93fc814 fix(runtime): harden reload and cleanup reporting · gordon

M docs/config/server.md +1 -0


@@ -438,6 +438,7 @@ gordon_domain = "gordon.mydomain.com"
438	438		```
439	439
440	440		With firewall port forwarding:
	441	+
441	442		```bash
442	443		sudo firewall-cmd --permanent --add-forward-port=port=80:proto=tcp:toport=8088
443	444		sudo firewall-cmd --reload


@@ -438,6 +438,7 @@ gordon_domain = "gordon.mydomain.com"
438	```	438		```
439		439
440	With firewall port forwarding:	440		With firewall port forwarding:
		441	+
441	```bash	442		```bash
442	sudo firewall-cmd --permanent --add-forward-port=port=80:proto=tcp:toport=8088	443		sudo firewall-cmd --permanent --add-forward-port=port=80:proto=tcp:toport=8088
443	sudo firewall-cmd --reload	444		sudo firewall-cmd --reload

M docs/index.md +0 -1


@@ -78,7 +78,6 @@ Gordon runs on your VPS and provides:
78	78		- Linux VPS (Ubuntu/Debian recommended)
79	79		- Docker or Podman runtime
80	80		- Domain pointing to your server
81		-	- Domain pointing to your server
82	81		- Optional Cloudflare account or other reverse proxy for edge TLS
83	82
84	83		> Note: Gordon can run behind Cloudflare/nginx, or terminate TLS directly via its internal CA (enabled by default on `server.tls_port`). Set `tls_port = 0` to disable.


@@ -78,7 +78,6 @@ Gordon runs on your VPS and provides:
78		- Linux VPS (Ubuntu/Debian recommended)	78	- Linux VPS (Ubuntu/Debian recommended)
79		- Docker or Podman runtime	79	- Docker or Podman runtime
80		- Domain pointing to your server	80	- Domain pointing to your server
81	-	- Domain pointing to your server
82		- Optional Cloudflare account or other reverse proxy for edge TLS	81	- Optional Cloudflare account or other reverse proxy for edge TLS
83			82
84		> Note: Gordon can run behind Cloudflare/nginx, or terminate TLS directly via its internal CA (enabled by default on `server.tls_port`). Set `tls_port = 0` to disable.	83	> Note: Gordon can run behind Cloudflare/nginx, or terminate TLS directly via its internal CA (enabled by default on `server.tls_port`). Set `tls_port = 0` to disable.

M internal/adapters/in/cli/auth.go +1 -1


@@ -238,7 +238,7 @@ Examples:
238	238		func resolveRequiredRemote(flagToken string) (*remote.ResolvedRemote, error) {
239	239		resolved, isRemote, err := remote.ResolveStrict(remoteFlag, flagToken, insecureTLSFlag)
240	240		if err != nil {
241		-	return nil, err
	241	+	return nil, fmt.Errorf("failed to resolve remote %q: %w", remoteFlag, err)
242	242		}
243	243		if !isRemote {
244	244		return nil, fmt.Errorf("no remote specified. Use --remote or 'gordon remotes use <name>'")


@@ -238,7 +238,7 @@ Examples:
238		func resolveRequiredRemote(flagToken string) (*remote.ResolvedRemote, error) {	238		func resolveRequiredRemote(flagToken string) (*remote.ResolvedRemote, error) {
239		resolved, isRemote, err := remote.ResolveStrict(remoteFlag, flagToken, insecureTLSFlag)	239		resolved, isRemote, err := remote.ResolveStrict(remoteFlag, flagToken, insecureTLSFlag)
240		if err != nil {	240		if err != nil {
241	-	return nil, err	241	+	return nil, fmt.Errorf("failed to resolve remote %q: %w", remoteFlag, err)
242		}	242		}
243		if !isRemote {	243		if !isRemote {
244		return nil, fmt.Errorf("no remote specified. Use --remote or 'gordon remotes use <name>'")	244		return nil, fmt.Errorf("no remote specified. Use --remote or 'gordon remotes use <name>'")

M internal/adapters/in/cli/routes.go +21 -1


@@ -1479,7 +1479,7 @@ func runRoutePurge(ctx context.Context, cp ControlPlane, domainName string, opts
1479	1479		if diag.Runtime != nil && diag.Runtime.ContainerID != "" {
1480	1480		report.OrphanedEntities = append(report.OrphanedEntities, domain.CleanupOrphanedEntity{Kind: "route_container", ID: diag.Runtime.ContainerID, Status: diag.Runtime.ContainerStatus, Reason: "route runtime state matched purge target"})
1481	1481		}
1482		-	report.OrphanedEntities = append(report.OrphanedEntities, diag.OrphanedEntities...)
	1482	+	report.OrphanedEntities = appendUniqueCleanupEntities(report.OrphanedEntities, diag.OrphanedEntities...)
1483	1483		if opts.Volumes {
1484	1484		for _, volume := range diag.Volumes {
1485	1485		report.PreservedVolumes = append(report.PreservedVolumes, domain.CleanupVolume{Name: volume.Name, Reason: "volume purge requires explicit --volumes and runtime support"})

