docs/ARCHITECTURE.md

@@ -226,7 +226,7 @@ automatic** (D-021).
   method, minutes_total or difficulty a member saved while the dinner was
   `pending` is kept; grounding fills only those left empty. At boot, after
   the job runner fails the jobs a previous process left running, every
-  dinner still `pending` with no `ground_dinner` job for its revision
+  unarchived dinner still `pending` with no `ground_dinner` job for its revision
   queued or running is set `failed` with `interrupted`, so no dinner stays
   pending with nothing left to ground it.
   `GET /api/v1/dinners/{id}/recipe-candidates`

internal/db/dinners.sql.go

@@ -150,6 +150,7 @@ SET grounding_status = 'failed',
 	grounding_error = ?1,
 	updated_at = ?2
 WHERE grounding_status = 'pending'
+	AND status <> 'archived'
 	AND NOT EXISTS (
 		SELECT 1 FROM jobs
 		WHERE jobs.kind = ?3
@@ -165,10 +166,11 @@ type FailOrphanedPendingDinnerGroundingsParams struct {
 	Kind           string
 }
 
-// Fails every dinner still pending whose grounding job for its current
-// grounding_revision is no longer queued or running: the process stopped
-// before the job, or before its handler recorded a result. Run at boot,
-// after the job runner fails the jobs a previous process left running.
+// Fails every unarchived dinner still pending whose grounding job for its
+// current grounding_revision is no longer queued or running: the process
+// stopped before the job, or before its handler recorded a result. An
+// archived dinner is left as it is; archiving stopped its grounding. Run at
+// boot, after the job runner fails the jobs a previous process left running.
 func (q *Queries) FailOrphanedPendingDinnerGroundings(ctx context.Context, arg FailOrphanedPendingDinnerGroundingsParams) (int64, error) {
 	result, err := q.db.ExecContext(ctx, failOrphanedPendingDinnerGroundings, arg.GroundingError, arg.UpdatedAt, arg.Kind)
 	if err != nil {

internal/db/querier.go

@@ -36,10 +36,11 @@ type Querier interface {
 	DeleteRecipeIngredients(ctx context.Context, recipeID string) error
 	DeleteRecipeSteps(ctx context.Context, recipeID string) error
 	FailJob(ctx context.Context, arg FailJobParams) (int64, error)
-	// Fails every dinner still pending whose grounding job for its current
-	// grounding_revision is no longer queued or running: the process stopped
-	// before the job, or before its handler recorded a result. Run at boot,
-	// after the job runner fails the jobs a previous process left running.
+	// Fails every unarchived dinner still pending whose grounding job for its
+	// current grounding_revision is no longer queued or running: the process
+	// stopped before the job, or before its handler recorded a result. An
+	// archived dinner is left as it is; archiving stopped its grounding. Run at
+	// boot, after the job runner fails the jobs a previous process left running.
 	FailOrphanedPendingDinnerGroundings(ctx context.Context, arg FailOrphanedPendingDinnerGroundingsParams) (int64, error)
 	// Records why the background grounding of an unarchived dinner still
 	// pending at the grounding_revision it was grounded for failed; a dinner

internal/dinners/grounding.go

@@ -227,10 +227,11 @@ func (s *Service) stopGroundings(dinnerID string, below int64) {
 	}
 }
 
-// FailInterruptedGroundings fails, with ReasonInterrupted, every dinner
-// still pending whose KindGroundDinner job is no longer queued or running:
-// a previous process stopped before enqueuing the job or before its
-// handler recorded a result. Call it once at boot, after the job runner's
+// FailInterruptedGroundings fails, with ReasonInterrupted, every unarchived
+// dinner still pending whose KindGroundDinner job is no longer queued or
+// running: a previous process stopped before enqueuing the job or before
+// its handler recorded a result. An archived dinner is left pending, since
+// archiving, not a restart, stopped its grounding. Call it once at boot, after the job runner's
 // Start has failed the jobs that process left running, so no dinner stays
 // pending with nothing left to ground it. It returns how many it failed.
 func (s *Service) FailInterruptedGroundings(ctx context.Context) (int64, error) {

internal/dinners/grounding_test.go

@@ -946,12 +946,64 @@ func TestRestartFailsInterruptedGrounding(t *testing.T) {
 	}
 }
 
