fix(preview): propagate cleanup errors and defer container removal

cleanupOrphanDomainResources now returns an error. If resource cleanup
fails, the container is not removed — it stays stopped and will be
rediscovered on the next GC scan for retry.

commit: 8b31c1c17e38858a9e1d35417b97b4456432d20d
author: brice · 7mo ago · 2026-03-25 08:00 UTC
parent: b32e7e8
signature: GPG VerifiedCD88F9E564E0102646FB6077A1A07C7C7C3E349D · arch-gpg Verified for the committer email

1 file · +19 -4

M internal/app/run.go +19 -4


@@ -652,7 +652,11 @@ func gcOrphanedPreviews(ctx context.Context, svc *services, previewConfig domain
652	652		}
653	653
654	654		if orphanDomain != "" {
655		-	cleanupOrphanDomainResources(ctx, svc, orphanDomain)
	655	+	if err := cleanupOrphanDomainResources(ctx, svc, orphanDomain); err != nil {
	656	+	log.Warn().Err(err).Str("container", c.Name).Str("domain", orphanDomain).
	657	+	Msg("orphan resource cleanup failed, deferring container removal to next scan")
	658	+	continue
	659	+	}
656	660		}
657	661
658	662		if err := svc.runtime.RemoveContainer(ctx, c.Name, true); err != nil {

@@ -663,20 +667,27 @@ func gcOrphanedPreviews(ctx context.Context, svc *services, previewConfig domain
663	667		}
664	668		}
665	669
666		-	// cleanupOrphanDomainResources removes volumes, network, and route for an orphaned preview domain.
667		-	func cleanupOrphanDomainResources(ctx context.Context, svc *services, orphanDomain string) {
	670	+	// cleanupOrphanDomainResources removes volumes, network, and route for an
	671	+	// orphaned preview domain. Returns an error if any step fails so the caller
	672	+	// can defer container removal until the next scan.
	673	+	func cleanupOrphanDomainResources(ctx context.Context, svc *services, orphanDomain string) error {
668	674		log := zerowrap.FromCtx(ctx)
669	675		domainSanitized := strings.ReplaceAll(orphanDomain, ".", "-")
	676	+	var cleanupErr error
670	677
671	678		// Remove volumes matching domain prefix (convention: {volPrefix}-{domain-sanitized}-).
672	679		_, volPrefix, _ := svc.configSvc.GetVolumeConfig()
673	680		fullVolumePrefix := volPrefix + "-" + domainSanitized + "-"
674	681		volumes, err := svc.runtime.ListVolumes(ctx)
675		-	if err == nil {
	682	+	if err != nil {
	683	+	log.Warn().Err(err).Msg("failed to list volumes for orphan cleanup")
	684	+	cleanupErr = err
	685	+	} else {
676	686		for _, v := range volumes {
677	687		if strings.HasPrefix(v.Name, fullVolumePrefix) {
678	688		if err := svc.runtime.RemoveVolume(ctx, v.Name, true); err != nil {
679	689		log.Warn().Err(err).Str("volume", v.Name).Msg("failed to remove orphan volume")
	690	+	cleanupErr = err
680	691		}
681	692		}
682	693		}

@@ -687,15 +698,19 @@ func cleanupOrphanDomainResources(ctx context.Context, svc *services, orphanDoma
687	698		networkName := networkPrefix + "-" + domainSanitized
688	699		if err := svc.runtime.RemoveNetwork(ctx, networkName); err != nil {
689	700		log.Warn().Err(err).Str("network", networkName).Msg("failed to remove orphan network")
	701	+	cleanupErr = err
690	702		}
691	703
692	704		// Remove route and invalidate proxy.
693	705		if err := svc.configSvc.RemoveRoute(ctx, orphanDomain); err != nil {
694	706		if !errors.Is(err, domain.ErrRouteNotFound) {
695	707		log.Warn().Err(err).Str("domain", orphanDomain).Msg("failed to remove orphan route")
	708	+	cleanupErr = err
696	709		}
697	710		}
698	711		svc.proxySvc.InvalidateTarget(ctx, orphanDomain)
	712	+
	713	+	return cleanupErr
699	714		}
700	715
701	716		// injectTelemetryMetrics creates and injects OTel metrics into services when

 		}
 		if orphanDomain != "" {
-			cleanupOrphanDomainResources(ctx, svc, orphanDomain)
+			if err := cleanupOrphanDomainResources(ctx, svc, orphanDomain); err != nil {
+				log.Warn().Err(err).Str("container", c.Name).Str("domain", orphanDomain).
+					Msg("orphan resource cleanup failed, deferring container removal to next scan")
+				continue
+			}
 		}
 		if err := svc.runtime.RemoveContainer(ctx, c.Name, true); err != nil {
 	}
 }
-// cleanupOrphanDomainResources removes volumes, network, and route for an orphaned preview domain.
+// cleanupOrphanDomainResources removes volumes, network, and route for an
-func cleanupOrphanDomainResources(ctx context.Context, svc *services, orphanDomain string) {
+// orphaned preview domain. Returns an error if any step fails so the caller
+// can defer container removal until the next scan.
+func cleanupOrphanDomainResources(ctx context.Context, svc *services, orphanDomain string) error {
 	log := zerowrap.FromCtx(ctx)
 	domainSanitized := strings.ReplaceAll(orphanDomain, ".", "-")
+	var cleanupErr error
 	// Remove volumes matching domain prefix (convention: {volPrefix}-{domain-sanitized}-).
 	_, volPrefix, _ := svc.configSvc.GetVolumeConfig()
 	fullVolumePrefix := volPrefix + "-" + domainSanitized + "-"
 	volumes, err := svc.runtime.ListVolumes(ctx)
-	if err == nil {
+	if err != nil {
+		log.Warn().Err(err).Msg("failed to list volumes for orphan cleanup")
+		cleanupErr = err
+	} else {
 		for _, v := range volumes {
 			if strings.HasPrefix(v.Name, fullVolumePrefix) {
 				if err := svc.runtime.RemoveVolume(ctx, v.Name, true); err != nil {
 					log.Warn().Err(err).Str("volume", v.Name).Msg("failed to remove orphan volume")
+					cleanupErr = err
 				}
 			}
 		}
 	networkName := networkPrefix + "-" + domainSanitized
 	if err := svc.runtime.RemoveNetwork(ctx, networkName); err != nil {
 		log.Warn().Err(err).Str("network", networkName).Msg("failed to remove orphan network")
+		cleanupErr = err
 	}
 	// Remove route and invalidate proxy.
 	if err := svc.configSvc.RemoveRoute(ctx, orphanDomain); err != nil {
 		if !errors.Is(err, domain.ErrRouteNotFound) {
 			log.Warn().Err(err).Str("domain", orphanDomain).Msg("failed to remove orphan route")
+			cleanupErr = err
 		}
 	}
 	svc.proxySvc.InvalidateTarget(ctx, orphanDomain)
+	return cleanupErr
 }
 // injectTelemetryMetrics creates and injects OTel metrics into services when