pulumi/pulumi · critical

failed to start loader server: %w

Error message

failed to start loader server: %w

What it means

The loader server is a gRPC server exposing the schema loader and provider mapper to the language host. `plugin.NewServer` failure is wrapped as `failed to start loader server` in `evalSource.Iterate`. This means the engine could not start the local gRPC service the program uses to fetch provider schemas and mapping info.

Source

Thrown at pkg/resource/deploy/source_eval.go:263

	)
	if err != nil {
		return nil, fmt.Errorf("failed to start resource monitor: %w", err)
	}

	// Also start up a schema loader and a provider mapper for the language runtime to use to fetch
	// schema and mapping information.
	loaderRegistration := schema.LoaderRegistration(
		schema.NewLoaderServer(schema.NewPluginLoader(src.plugctx)))

	mapper, err := newRunMapper(ctx, src.plugctx)
	if err != nil {
		return nil, fmt.Errorf("failed to create provider mapper: %w", err)
	}
	mapperRegistration := pconvert.MapperRegistration(pconvert.NewMapperServer(mapper))

	loaderServer, err := plugin.NewServer(src.plugctx, loaderRegistration, mapperRegistration)
	if err != nil {
		return nil, fmt.Errorf("failed to start loader server: %w", err)
	}

	// Create a new iterator with appropriate channels, and gear up to go!
	iter := &evalSourceIterator{
		loaderServer:    loaderServer,
		mon:             mon,
		src:             src,
		regChan:         regChan,
		regOutChan:      regOutChan,
		regReadChan:     regReadChan,
		finChan:         finChan,
		programComplete: programComplete,
		panicErrs:       src.panicErrs,
	}

	// Now invoke Run in a goroutine.  All subsequent resource creation events will come in over the gRPC channel,
	// and we will pump them through the channel.  If the Run call ultimately fails, we need to propagate the error.
	iter.forkRun(config, configSecretKeys)

View on GitHub (pinned to 793f7b2e16)

Solutions

1. Re-run with `--logtostderr -v=9` for the wrapped root cause
2. Reinstall plugins (`pulumi plugin rm --all`, then retry) and/or reinstall the CLI
3. Check that the environment allows binding local sockets/ports and has free resources
4. Retry the deployment; transient startup failures may clear

Example fix

// before
pulumi up
// after
pulumi plugin rm --all && pulumi up

// shell: confirm environment permits local listeners
pulumi preview --logtostderr -v=3 --dry-run > /dev/null 2>&1 || echo 'engine startup problem'

// TS automation API
try {
  await stack.up();
} catch (err) {
  if (String(err).includes("failed to start loader server")) {
    // transient local gRPC startup failure: retry once
    return stack.up();
  }
  throw err;
}

When it happens

Trigger: `pulumi up`/`preview` when `plugin.NewServer(src.plugctx, loaderRegistration, mapperRegistration)` fails — usually gRPC server/socket startup failure within the plugin context.

Common situations: Resource exhaustion or restricted networking preventing local listeners; corrupted plugin infrastructure; unusual sandboxed CI environments.

Related errors

• load %q: %w
• load provider: %w
• could not run plugin at %q: %w
• getting schema: %w
• could not execute plugin %s (%s): %w

AI-assisted analysis of pulumi/pulumi@793f7b2e16 (2026-08-31). Data as JSON: /api/errors/1b2193332ce26b96. Report an issue: GitHub.