mechmotum / jupyterhub-deploy-teaching Goto Github PK

Need to grab it from the live server.

This is referring to the script users can run to archive their home directory:

#!/bin/bash
# Back up current user's home directory.
homedir=$(getent passwd "$LOGNAME" | cut -d: -f6)
backupfile="backup.zip"
find "$homedir" ! -name "$backupfile" | zip -u "$homedir/$backupfile" -@

Also make sure logging from two different services (eng122 and mae223) works.

apt:

• aptitude
• screen

Kevin's note:

conda issues with the ansible were resolved. you have to make sure to run the ansible playbook from the beginning each time.

This needs more clarification. It should work and ansible should be able to handle dependent tasks properly, as far as I know.

@mechmotum/libretexts

• Running with DHCP is a pain--got bit hard by connecting to a stale VM instance when I was deploying to a new one. There goes half a day.
• Would be nice to just forward ports rather than running on its own little network. Got cross-origin API requests when I tried it a while back though.

May not be able to check this task in.

I tried downloading the Washing Machine notebook as a PDF and it returns only 2 pages. If I download the LaTeX source, it has errors finding the images listed in the markdown cells. But you can compile the document fully by pressing enter at the errors. This gives a full PDF with missing images. The first missing image is at the top of page 3, so I suspect this issue is causing problems:

jupyter/nbconvert#552

But, I also tried to use IPython.display.Image instead, but after saving the IPYNB file, the download as latex and download as PDF still return the same files pre-updates and saves. That's more confusing.

I did this on the server:

conda install jupyter_contrib_nbextensions
jupyter nbextension enable exercise2/main
jupyter nbextension enable rubberband/main

Tasks:

• Tag most recent commit at the 16.04 tag so we can use it if needed.
• Review #20 to incorporate any issues there that are relevant.
• Consider merging in changes that are now in https://github.com/jupyterhub/jupyterhub-deploy-teaching
• Update this repo to work on 18.04: VM and bare metal server.
• Practice deploying to individual servers (each student).
• Upgrade the new bicycle server to 18.04.
• Deploy to the new bicycle server.
• Correct the SSL certificate.
• Wipe the old bicycle server.

@celine168 @kkrausse @lux12337 @zxcGrace @rfeltstykket here is a basic issue to track the updates we discussed today. Check the boxes as you tackle the tasks and fee free to edit this issue or discuss here.

Not sure what to make of this:

Can you get the pdf download working? Should just be some texlive apt get installs.

Install ffmpeg to convert animations to HTML5 videos.

Currently just keeping them local and not checking in.

IIRC Jess Hamrick mentioned the database needs to be postgres or something else non-default

Also not clear if this will cause errors or fail silently/corrupt the database.

My friend crashed my JupyterHub server by using up all of the RAM (he did it by running an infinite loop with a recursive Fibonacci function and a new process spawning on each iteration of the loop). Richard suggested creating cgroups so that we could limit memory and CPU usage.

Kenny would you be wiling to move this fork to the mechmotum user?

...but leave it non-default for now.

The math that rendered in the normal jupyter interface is correctly rendered mathjax, but when you open that up in the grading interface lots of things seem removed and fonts are wonky.

Need to handle deploying, grabbing backed up home dirs, and making sure students can log in. Might need to create all student accounts upon deployment.

• Not clear about copying hosts.example and jupyterhub_hosts.example. Edit then rename? Copy to new name and edit?
• use 'ip addr show` on the server to get the IP
• You need to git submodule update --init -- recursive after cloning the repo.
• ansible asks for a sudo password, it seems this is the sudo password for your local machine that is running ansible
• install ansible locally: apt get install ansible
• wasn't sure if we were using default ssh port on server or not
• didn't understand names of bicycle1 bicycle2 folders on drive that held cert stuff, is 2 just the newest?

Got this error first time:

TASK [nginx : start the nginx service] ******************************************************************
fatal: [bicycle]: FAILED! => {"changed": false, "msg": "Unable to start service nginx: Job for nginx.service failed because the control process exited with error code. See \"systemctl status nginx.service\" and \"journalctl -xe\" for details.\n"}

RUNNING HANDLER [nginx : reload nginx] ******************************************************************
	to retry, use: --limit @/home/moorepants/Websites/bicycle.ucdavis.edu/jupyterhub-deploy-teaching/deploy.retry

PLAY RECAP **********************************************************************************************
bicycle                    : ok=30   changed=16   unreachable=0    failed=1   

On the server:

root@bicycle:~# systemctl status nginx.service
● nginx.service - A high performance web server and a reverse proxy server
   Loaded: loaded (/lib/systemd/system/nginx.service; enabled; vendor preset: enabled)
   Active: failed (Result: exit-code) since Mon 2018-10-15 15:57:09 PDT; 1min 35s ago
  Process: 32440 ExecStartPre=/usr/sbin/nginx -t -q -g daemon on; master_process on; (code=exited, status

Oct 15 15:57:09 bicycle systemd[1]: Starting A high performance web server and a reverse proxy server...
Oct 15 15:57:09 bicycle nginx[32440]: nginx: [emerg] PEM_read_bio_X509_AUX("/etc/nginx/ssl/ssl.crt") fail
Oct 15 15:57:09 bicycle nginx[32440]: nginx: configuration file /etc/nginx/nginx.conf test failed
Oct 15 15:57:09 bicycle systemd[1]: nginx.service: Control process exited, code=exited status=1
Oct 15 15:57:09 bicycle systemd[1]: Failed to start A high performance web server and a reverse proxy ser
Oct 15 15:57:09 bicycle systemd[1]: nginx.service: Unit entered failed state.
Oct 15 15:57:09 bicycle systemd[1]: nginx.service: Failed with result 'exit-code'.

After running JupterHub, it seems like there is an error originating in the 'cull_idle_servers.py' script. The error doesn't seem to be affecting the server from running: