Set Up NYU RCN: Dan's Notes

Overview

• Make sure you use your NYU account, not your Stern account.
• NYU recommends Google Chrome.
• I am setting up a simple Python setup. My notes reflect that.
• Reach out to the support group. The link to it in MANUAL.HTML does not work. Reach it directly by clicking below.
  • Link to the support group
  • Link to FAQ
• My class sites
  • Admin: FSA using Python
  • Admin: Programming
  • Admin: Tandon
  • Student: FSA using Python
  • Student: Programming
  • Student: Tandon

Issues common to all steps

• Login to the instructor site. For example, login to the RCN Instructor Site for the class. For example, https://techub-23003-instructor.rcnyu.org/
• All setup and help files are in the tuning folder.
• Stopping and restarting the server does not apply the settings immediately. It can take anywhere between 3-10 minutes. Therefore, do these settings changes before class.
• Useful files
  • MANUAL.HTML: Instructions
  • settings.txt: Settings. Crucial file. See below.
  • settings_description.html: Description of settings.txt
  • keep_open.txt: See granting open access below.
  • default-settings.txt: Do not change this file. Keep it as a backup of the original settings.
  • settings.log: shows which settings have been applied
    • Delete this file from time to time to clean it up
• MANUAL.HTML says that the settings are applied in 10 minutes. There is no need to stop and start the server. I have been stopping and restarting my server just to be safe.
  • To restart the server, go to the Control Panel button on the top right.
  • Stop the server.
  • Start the server.
• If you create a new environment, you must stop and start the server.
• MANUAL.HTML says that students have to stop and stop and start their servers for the changes to be applicable.
• You can force a restart on their servers by going to Control Panel.
  • Pick Admin from the top menu.
  • Not a good idea to force restart since this may disrupt students. However, student Notebooks are autosaved by the system.
• MANUAL.HTML asks to close the settings file after adjusting settings. "Do not keep settings file open (Jupiter will try to resave it at some point, and this may replace content with an error message)."

Step 1: Grant access

Instructor/TA access

You have to update INSTR_LIST in settings.txt.

• Login to the RCN Instructor Site for the class. For example, https://techub-23003-instructor.rcnyu.org/
• Go to the tuning folder.
• Click on Settings.txt to open it in a text editor.
• Go to INSTR_LIST
• Mine reads: INSTR_LIST:ss13638,eb167,dg46
  • The first two names are sysadmins from NYU.
  • To add any TA netids, add comma separated netids, no blanks, to INSTR_LIST
• Save Settings.txt
• Not necessary to do this, but I shut down and restart my server.

First class: Grant temporary student access to any NYU netid

Update keep_open.txt on NYU RCN

• Login to the RCN Instructor Site for the class. For example, https://techub-23003-instructor.rcnyu.org/
• Go to the tuning folder.
• Click on keep_open.txt in the tuning folder to edit it. (The same folder as settings.txt.)
• Specify the number of minutes that you want to allow open access. Max time allowed is 2160 minutes or 36 hours per settings_description.html. If you set this time to 0, you are shutting off open access.
• If the access is open, then ACCESS_INDICATOR:open will appear in settings.txt. If the access is not open, then ACCESS_INDICATOR:whitelist will appear in settings.txt.
  • It may take several minutes for the status to be updated in settings.txt.
• Apparently, the time remaining in minutes in keep_open.txt winds down over time. Therefore, if you Login later, you will see that time remaining in minutes is less than what you entered.
• If you want to keep granting open access, then you have to keep resetting this time limit.
• Not necessary to do this, but I shut down and restart my server.

Granting permanent access only to registered students after the roster is stable

Update keep_open.txt on NYU RCN

• Login to the RCN Instructor Site for the class. For example, https://techub-23003-instructor.rcnyu.org/
• Go to the tuning folder.
• Click on keep_open.txt in the tuning folder to edit it. (The same folder as settings.txt.)
• Set the number of minutes to 0.
• If the access is open, then ACCESS_INDICATOR:open will appear in settings.txt. If the access is not open, then ACCESS_INDICATOR:whitelist will appear in settings.txt.

