typedef struct TF_Server TF_Server;

// Creates a new in-process TensorFlow server configured using a serialized
// ServerDef protocol buffer provided via `proto` and `proto_len`.
//
// The server will not serve any requests until TF_ServerStart is invoked.
// The server will stop serving requests once TF_ServerStop or
// TF_DeleteServer is invoked.
TF_CAPI_EXPORT extern TF_Server* TF_NewServer(const void* proto,

    TFE_Context* ctx, int keep_alive_secs, const void* proto, size_t proto_len,
    int64_t init_timeout_in_ms, TF_Status* status,
    bool clear_existing_contexts);

// Set server def with retries and timeout. This is helpful for fault-tolerant
// initial connection in high-preemption environments, such as
// ParameterServerStrategy training.

TFE_Context* CreateContext(const std::string& serialized_server_def,
                           bool isolate_session_state,
                           int64_t init_timeout_in_ms);

tensorflow::ServerDef ReplaceTaskInServerDef(
    const tensorflow::ServerDef& server_def, int task_index);

void ReplaceTaskInServerDef(tensorflow::ServerDef* server_def, int task_index,

  // on cluster configurations provided in `server_def`; otherwise, update
  // existing context state with the provided `server_def`. Contexts created
  // on remote tasks will be considered stale and garbage collected after
  // `keep_alive_secs` of inactivity.
  virtual Status SetOrUpdateServerDef(const ServerDef& server_def,
                                      bool reset_context, int keep_alive_secs,

  // all tasks in the cluster.
  // This call internally coordinates with other tasks to initialize the eager
  // context and TF server for multi-client execution.
  virtual Status EnableCollectiveOps(const ServerDef& server_def) = 0;

  // Set a distributed manager that helps set up, update, and check liveness
  // of member tasks in the cluster.
  virtual void SetDistributedManager(