Intended for folks in genome sciences at UW.
This script should only be run on the remote host.
It can be used to submit a JupyterLab job to the grid engine using the specified
parameters.
The name of this job is determined from '-N', -l', '-pe', and '-P' options
(see run_jupyter -help for details).
If a job with this same name already exists, the script merely prints
information for the preexisting job - no new job will be submitted.
If you typically use the same options for your jobs and don't want to have to
type them out each time you use this script, modify
~/.run_jupyter.qsub_option_defaults on the remote host.
This script should only be run on your local machine.
It submits a JupyterLab job to the grid engine on the remote host, then
activates JupyterLab by sshing to the remote host at the JupyterLab port and
grid engine node.
This script calls run_jupyter on the remote host - most command line options
for ssh_jupyter are forwarded to run_jupyter.
Put run_jupyter on your remote host and ssh_jupyter on your local machine.
Add both to the PATH environment variable in your shell configuration files.
For ssh_jupyter, I recommend setting ~/.ssh/config on your local machine
such that you don't need to type your password every time you ssh to the remote
host.
If you have to connect to a gateway (ie nexus) to reach the head node (ie
grid), you'll probably want to set up local port forwarding as well.
Most qsub options are supported.
See run_jupyter -help and ssh_jupyter -help for full usage information.
Feel free to reach out if you run into any problems! Use at your own risk, ect ect.