caddyserver/caddy · error

creating storage value: %v

Error message

creating storage value: %v

What it means

After the storage module is loaded, its StorageConverter.CertMagicStorage() method constructs the underlying certmagic.Storage value. If that constructor fails (e.g. cannot reach the backend, cannot open a database), the error is wrapped as 'creating storage value'. This is distinct from 'loading storage module': the module was found and provisioned, but instantiating the backend connection failed.

Source

Thrown at caddy.go:548

	}

	// create the new filesystem map
	newCfg.fileSystems = &filesystems.FileSystemMap{}

	// prepare the new config for use
	newCfg.apps = make(map[string]App)
	newCfg.failedApps = make(map[string]error)

	// set up global storage and make it CertMagic's default storage, too
	err = func() error {
		if newCfg.StorageRaw != nil {
			val, err := ctx.LoadModule(newCfg, "StorageRaw")
			if err != nil {
				return fmt.Errorf("loading storage module: %v", err)
			}
			stor, err := val.(StorageConverter).CertMagicStorage()
			if err != nil {
				return fmt.Errorf("creating storage value: %v", err)
			}
			newCfg.storage = stor
		}

		if newCfg.storage == nil {
			newCfg.storage = DefaultStorage
		}
		certmagic.Default.Storage = newCfg.storage

		return nil
	}()
	if err != nil {
		return ctx, err
	}

	// start the admin endpoint (and stop any prior one)
	if replaceAdminServer {
		err = replaceLocalAdminServer(newCfg, ctx)

View on GitHub (pinned to 50e54ee279)

Solutions

1. Verify the storage backend is reachable from Caddy's network namespace (curl/nc the endpoint).
2. Fix credentials/endpoint in the storage module config per the plugin's docs.
3. Bring the backend up, then reload the config; the failure is at connect time, so retry after fixing.
4. If HA is not needed, drop the storage block to use the default local storage while debugging.

Example fix

# before: backend down
caddy run --config Caddyfile  # creating storage value: dial tcp ...:6379

# after: start backend first
redis-server --daemonize yes && caddy run --config Caddyfile

// Health-check the storage backend before Caddy loads the config.
conn, err := net.DialTimeout("tcp", storageEndpoint, 2*time.Second)
if err != nil {
    return fmt.Errorf("storage backend unreachable, load would fail: %w", err)
}
conn.Close()

for attempt := 1; attempt <= 3; attempt++ {
    err := startCaddy(cfg)
    if err == nil || !strings.Contains(err.Error(), "creating storage value") {
        break
    }
    time.Sleep(time.Duration(attempt) * 5 * time.Second) // backend may still be booting
}

When it happens

Trigger: A storage plugin whose CertMagicStorage() dials its backend eagerly — Redis/etcd/Postgres plugins failing to connect, S3 plugins failing credential checks — while Caddy loads the config. Any admin config load or startup carrying that storage block reproduces it.

Common situations: Backend service down or firewalled when Caddy starts; wrong credentials/secrets; DNS for the storage endpoint failing in the container; TLS CA mismatch between Caddy and the storage backend.

Related errors

• loading storage module: %v
• creating storage configuration: %v
• WebSocket connections aren't allowed.
• Disabling same-origin restrictions is not allowed.
• Buggy browser is sending null Origin header.

AI-assisted analysis of caddyserver/caddy@50e54ee279 (2026-08-15). Data as JSON: /api/errors/1dcff8618bef4a2e. Report an issue: GitHub.