+// TestRestartKeepsArchivedPendingDinner: a pending Regular archived while
+// its job runs stays pending, its job succeeded; the boot cleanup leaves it
+// as it is rather than failing it as interrupted.
+func TestRestartKeepsArchivedPendingDinner(t *testing.T) {
+	t.Parallel()
+	e := newEnv(t)
+	e.search.Gate = make(chan struct{})
+	c, err := e.svc.CreateRegular(t.Context(), dinners.NewRegular{Name: ptr("Chicken curry")}, member)
+	if err != nil {
+		t.Fatal(err)
+	}
+	e.waitQuery(t, "Chicken curry")
+	if _, err := e.svc.Archive(t.Context(), c.ID, member); err != nil {
+		t.Fatal(err)
+	}
+	job := e.waitJob(t, *c.GroundingJobID)
+	close(e.search.Gate)
+	if job.Status != jobs.StatusSucceeded {
+		t.Fatalf("job = %+v, want succeeded", job)
+	}
+
+	// The database is now as the next boot finds it.
+	n, err := e.svc.FailInterruptedGroundings(t.Context())
+	if err != nil {
+		t.Fatal(err)
+	}
+	if n != 0 {
+		t.Errorf("failed %d dinners, want 0", n)
+	}
+	if row := e.dinner(t, c.ID); row.Status != dinners.StatusArchived ||
+		row.GroundingStatus != dinners.GroundingPending || row.GroundingError.Valid {
+		t.Errorf("dinner = %s, %s (%s), want archived and pending with no error",
+			row.Status, row.GroundingStatus, row.GroundingError.String)
+	}
+}
+
 // TestUpgradeKeepsLegacyQueuedGrounding: a pending Regular whose
 // ground_dinner job was queued before migration 0010, so its payload has
 // no revision, is not failed as interrupted at the first boot after the
-// upgrade; its job still grounds it.
+// upgrade; its job still grounds it. The upgrade starts either from 0009
+// or from a database that has already applied 0010.
 func TestUpgradeKeepsLegacyQueuedGrounding(t *testing.T) {
 	t.Parallel()
+	for name, below := range map[string]string{
+		"from 0009": "0010",
+		"from 0010": "0011",
+	} {
+		t.Run(name, func(t *testing.T) {
+			t.Parallel()
+			upgradeWithLegacyQueuedGrounding(t, below)
+		})
+	}
+}
+
+// upgradeWithLegacyQueuedGrounding migrates a database to every migration
+// named before below, adds a pending Regular with a queued ground_dinner
+// job whose payload has no revision, then boots on the latest migrations.
+func upgradeWithLegacyQueuedGrounding(t *testing.T, below string) {
 	ctx := t.Context()
 	dataDir := t.TempDir()
 	st := openStore(t, dataDir)
@@ -961,7 +1013,7 @@ func TestUpgradeKeepsLegacyQueuedGrounding(t *testing.T) {
 		t.Fatal(err)
 	}
 	for _, name := range files {
-		if name >= "0010" {
+		if name >= below {
 			continue
 		}
 		body, err := fs.ReadFile(migrations.FS, name)

internal/store/migrate_test.go

@@ -175,8 +175,8 @@ func TestMigrateEmbeddedMigrations(t *testing.T) {
 	if err := Migrate(context.Background(), s.Writer, migrations.FS); err != nil {
 		t.Fatal(err)
 	}
-	if got := versions(migrationRows(t, s)); got != "0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision" {
-		t.Errorf("versions = %q, want 0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision", got)
+	if got := versions(migrationRows(t, s)); got != "0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision,0011_ground_dinner_job_revision" {
+		t.Errorf("versions = %q, want 0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision,0011_ground_dinner_job_revision", got)
 	}
 }
 
