apache/beam · error

unable to provide %v to worker

Error message

unable to provide %v to worker

What it means

GetArtifact serves staged artifacts to the worker from the server's in-memory artifacts map, keyed by the artifact's type payload. If the requested artifact's key is not present, the server logs a warning and returns 'unable to provide <prototext> to worker'. This means the prism server never staged (or already discarded) the artifact the SDK harness asked for.

Source

Thrown at sdks/go/pkg/beam/runners/prism/internal/jobservices/artifact.go:107

			s.artifacts[string(dep.GetTypePayload())] = buf.Bytes()
		}
	}
	return nil
}

func (s *Server) ResolveArtifacts(_ context.Context, req *jobpb.ResolveArtifactsRequest) (*jobpb.ResolveArtifactsResponse, error) {
	return &jobpb.ResolveArtifactsResponse{
		Replacements: req.GetArtifacts(),
	}, nil
}

func (s *Server) GetArtifact(req *jobpb.GetArtifactRequest, stream jobpb.ArtifactRetrievalService_GetArtifactServer) error {
	info := req.GetArtifact()
	buf, ok := s.artifacts[string(info.GetTypePayload())]
	if !ok {
		pt := prototext.Format(info)
		slog.Warn("unable to provide artifact to worker", "artifact_info", pt)
		return fmt.Errorf("unable to provide %v to worker", pt)
	}
	chunk := 128 * 1024 * 1024 // 128 MB
	var i int
	for i+chunk < len(buf) {
		stream.Send(&jobpb.GetArtifactResponse{
			Data: buf[i : i+chunk],
		})
		i += chunk
	}
	stream.Send(&jobpb.GetArtifactResponse{
		Data: buf[i:],
	})
	return nil
}

View on GitHub (pinned to 12126d8942)

Solutions

1. Check the warning log line 'unable to provide artifact to worker' for the exact artifact_info requested and compare it to what was staged.
2. Ensure the artifact staging step ran before workers request dependencies (e.g. correct --artifact/staging flags).
3. Verify prism server and harness agree on the artifact type payload encoding; upgrade both to the same version.

Example fix

// before: artifact never staged
// ensure staging before running:
jobStager.Stage(artifacts...)

// after: server has s.artifacts[typePayload] populated, GetArtifact succeeds

// Pre-check artifact availability before requesting from the worker
if _, ok := stagedArtifacts[string(info.GetTypePayload())]; !ok {
    return fmt.Errorf("artifact with type payload %q was never staged", info.GetTypePayload())
}

err := stream.Recv() // or the client-side equivalent
if err != nil && strings.Contains(err.Error(), "unable to provide") {
    log.Fatalf("artifact missing from prism server staging; re-stage artifacts: %v", err)
}

When it happens

Trigger: A worker requests an artifact whose type payload has no entry in s.artifacts — e.g. the artifact was never registered via the staging path, or the request's type payload doesn't match what was stored.

Common situations: Reverse-artifact retrieval setups where the harness computes a different artifact type payload than the server stored; jobs restarted against a fresh prism server that lost staged artifacts; custom artifact staging omitted.

Understand the failure class

Background: 'Could not be found', 'does not exist', 'not found in database': the resource-not-found family when an ID, slug, key, or URI lookup comes back empty — this error's family across 20 libraries.

Related errors

• unexpected ResolveArtifactResponse to GetArtifact: %v
• Unknown context parameter: ${param.parDoParamName}
• Unsupported window mapping fn: ${sideInput.windowMappingFn.u
• Unimplemented access pattern: ${accessPattern}
• State stream is closed.

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/34f0a533ad72941b. Report an issue: GitHub.