Jupyterhub: Fractional memory / CPU limits / guarantees fail

[yuvipanda]

How to reproduce the issue

Set memory limit (or guarantee, or cpu limit / guarantee) to a non-integral spec:

c.Spawner.mem_limit = "1.5G"

What you expected to happen

(In supported spawners) memory limit is set to 1.5 gigabytes of RAM

What actually happens

JupyterHub refuses to start, with:

[E 2017-04-18 05:39:02.270 JupyterHub app:1527]
    Traceback (most recent call last):
      File "/usr/local/lib/python3.4/dist-packages/jupyterhub/app.py", line 1524, in launch_instance_async
        yield self.initialize(argv)
      File "/usr/local/lib/python3.4/dist-packages/jupyterhub/app.py", line 1315, in initialize
        yield self.init_spawners()
      File "/usr/local/lib/python3.4/dist-packages/jupyterhub/app.py", line 1084, in init_spawners
        self.users[orm_user.id] = user = User(orm_user, self.tornado_settings)
      File "/usr/local/lib/python3.4/dist-packages/jupyterhub/user.py", line 128, in __init__
        config=self.settings.get('config'),
      File "/usr/local/lib/python3.4/dist-packages/kubespawner/spawner.py", line 29, in __init__
        super().__init__(*args, **kwargs)
      File "/usr/local/lib/python3.4/dist-packages/jupyterhub/spawner.py", line 345, in __init__
        super(Spawner, self).__init__(**kwargs)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/config/configurable.py", line 84, in __init__
        self.config = config
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 585, in __set__
        self.set(obj, value)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 574, in set
        obj._notify_trait(self.name, old_value, new_value)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 1139, in _notify_trait
        type='change',
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 1176, in notify_change
        c(change)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 819, in compatible_observer
        return func(self, change)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/config/configurable.py", line 186, in _config_changed
        self._load_config(change.new, traits=traits, section_names=section_names)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/config/configurable.py", line 153, in _load_config
        setattr(self, name, deepcopy(config_value))
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 585, in __set__
        self.set(obj, value)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 559, in set
        new_value = self._validate(obj, value)
      File "/usr/local/lib/python3.4/dist-packages/traitlets/traitlets.py", line 591, in _validate
        value = self.validate(obj, value)
      File "/usr/local/lib/python3.4/dist-packages/jupyterhub/traitlets.py", line 71, in validate
        return int(num) * ByteSpecification.UNIT_SUFFIXES[suffix]
    ValueError: invalid literal for int() with base 10: '1.5'

Share what version of JupyterHub you are using

0.72.

[willingc]

@yuvipanda @minrk Has recent work closed this?

[minrk]

I think this bug still exists. Should be an easy enough fix to use float instead of int here, and then cast to int after multiplying.

[yuvipanda]

I'm going to try to fix this for 0.8.