-
-
Notifications
You must be signed in to change notification settings - Fork 994
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
PBS job templates not picked up? #37
Comments
The unstable ordering is a bug that will be fixed in traitlets 4.1. |
Okay, so I was using the wrong configuration item. But the template file name needs to be a full path to a file, its seems. As otherwise it won't be found. I.e. with
and both template files located in
|
And if you specify them as full paths? |
Well, yes, that's what I wrote. I need to specify their full paths, and then they are found. |
Then this issue can be closed? |
Well, the documentation might still need a remark added, as it doesn't show full paths to the template files. |
Hi! I’m going through and cleaning up old/stale issues on this repo. Sorry for leaving it forever. I'm hoping to bring this repo back to a healthier state. Template paths are evaluated relative to the cwd, like most file path arguments (e.g. |
I'm using ipyparallel 4.0.2 and friends (with Python 3.4) installed in a virtualenv with pip. I'm trying to get a set of engines+controller running on a system with the Torque batch system, i.e. PBS. I've followed the documentation at https://ipyparallel.readthedocs.org/en/stable/process.html#using-ipcluster-in-pbs-mode to get the configuration set up.
When I launch with "ipcluster start --profile=1engine-per-node -n 2" my job templates are not being used (the actual job name is different from what I specify in the templates with -N and the walltime is incorrect).
The steps I've done:
a) ipcluster_config.py:
c.IPClusterEngines.engine_launcher_class = 'PBS'
c.IPClusterStart.controller_launcher_class = 'PBS'
c.IPClusterEngines.n = 4
c.PBSControllerLauncher.batch_file_name = 'controller.template'
c.PBSEngineSetLauncher.batch_file_name = 'engine.template'
b) ipcontroller_config.py:
c.HubFactory.ip = '*'
a) .ipython/profile_1engine-per-node/controller.template
#PBS -N ipyparallel-controller
#PBS -j oe
#PBS -l walltime=01:00:00
#PBS -l nodes=1:ppn=4
cd $PBS_O_WORKDIR
source $HOME/pyenv/3.4/bin/activate
ipcontroller --profile-dir={profile_dir}
b) .ipython/profile_1engine-per-node/engine.template
PBS -N ipyparallel-engine
PBS -j oe
PBS -l walltime=01:00:00
PBS -l nodes={n}:ppn=1
cd $PBS_O_WORKDIR
source $HOME/pyenv/3.4/bin/activate
module load openmpi/gnu/1.6.5
which mpiexec
-n {n} ipengine --profile-dir={profile_dir}Btw, are the profiles that get generated with "profile create" written based on inspecting Python classes or something? I generated a new dummy parallel profile so I could compare what I had changed in my 1engine-per-node profile, but a diff show vastly different order of the config items, making direct comparison hard.
The text was updated successfully, but these errors were encountered: