[RL] fix skip_server_warmup and rl health_generate logic (#8757)

python/sglang/srt/entrypoints/http_server.py

@@ -1172,6 +1172,8 @@ def _wait_and_warmup(
             pipe_finish_writer,
         ):
             return
+    else:
+        _global_state.tokenizer_manager.server_status = ServerStatus.Up
 
     logger.info("The server is fired up and ready to roll!")