@@ -1516,6 +1516,26 @@ func runRoutePurge(ctx context.Context, cp ControlPlane, domainName string, opts
1516	1516		return report, nil
1517	1517		}
1518	1518
	1519	+	func appendUniqueCleanupEntities(existing []domain.CleanupOrphanedEntity, candidates ...domain.CleanupOrphanedEntity) []domain.CleanupOrphanedEntity {
	1520	+	seen := make(map[string]struct{}, len(existing))
	1521	+	for _, entity := range existing {
	1522	+	seen[cleanupEntityKey(entity)] = struct{}{}
	1523	+	}
	1524	+	for _, entity := range candidates {
	1525	+	key := cleanupEntityKey(entity)
	1526	+	if _, ok := seen[key]; ok {
	1527	+	continue
	1528	+	}
	1529	+	existing = append(existing, entity)
	1530	+	seen[key] = struct{}{}
	1531	+	}
	1532	+	return existing
	1533	+	}
	1534	+
	1535	+	func cleanupEntityKey(entity domain.CleanupOrphanedEntity) string {
	1536	+	return entity.Kind + "\|" + entity.ID + "\|" + entity.Name
	1537	+	}
	1538	+
1519	1539		func writeRoutePurgeText(out io.Writer, report *domain.CleanupReport, force bool) error {
1520	1540		mode := "Purge dry-run"
1521	1541		if force {


@@ -1479,7 +1479,7 @@ func runRoutePurge(ctx context.Context, cp ControlPlane, domainName string, opts
1479		if diag.Runtime != nil && diag.Runtime.ContainerID != "" {	1479		if diag.Runtime != nil && diag.Runtime.ContainerID != "" {
1480		report.OrphanedEntities = append(report.OrphanedEntities, domain.CleanupOrphanedEntity{Kind: "route_container", ID: diag.Runtime.ContainerID, Status: diag.Runtime.ContainerStatus, Reason: "route runtime state matched purge target"})	1480		report.OrphanedEntities = append(report.OrphanedEntities, domain.CleanupOrphanedEntity{Kind: "route_container", ID: diag.Runtime.ContainerID, Status: diag.Runtime.ContainerStatus, Reason: "route runtime state matched purge target"})
1481		}	1481		}
1482	-	report.OrphanedEntities = append(report.OrphanedEntities, diag.OrphanedEntities...)	1482	+	report.OrphanedEntities = appendUniqueCleanupEntities(report.OrphanedEntities, diag.OrphanedEntities...)
1483		if opts.Volumes {	1483		if opts.Volumes {
1484		for _, volume := range diag.Volumes {	1484		for _, volume := range diag.Volumes {
1485		report.PreservedVolumes = append(report.PreservedVolumes, domain.CleanupVolume{Name: volume.Name, Reason: "volume purge requires explicit --volumes and runtime support"})	1485		report.PreservedVolumes = append(report.PreservedVolumes, domain.CleanupVolume{Name: volume.Name, Reason: "volume purge requires explicit --volumes and runtime support"})

@@ -1516,6 +1516,26 @@ func runRoutePurge(ctx context.Context, cp ControlPlane, domainName string, opts
1516		return report, nil	1516		return report, nil
1517		}	1517		}
1518			1518
			1519	+	func appendUniqueCleanupEntities(existing []domain.CleanupOrphanedEntity, candidates ...domain.CleanupOrphanedEntity) []domain.CleanupOrphanedEntity {
			1520	+	seen := make(map[string]struct{}, len(existing))
			1521	+	for _, entity := range existing {
			1522	+	seen[cleanupEntityKey(entity)] = struct{}{}
			1523	+	}
			1524	+	for _, entity := range candidates {
			1525	+	key := cleanupEntityKey(entity)
			1526	+	if _, ok := seen[key]; ok {
			1527	+	continue
			1528	+	}
			1529	+	existing = append(existing, entity)
			1530	+	seen[key] = struct{}{}
			1531	+	}
			1532	+	return existing
			1533	+	}
			1534	+
			1535	+	func cleanupEntityKey(entity domain.CleanupOrphanedEntity) string {
			1536	+	return entity.Kind + "\|" + entity.ID + "\|" + entity.Name
			1537	+	}
			1538	+
1519		func writeRoutePurgeText(out io.Writer, report *domain.CleanupReport, force bool) error {	1539		func writeRoutePurgeText(out io.Writer, report *domain.CleanupReport, force bool) error {
1520		mode := "Purge dry-run"	1540		mode := "Purge dry-run"
1521		if force {	1541		if force {

M internal/app/run.go +9 -9


@@ -701,8 +701,8 @@ func (si *serviceInit) registerReloadCoordinatorHooks() {
701	701		return
702	702		}
703	703
704		-	si.svc.reloadCoordinator.SetContainerConfigApplier(func(reloadCfg Config) error {
705		-	containerCfg, err := buildContainerServiceConfig(si.ctx, si.v, reloadCfg, si.svc, si.log)
	704	+	si.svc.reloadCoordinator.SetContainerConfigApplier(func(reloadCtx context.Context, reloadCfg Config) error {
	705	+	containerCfg, err := buildContainerServiceConfig(reloadCtx, si.v, reloadCfg, si.svc, si.log)
706	706		if err != nil {
707	707		return err
708	708		}

@@ -1627,7 +1627,7 @@ type reloadCoordinator struct {
1627	1627		configSvc configReloader
1628	1628		v *viper.Viper
1629	1629		proxySvc proxyConfigUpdater
1630		-	applyContainerConfig func(Config) error
	1630	+	applyContainerConfig func(context.Context, Config) error
1631	1631		registryLimits interface {
1632	1632		UpdateBlobLimits(maxBlobChunkSize, maxBlobSize int64)
1633	1633		}

@@ -1659,7 +1659,7 @@ func (c *reloadCoordinator) SetRegistryLimits(limits interface {
1659	1659		c.registryLimits = limits
1660	1660		}
1661	1661
1662		-	func (c *reloadCoordinator) SetContainerConfigApplier(apply func(Config) error) {
	1662	+	func (c *reloadCoordinator) SetContainerConfigApplier(apply func(context.Context, Config) error) {
1663	1663		c.mu.Lock()
1664	1664		defer c.mu.Unlock()
1665	1665		c.applyContainerConfig = apply

@@ -1713,16 +1713,16 @@ func (c *reloadCoordinator) applyLoadedConfig(ctx context.Context, now time.Time
1713	1713		return fmt.Errorf("failed to parse proxy config on reload: %w", err)
1714	1714		}
1715	1715
1716		-	c.proxySvc.UpdateConfig(reloadedProxy.proxyConfig)
1717		-	if c.registryLimits != nil {
1718		-	c.registryLimits.UpdateBlobLimits(reloadedProxy.maxBlobChunkSize, reloadedProxy.maxBlobSize)
1719		-	}
1720	1716		if c.applyContainerConfig != nil {
1721		-	if err := c.applyContainerConfig(reloadCfg); err != nil {
	1717	+	if err := c.applyContainerConfig(ctx, reloadCfg); err != nil {
1722	1718		c.log.Error().Err(err).Msg("failed to apply container config on reload")
1723	1719		return fmt.Errorf("failed to apply container config on reload: %w", err)
1724	1720		}
1725	1721		}
	1722	+	c.proxySvc.UpdateConfig(reloadedProxy.proxyConfig)
	1723	+	if c.registryLimits != nil {
	1724	+	c.registryLimits.UpdateBlobLimits(reloadedProxy.maxBlobChunkSize, reloadedProxy.maxBlobSize)
	1725	+	}
1726	1726
1727	1727		if c.eventBus != nil {
1728	1728		if err := c.eventBus.Publish(domain.EventConfigReload, nil); err != nil {

