ssh到服务器使用ai2thor遇到的问题

最新推荐文章于 2025-04-09 11:50:00 发布

原创最新推荐文章于 2025-04-09 11:50:00 发布 · 942 阅读

6 ·

CC 4.0 BY-SA版权

文章标签：

#ssh #机器人

机器人仿真专栏收录该内容

2 篇文章

订阅专栏

部署运行你感兴趣的模型镜像

服务器上装了ai2thor，ssh到服务器上跑程序，程序开了多进程调用ai2thor进行仿真。
结果出现如下提示，然后一直卡住不动。

/usr/lib/xorg/Xorg.wrap: Only console users are allowed to run the X server

Ctrl+C结束程序发现所有进程都卡在fifo_server.py的open函数上

    controller = ai2thor.controller.Controller(**controller_kwargs)
  File "/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/controller.py", line 498, in __init__
    host=host,
  File "/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/controller.py", line 1299, in start
    self.last_event = self.server.receive()
  File "/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/fifo_server.py", line 182, in receive
    metadata, files = self._recv_message()
  File "/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/fifo_server.py", line 103, in _recv_message
    self.server_pipe = open(self.server_pipe_path, "rb")

搜索卡住的函数没找到解决方案，试着直接在python里面初始化一个Controller，出现如下报错。

>>> from ai2thor.controller import Controller
>>> c=Controller()
/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/platform.py:155: UserWarning: could not connect to X Display: 0, Can't connect to display ":0": b'No protocol specified\n'
  "could not connect to X Display: %s, %s" % (display_str, e)
/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/platform.py:155: UserWarning: could not connect to X Display: 1, Can't connect to display ":1": b'No protocol specified\n'
  "could not connect to X Display: %s, %s" % (display_str, e)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/controller.py", line 487, in __init__
    self._build = self.find_build(local_build, commit_id, branch, platform)
  File "/home/zzy/.conda/envs/cow/lib/python3.7/site-packages/ai2thor/controller.py", line 1214, in find_build
    raise Exception("\n".join(error_messages))
Exception: The following builds were found, but had missing dependencies. Only one valid platform is required to run AI2-THOR.
Platform Linux64 failed validation with the following errors: No valid X display found
  Linux64 requires a X11 server to be running with GLX. If you have a NVIDIA GPU, please run: sudo ai2thor-xorg start

查了说是要设置DISPLAY环境变量，echo $DISPLAY确实是没有设置的，但是试了:0，:1，:2等都一样是报错，但有额外的报错信息：
对于:0，:1已经在服务器上登录了的图形桌面（其它账号登录的）所对应的号码，是如下提示：

Platform Linux64 failed validation with the following errors: Invalid display: :0. Failed to connect Can't connect to display ":0": b'No protocol specified\n'

对于:2没有图形桌面对应的号码，是如下提示：

Platform Linux64 failed validation with the following errors: Invalid display: :2. Failed to connect Can't connect to display ":2": [Errno 111] Connection refused

但根据这些新提示也没有查到什么有用的解决方法。
想着可能需要ssh时加上x11转发有图形界面才能开启Controller，但网络问题支持不了开启转发，于是根据ai2thor官网的说法准备使用CloudRendering看看。等着下载的时候想着可能是ai2thor配置不正确，决定到机房服务器上直接登录再试，发现的确是可以运行的，同时也查看了一下$DISPLAY变量，发现是:2。
在不登出桌面环境的情况下，再通过ssh连接到服务器并且设置DISPLAY为:2发现能够初始化Controller了，桌面环境上看的话会出现ai2thor的仿真环境。
PS1：后面试了如果不设置DISPLAY环境变量或者传参x_display的话，初始化的时候会自动遍历所有现有的display，如果有能运行的就会自动选取进行初始化，没有就会报错。
PS2：设置DISPLAY为空字符串的话也是会卡在open函数上面。
PS3：开多进程跑ai2thor的时候如果初始化出错的话好像并不会导致进程崩溃，而是一直卡在open函数上面没有任何提示。
因此，如果要ssh到服务器上面跑ai2thor的话可以尝试一下先在服务器开一个图形界面，查一下界面对应的display（echo $DISPLAY），然后ssh登录后设置DISPLAY为对应的值由或者初始化Controller时设置传参x_display=“:2” #相应的display值
又或者可以试试x11转发或者CloudRendering