A more comprehensive user guide can be found here - https://arc.umich.edu/open-ondemand/

• When clicking on the 'Files', 'jobs', or 'Cluster' drop-down menu choices, those choices will open a new tab rather than opening in the current window.
• In the window for Job Composer, editing the script for a job that has been highlighted in blue and resubmitting that job for a new job id number will make OnDemand place the output in the same folder as all previous output for that specific job. All of the output for that specific job title will be shown on the right in the 'Folder Contents' section in the 'Job Details' box:

• After using the Job Composer feature, if you create a new job and have not specified a specific folder for the output, OnDemand will create a new folder for that job which will  contain both the output of the job and a copy of the original script. All the output will be in your My Jobs folder:

• From the Dashboard, click on the "Jobs" drop-down menu option, and select "Job Composer"
• To create a new job, select the "+ New Job" drop-down menu option, and select "From Default Template"
  • To change your job options, select the "Job Options" button
  • From there, you can change the job name, cluster, job script, and account.
• You can edit your "Submit Script" by clicking on the "Open Editor" button (bottom right hand side) and it will open a new tab where you can edit the bash script.
  • After you've made your edits, click the "Save" button (top left hand side) and it will save the bash script.
  • Close out of the Editor tab and return to Job Composer
• When your job is ready for submission, click the green "Submit" button.
• Your job will submit and your "Status" will change to "Running"
  • Status is updated in real-time, so when it completes, you will not need to refresh the page.
• Your job will spin up an output file (slurm-<job_id>.out) in your Folder Contents. Clicking on the output file will let you view your job's output

• From the Dashboard, click on the "Jobs" drop-down menu option, and select "Job Composer"
• To create a new job, select the "+ New Job" drop-down menu option, and select "From Specified Path"
• It will redirect you to a page to input your job settings
  • Source path → enter the path to a directory on the file system
    • If there is an invalid file or directory, you will receive "Error with status pid 1868 exit 1 occurred when trying to determind directory size: du: cannot access <`specified_path`>: No such file or directory
    • It will not create the missing directories or files
  • Job Attributes → this is optional
    • Name → job name
    • Script Name → name of your bash script (from the specified path location)
    • Cluster → cluster of your choice
    • Account → account of your choice
• When you have the job settings you want, click the "Save" button at the bottom of the page to continue
• If you need to make adjustments:
  • To change your job options, select the "Job Options" button
  • From there, you can change the job name, cluster, job script, and account.
  • Instead of editing your bash script from the command line, you can directly edit your "Submit Script" by clicking on the "Open Editor" button (bottom right hand side), and it will open a new tab where you can edit the bash script
    • After you've made your edits, click the "Save" button (top left hand side) and it will save the bash script.
    • Close out of the Editor tab and return to Job Composer
• When your job is ready for submission, click the green "Submit" button.
• Your job will submit and your "Status" will change to "Running"
  • Status is updated in real-time, so when it completes, you will not need to refresh the page.
• Your job will spin up an output file (slurm-<job_id>.out) in your Folder Contents. Clicking on the output file will let you view your job's output

• From the Dashboard, click on the "Jobs" drop-down menu option, and select "Job Composer"
• To create a new job, select the "+ New Job" drop-down menu option, and select "From Template"
  • It will redirect you to your "Templates" page.
• Select the job template you would like to use.
• On the right-hand side of the page, enter in:
  • Job name
  • Cluster
  • Script Name → this is greyed out as it uses the script from the template
• When you've entered in the available options, click the "Create New Job" button
  • It will redirect you back to Job Composer
• If you need to make adjustments:
  • To change your job options, select the "Job Options" button
  • From there, you can change the job name, cluster, job script, and account.
  • Instead of editing your bash script from the command line, you can directly edit your "Submit Script" by clicking on the "Open Editor" button (bottom right hand side), and it will open a new tab where you can edit the bash script
    • After you've made your edits, click the "Save" button (top left hand side) and it will save the bash script.
    • Close out of the Editor tab and return to Job Composer
• When your job is ready for submission, click the green "Submit" button.
• Your job will submit and your "Status" will change to "Running"
  • Status is updated in real-time, so when it completes, you will not need to refresh the page.
• Your job will spin up an output file (slurm-<job_id>.out) in your Folder Contents. Clicking on the output file will let you view your job's output

• After you submit a job in Job Composer, the "Submit" button will become greyed out and the "Stop" button will become un-greyed.
• All you need to do is click on the "Stop" button and it will stop your job.
  • Just a quick note, it will show your job as "completed"

• While in Job Composer, if you need to delete a job, select the job you wish to delete.
• Then, click the red "Delete" button.
• A popup will ask you if you are sure you want to delete the job.
  • Ok to continue
  • Cancel to cancel
• Your job will then be deleted

• From the Dashboard, click on the "Files" drop-down menu, and select "Home Directory"
  • It will open a new tab with your /home directory
• While in Job Composer, select the "Edit Files" button
  • It will open a new tab with your /home directory and path you directly to where your files are being stored for your specific job

• Select the job you would like to create a template for, then select the "Create Template" button
  • It will create a new tab where you can enter in the Template options
    • Path → the template will be created by copying file from this source path.
    • Name → name of the job template
    • Cluster → cluster the template will be for
    • Notes → (optional) input any notes you might want. You can change these notes later by editing the manifest.yml in this templates's directory.
  • When your template is ready to be created, hit the "Save" button
  • It will redirect you to your "Templates" page where you can see your newly created job template.

1. Load both modules you will be using from Open OnDemand module load python3.7-anaconda/2019.07 julia/1.1.1
2. Start julia and enter the following
   1. julia> using Pkg
   2. julia> Pkg.add("IJulia")
   3. ...\
   4. julia> exit()
3. Start a Jupyter session from Open OnDemand
4. From inside jupyter under the "New" menu you should now see an option for "Julia 1.1.1"
5. Create a new notebook with that kernel and test
6. Following the process above will install the julia kernel to: ~/.local/share/jupyter/kernels/julia-1.1/
   1. This location is searched by Jupyter by default. If you ever need to reset or reinstall, then this is the directory you would remove.