e` / version `157`; its checkout has a pre-existing mode-only `deploy/deploy.sh` change that must be preserved/reviewed.
- **Uncommitted changes:** documentation and the plan-only Ollama benchmark harness/tests following pushed SEC-009 commit `842e793`; no dependency, model, application runtime, schema or production state change in this checkpoint.
- **Known failures:** PR 28's newest remote CI is not yet confirmed. Production publishes frontend 3000 and JobTracker Ollama 11434 on all interfaces; latest observed database-only backup is 2026-08-02; no complete files/keys/tombstone restore proof; root is 83% used; deployed AI sidecar is old direct-Gemini behavior; production script mode is dirty. SEC-006 internet access, SEC-007 dependency, provider/re-consent, Stripe price, signup, retention/legal, backup/restore, model execution/deployment and legacy cutover decisions remain recorded blockers. Historical JT-019 and Jest open handles remain.
- **Known failures:** PR 28's newest remote CI is not yet confirmed. Production publishes frontend 3000 and JobTracker Ollama 11434 on all interfaces; latest observed database-only backup is 2026-08-02; no complete files/keys/tombstone restore proof; root is 83% used; deployed AI sidecar is old direct-Gemini behavior; production script mode is dirty. SEC-006 internet access, SEC-007 dependency, provider/re-consent, Stripe price, signup, retention/legal, backup/restore, model execution/deployment and legacy cutover decisions remain recorded blockers. JT-019's blank-chain defect is fixed, while dual ownership remains debt; Jest open handles remain.
- **Exact next action:** after this documentation/harness checkpoint is committed and pushed, stop. Resume from the highest-priority blocker the user authorizes: recommended first is production network plus complete backup/scratch-restore safety, then bounded synthetic model benchmarking.
- **Work that can continue independently:** none identified after the PROD-003 harness. Do not bypass blockers by pulling models, changing ports/firewalls, restoring data, contacting providers, or using package indexes without explicit authority.
- **Decisions still required from the user:** production network/port mutation; complete backup and scratch restore; retention/tombstone/legal policy; model pull/synthetic production inference; deployment/worker activation; parser package-index access; provider/Stripe/signup actions; legacy cutover timing.