@@ -196,8 +196,8 @@ func TestMigrateBootstrapThenEmbeddedInitApplies(t *testing.T) {
 	if err := Migrate(ctx, s.Writer, migrations.FS); err != nil {
 		t.Fatal(err)
 	}
-	if got := versions(migrationRows(t, s)); got != "0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision" {
-		t.Errorf("versions = %q, want 0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision", got)
+	if got := versions(migrationRows(t, s)); got != "0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision,0011_ground_dinner_job_revision" {
+		t.Errorf("versions = %q, want 0000_bootstrap,0001_init,0002_household_settings_updated_by,0003_jobs_status,0004_ai_provider_presets,0005_ai_provider_xai,0006_ai_runs_status,0007_recipe_grounding,0008_dinner_recipe_url,0009_dinner_grounding_error,0010_dinner_grounding_revision,0011_ground_dinner_job_revision", got)
 	}
 	if !tableExists(t, s, "dinners") {
 		t.Error("table from 0001_init was not created")

migrations/0010_dinner_grounding_revision.sql

@@ -5,12 +5,3 @@
 -- earlier job — queued or running — can no longer write the dinner.
 
 ALTER TABLE dinners ADD COLUMN grounding_revision INTEGER NOT NULL DEFAULT 0;
-
--- Every dinner starts at revision 0, so a ground_dinner job enqueued
--- before this migration, whose payload has only dinner_id, is for revision
--- 0. Recording it lets the boot cleanup, which matches a pending dinner to
--- its job by revision, see a job still queued for its dinner.
-UPDATE jobs
-SET payload = json_set(payload, '$.revision', 0)
-WHERE kind = 'ground_dinner'
-	AND json_type(payload, '$.revision') IS NULL;

migrations/0011_ground_dinner_job_revision.sql

@@ -0,0 +1,11 @@
+-- 0011_ground_dinner_job_revision: every dinner starts at grounding
+-- revision 0 (0010), so a ground_dinner job enqueued before 0010, whose
+-- payload has only dinner_id, is for revision 0. Recording it lets the boot
+-- cleanup, which matches a pending dinner to its job by revision, see a job
+-- still queued for its dinner. This is a migration of its own, not part of
+-- 0010, so it also runs on a database that has already applied 0010.
+
+UPDATE jobs
+SET payload = json_set(payload, '$.revision', 0)
+WHERE kind = 'ground_dinner'
+	AND json_type(payload, '$.revision') IS NULL;

migrations/queries/dinners.sql

@@ -134,15 +134,17 @@ WHERE id = ?
 	AND grounding_revision = ?;
 
 -- name: FailOrphanedPendingDinnerGroundings :execrows
--- Fails every dinner still pending whose grounding job for its current
--- grounding_revision is no longer queued or running: the process stopped
--- before the job, or before its handler recorded a result. Run at boot,
--- after the job runner fails the jobs a previous process left running.
+-- Fails every unarchived dinner still pending whose grounding job for its
+-- current grounding_revision is no longer queued or running: the process
+-- stopped before the job, or before its handler recorded a result. An
+-- archived dinner is left as it is; archiving stopped its grounding. Run at
+-- boot, after the job runner fails the jobs a previous process left running.
 UPDATE dinners
 SET grounding_status = 'failed',
 	grounding_error = sqlc.arg(grounding_error),
 	updated_at = sqlc.arg(updated_at)
 WHERE grounding_status = 'pending'
+	AND status <> 'archived'
 	AND NOT EXISTS (
 		SELECT 1 FROM jobs
 		WHERE jobs.kind = sqlc.arg(kind)