Get comma separated netid list from NYU Brightspace

• Go to NYU Brightspace.
• Click on Roster.
• Click the Export Roster tab to download an Excel file with the class list.
• Open the Excel file.
• Use =TEXTJOIN(",",True,C1:C50) function to join the netids with a comma.
• Make sure that comma separated list has no spaces.

Update settings.txt on NYU RCN

• Login to the RCN Instructor Site for the class. For example, https://techub-23003-instructor.rcnyu.org/
• Go to the tuning folder.
• Click on settings.txt to open it in a text editor.
• STUD_WHITELIST should read as follows: (The first two netids are NYU RCN Admins, the third one is mine.
  • STUD_WHITELIST:ss13638,eb167,dg46
• Paste the comma separated (no spaces) netid list AFTER these names.
  • Painful to update this as students add/drop your course.
• Not necessary to do this, but I shut down and restart my server.

Step 2: Set up Python environment

Do not use the default conda environment. Set up your own and tell students to use it.

• Go to "New" drop down button and pick Terminal
• I executed the following commands because I need a simple Python setup. Replace "gode" by whatever name you prefer.
  • conda create --name gode python=3.7
    • Notice the two hyphens before name
  • conda install -n gode ipykernel
  • Install whatever packages you need.
  • It is a good idea to activate the environment and then install packages.
    • conda activate gode
  • Required: Stop and start the server.
  • Tell students to use this custom environment when creating and saving their files.
• You can always change environment by going to Menu-Kernel-Change Kernel while in a Notebook.

TURBO mode: When installing packages requires more memory and computing power

Sometimes, when one is trying to install new conda/python/R packages on instructor site, conda begins to do operations and then gives an error ("killed..." or similar). The reason for this error is small RAM/CPU allocation.

For that purpose you can activate a so called "TURBO" mode for instructor site (settings_description file describes how, under TURBO_INDICATOR). In this mode a whole VM (2cpu, 10GB RAM) will be allocated to your environment and then you can install (hopefully :) ) everything you need.

Notes:

• you need to wait for new settings to apply (look into log file as usual)
• When they applied, stop and start your server again
• This time it may load slower than usual (a new VM will be deployed for you)
• After a certain time this mode will be automatically switched to off

Step 3: Persistent file storage, sharing files, and interface

Instructor storage

By default, instructor files are saved for the duration of the semester. They will be deleted at the end. Therefore, make sure you back up your work. I do not think the system does daily backups. Therefore, back up your files anyway.

Student storage

By default student files are DELETED after every session. You have to grant them permanent storage for the duration of the semester. NYU will automatically delete their files at the end of the semester. Therefore, tell them to back up their work.

Update settings.txt on NYU RCN

• Login to the RCN Instructor Site for the class. For example, https://techub-23003-instructor.rcnyu.org/
• Go to the tuning folder.
• Click on Settings.txt to open it in a text editor.
• Look for the following entry: PERSISTENTSTORAGE:no
• Change it to: PERSISTENTSTORAGE:yes
• Not necessary to do this, but I shut down and restart my server.

How to share files with students?

• Any files saved in the shared directly become immediately available to the students.
• Ask them to refresh their browsers.

Where can students save their files?

Students cannot save their files in the shared directory. After persistent storage has been enabled for them, they can save their files in their home directory or any other subdirectory that they create.

Pick Classic Notebook versus JupyterLab interface

JupyterLab interface is better than the Classic Notebook. However, nbgrader works only with Classic Notebooks. Therefore, stick with Classic. Need to look into this more.

Step 5: Extensions

After NYU has enabled extensions, run the following commands:

• conda activate gode
• conda install -y autopep8
• conda install -y isort

Step 5: NBGrader

Need to figure this out.