我正在尝试在我的 HPC 上使用 clustershell,并且我想使用 munge 作为通信基础,因为它已经用于 SLURM。
我按照官网的配置,据说只需要添加一个defaults.conf文件
[task.default]
distant_workername: mrsh
但后来我有一个 python 错误说
Traceback (most recent call last):
File "/usr/lib/python3/dist-packages/ClusterShell/CLI/Clush.py", line 765, in clush_excepthook
raise exp
File "/usr/bin/clush", line 11, in <module>
load_entry_point('ClusterShell==1.8.3', 'console_scripts', 'clush')()
File "/usr/lib/python3/dist-packages/ClusterShell/CLI/Clush.py", line 879, in main
task = task_self()
File "/usr/lib/python3/dist-packages/ClusterShell/Task.py", line 1373, in task_self
return Task(thread=threading.currentThread(), defaults=defaults)
File "/usr/lib/python3/dist-packages/ClusterShell/Task.py", line 286, in __init__
"distant_worker": _distant_workerclass(defaults)})
File "/usr/lib/python3/dist-packages/ClusterShell/Defaults.py", line 76, in _distant_workerclass
return _load_workerclass(defaults.distant_workername)
File "/usr/lib/python3/dist-packages/ClusterShell/Defaults.py", line 65, in _load_workerclass
__import__(modname)
ModuleNotFoundError: No module named 'ClusterShell.Worker.Mrsh'
我使用 pip 安装了 python 包 mrsh 但它没有帮助。
所有节点具有相同的配置
我不知道如何将 munge (如果可能的话)用于 clustershell