Fix ClusterShell NodeSet import

[faganihajizada]

Summary

The current code uses import ClusterShell followed by ClusterShell.NodeSet.NodeSet(), which causes an AttributeError because the ClusterShell module doesn't expose NodeSet as a direct attribute when imported this way.
This bug prevents users from viewing jobs filtered by a specific node in the web UI.

examples: https://github.com/cea-hpc/clustershell/blob/master/doc/examples/check_nodes.py

Solution

Changed to import NodeSet directly:

from ClusterShell.NodeSet import NodeSet

This fix has been applied to:

• Main code (slurmweb/slurmrestd/init.py)
• Test files (slurmweb/tests/slurmrestd/test_slurmrestd.py, slurmweb/tests/views/test_agent.py)

Test

• The import works correctly with ClusterShell
• Only NodeSet functionality is used in the codebase
• No other ClusterShell modules are required

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[faganihajizada]

I have read the CLA Document and I hereby sign the CLA

[faganihajizada]

recheck

[rezib]

In what environment did you get this attribute error? Can you please provide the exception backtrace?

[faganihajizada]

Environment:

I see this behavior when RacksDB is disabled and patched RacksDB imports to make it optional.

Slurm-web v6.0.0
Python 3.11
ClusterShell 1.9.3

How to Reproduce:

1. Access the Slurm-web UI
2. Navigate to the "Jobs" page
3. Click on a node name where a job is running
4. This triggers a request to /v6.0.0/jobs?node=<node_name>

Trace:

Thread-80 (process_request_thread): [INFO] 10.0.147.123 - - [10/Dec/2025 12:48:38] "GET /v6.0.0/nodes HTTP/1.1" 200 -
Thread-81 (process_request_thread): [INFO] 10.0.147.123 - - [10/Dec/2025 12:48:42] "GET /v6.0.0/node/awsgpu-0 HTTP/1.1" 200 -
Thread-83 (process_request_thread): [ERROR] Exception on /v6.0.0/jobs [GET]
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/site-packages/flask/app.py", line 1511, in wsgi_app
    response = self.full_dispatch_request()
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/flask/app.py", line 919, in full_dispatch_request
    rv = self.handle_user_exception(e)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/flask/app.py", line 917, in full_dispatch_request
    rv = self.dispatch_request()
         ^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/flask/app.py", line 902, in dispatch_request
    return self.ensure_sync(self.view_functions[rule.endpoint])(**view_args)  # type: ignore[no-any-return]
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/rfl/web/tokens.py", line 102, in wrapped
    return view(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/slurmweb/views/agent.py", line 162, in jobs
    return jsonify(slurmrest("jobs_by_node", node))
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/slurmweb/views/agent.py", line 73, in wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/slurmweb/views/agent.py", line 123, in slurmrest
    return getattr(current_app.slurmrestd, method)(*args)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/slurmweb/slurmrestd/__init__.py", line 234, in jobs_by_node
    return [job for job in self.jobs() if on_node(job) and not terminated(job)]
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/slurmweb/slurmrestd/__init__.py", line 234, in <listcomp>
    return [job for job in self.jobs() if on_node(job) and not terminated(job)]
                                          ^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/slurmweb/slurmrestd/__init__.py", line 225, in on_node
    return node in ClusterShell.NodeSet.NodeSet(job["nodes"])
                   ^^^^^^^^^^^^^^^^^^^^
AttributeError: module 'ClusterShell' has no attribute 'NodeSet'

[rezib]

OK, I see!

I wrote a simple reproducer:

$ cat main.py 
import ClusterShell
print(ClusterShell)
print(ClusterShell.NodeSet)
print(ClusterShell.NodeSet.NodeSet)
$ python3 main.py 
<module 'ClusterShell' from '/…/.venv/lib/python3.12/site-packages/ClusterShell/__init__.py'>
Traceback (most recent call last):
  File "/…/main.py", line 3, in <module>
    print(ClusterShell.NodeSet)
          ^^^^^^^^^^^^^^^^^^^^
AttributeError: module 'ClusterShell' has no attribute 'NodeSet'
$ cat pre.py 
from ClusterShell.NodeSet import NodeSet
$ sed -i '1s/^/import pre\n/' main.py
$ python3 main.py 
<module 'ClusterShell' from '/…/.venv/lib/python3.12/site-packages/ClusterShell/__init__.py'>
<module 'ClusterShell.NodeSet' from '/…/.venv/lib/python3.12/site-packages/ClusterShell/NodeSet.py'>
<class 'ClusterShell.NodeSet.NodeSet'>

Indeed, the bug is hidden by RacksDB which imports NodeSet correctly. That explains why I never met the bug in CI and in supported installations.

Thank you @faganihajizada for reporting!

[faganihajizada]

Thanks @rezib

If you are fine, I am planning to submit another PR to update RacksDB imports to make it optional.

[rezib]

If you are fine, I am planning to submit another PR to update RacksDB imports to make it optional.

I'm fine with it!