Thanks!
I have a potential addition to my config file. I had to subclass the
LsfSpawner class because it appears incomplete and doesn't have all of the
options required for our cluster. Does this look right?
c.JupyterHub.spawner_class = 'wrapspawner.ProfilesSpawner'
c.Spawner.http_timeout = 120
#------------------------------------------------------------------------------
# BatchSpawnerBase configuration
# Providing default values that we may omit in the profiles
#------------------------------------------------------------------------------
c.BatchSpawnerBase.req_runtime = '12:00'
#------------------------------------------------------------------------------
# ProfilesSpawner configuration
#------------------------------------------------------------------------------
# List of profiles to offer for selection. Signature is:
# List(Tuple( Unicode, Unicode, Type(Spawner), Dict ))
# corresponding to profile display name, unique key, Spawner class,
# dictionary of spawner config options.
#
# The first three values will be exposed in the input_template as {display},
# {key}, and {type}
#
## Override LsfSpawner to be compatible with Minerva
class MinervaSpawner(batchspawner.LsfSpawner):
batch_script = Unicode('''#!/bin/sh
#BSUB -R "span[hosts=1]" # Only spawn job on one server
#BSUB -P {account}
#BSUB -q {queue}
#BSUB -m {partition}
#BSUB -R rusage[mem={memory}]
#BSUB -n {nprocs}
#BSUB -J spawner-jupyterhub
#BSUB -o {homedir}/.jupyterhub.lsf.out
#BSUB -e {homedir}/.jupyterhub.lsf.err''')
req_partition = Unicode('', \
help="Partition name to submit job to resource manager"
).tag(config=True)
req_account = Unicode('', \
help="Account name to submit job to resource manager"
).tag(config=True)
c.ProfilesSpawner.profiles = [
( "Local server", 'local', 'jupyterhub.spawner.SudoSpawner', {'ip':
'0.0.0.0'} ),
('Bode - 2 cores, 4 GB, 8 hours', 'bode2c4g12h', 'MinervaSpawner',
dict(req_nprocs='2', req_queue='premium', req_runtime='8:00', req_memory
='4000',
req_partition='bode', req_account='acc_LOAD')),
('Manda - 8 cores, 128 GB, 4 hours', 'manda128gb', 'MinervaSpawner',
dict(req_nprocs='8', req_queue='premium', req_runtime='4:00', req_memory
='16000',
req_partition='bode', req_account='acc_LOAD')),
('Bode - 2 cores, 4 GB, 24 hours', 'mesabi2c4gb24h',
'batchspawner.TorqueSpawner',
dict(req_nprocs='2', req_queue='premium', req_runtime='24:00',
req_memory='4gb',
req_partition='bode', req_account='acc_LOAD')),
]
On Thursday, March 15, 2018 at 6:47:28 PM UTC-4, Michael Milligan wrote:
>
> Hi Brian -
>
> It is possible, and it sounds like you already mostly have the solution!
> The example config at the end of the batchspawner README should be roughly
> what you want, except that you will use the SudoSpawner instead of
> LocalProcessSpawner. Specifically, you need to install all of the spawners
> you are using alongside Jupyterhub (e.g. pip install wrapspawner,
> batchspawner, and sudospawner into the same virtualenv as jupyterhub) --
> the ProfilesSpawner configuration will tell Jupyterhub which spawner to use
> in response to the user's menu selection.
>
> Sorry about the confusing state of the documentation about how to do this.
> Batchspawner and wrapspawner were originally distributed in the same
> repository, and when we split them apart it sounds like we may not have
> included enough context on one side or the other. We're happy to get
> suggestions about how to improve the docs if you'd like to give us some
> feedback when you get a chance.
>
> Regards,
> Michael
>
> On Thu, Mar 15, 2018 at 4:58 PM, Brian Fulton-Howard <[email protected]
> <javascript:>> wrote:
>
>> As I understand it, I need to use some combination of wrapspawner,
>> localspawner, sudospawner and batchspawner. I want to run jupyterhub as an
>> unprivileged user, but the admins are happy to edit the sudoers file to
>> make it work. What I'd like to do is give users in my lab a choice to start
>> on the LSF cluster (batchspawner) or locally (localspawner). Wrapspawner
>> would provide the interface to make a selection.
>>
>> Is this possible, and how would I go about doing it? I looked in the
>> documentation for the various spawners and didn't find any detailed
>> information.
>>
>> The cluster supports PAM, so I don't need any special authenticator.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Project Jupyter" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected] <javascript:>.
>> To post to this group, send email to [email protected]
>> <javascript:>.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/jupyter/a3f21888-0495-4fe0-b598-ca7e226bd61f%40googlegroups.com
>>
>> <https://groups.google.com/d/msgid/jupyter/a3f21888-0495-4fe0-b598-ca7e226bd61f%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>
>
--
You received this message because you are subscribed to the Google Groups
"Project Jupyter" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/jupyter/1f26020c-41cf-4dc4-84d7-41a879c68bea%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