 		return
 	}
-	si.svc.reloadCoordinator.SetContainerConfigApplier(func(reloadCfg Config) error {
+	si.svc.reloadCoordinator.SetContainerConfigApplier(func(reloadCtx context.Context, reloadCfg Config) error {
-		containerCfg, err := buildContainerServiceConfig(si.ctx, si.v, reloadCfg, si.svc, si.log)
+		containerCfg, err := buildContainerServiceConfig(reloadCtx, si.v, reloadCfg, si.svc, si.log)
 		if err != nil {
 			return err
 		}
 	configSvc            configReloader
 	v                    *viper.Viper
 	proxySvc             proxyConfigUpdater
-	applyContainerConfig func(Config) error
+	applyContainerConfig func(context.Context, Config) error
 	registryLimits       interface {
 		UpdateBlobLimits(maxBlobChunkSize, maxBlobSize int64)
 	}
 	c.registryLimits = limits
 }
-func (c *reloadCoordinator) SetContainerConfigApplier(apply func(Config) error) {
+func (c *reloadCoordinator) SetContainerConfigApplier(apply func(context.Context, Config) error) {
 	c.mu.Lock()
 	defer c.mu.Unlock()
 	c.applyContainerConfig = apply
 		return fmt.Errorf("failed to parse proxy config on reload: %w", err)
 	}
-	c.proxySvc.UpdateConfig(reloadedProxy.proxyConfig)
-	if c.registryLimits != nil {
-		c.registryLimits.UpdateBlobLimits(reloadedProxy.maxBlobChunkSize, reloadedProxy.maxBlobSize)
-	}
 	if c.applyContainerConfig != nil {
-		if err := c.applyContainerConfig(reloadCfg); err != nil {
+		if err := c.applyContainerConfig(ctx, reloadCfg); err != nil {
 			c.log.Error().Err(err).Msg("failed to apply container config on reload")
 			return fmt.Errorf("failed to apply container config on reload: %w", err)
 		}
 	}
+	c.proxySvc.UpdateConfig(reloadedProxy.proxyConfig)
+	if c.registryLimits != nil {
+		c.registryLimits.UpdateBlobLimits(reloadedProxy.maxBlobChunkSize, reloadedProxy.maxBlobSize)
+	}
 	if c.eventBus != nil {
 		if err := c.eventBus.Publish(domain.EventConfigReload, nil); err != nil {

M internal/app/run_reload_test.go +4 -2


@@ -116,11 +116,13 @@ func (r *registryLimitsRecorder) UpdateBlobLimits(maxBlobChunkSize, maxBlobSize
116	116
117	117		type containerConfigApplyRecorder struct {
118	118		calls int
	119	+	ctx context.Context
119	120		cfg Config
120	121		}
121	122
122		-	func (r *containerConfigApplyRecorder) Apply(cfg Config) error {
	123	+	func (r *containerConfigApplyRecorder) Apply(ctx context.Context, cfg Config) error {
123	124		r.calls++
	125	+	r.ctx = ctx
124	126		r.cfg = cfg
125	127		return nil
126	128		}

@@ -256,7 +258,7 @@ func TestServiceInit_RegisterReloadCoordinatorHooks_WiresContainerConfigApplier(
256	258		reloadCfg.Server.RegistryPort = 5000
257	259		reloadCfg.Images.AllowedRegistries = []string{"docker.io"}
258	260
259		-	require.NoError(t, si.svc.reloadCoordinator.applyContainerConfig(reloadCfg))
	261	+	require.NoError(t, si.svc.reloadCoordinator.applyContainerConfig(ctx, reloadCfg))
260	262		}
261	263		func TestReloadCoordinator_DebouncesRepeatedWatchCallbacks(t *testing.T) {
262	264		ctx := context.Background()

 type containerConfigApplyRecorder struct {
 	calls int
+	ctx   context.Context
 	cfg   Config
 }
-func (r *containerConfigApplyRecorder) Apply(cfg Config) error {
+func (r *containerConfigApplyRecorder) Apply(ctx context.Context, cfg Config) error {
 	r.calls++
+	r.ctx = ctx
 	r.cfg = cfg
 	return nil
 }
 	reloadCfg.Server.RegistryPort = 5000
 	reloadCfg.Images.AllowedRegistries = []string{"docker.io"}
-	require.NoError(t, si.svc.reloadCoordinator.applyContainerConfig(reloadCfg))
+	require.NoError(t, si.svc.reloadCoordinator.applyContainerConfig(ctx, reloadCfg))
 }
 func TestReloadCoordinator_DebouncesRepeatedWatchCallbacks(t *testing.T) {
 	ctx := context.Background()

M internal/usecase/publictls/service_test.go +6 -2


@@ -663,8 +663,12 @@ func TestServiceReconcile_SerializesConcurrentRuns(t *testing.T) {
663	663		t.Fatal("first reconcile did not start certificate obtain")
664	664		}
665	665
666		-	go func() { errCh <- svc.Reconcile(ctx) }()
667		-	time.Sleep(50 * time.Millisecond)
	666	+	secondStarted := make(chan struct{})
	667	+	go func() {
	668	+	close(secondStarted)
	669	+	errCh <- svc.Reconcile(ctx)
	670	+	}()
	671	+	<-secondStarted
668	672		close(releaseObtain)
669	673
670	674		require.NoError(t, <-errCh)

 		t.Fatal("first reconcile did not start certificate obtain")
 	}
-	go func() { errCh <- svc.Reconcile(ctx) }()
+	secondStarted := make(chan struct{})
-	time.Sleep(50 * time.Millisecond)
+	go func() {
+		close(secondStarted)
+		errCh <- svc.Reconcile(ctx)
+	}()
+	<-secondStarted
 	close(releaseObtain)
 	require.NoError(t, <-errCh)