Posts made by Camille

@girish Actually, I'm not really sure how we got there
I have installed the https://github.com/jupyterlab-contrib/jupyterlab-link-share extension, which kind of works, but we have to manually add the token to the URL it provides.
So maybe it enabled settings in the background.
I will try to redo it from a fresh install and keep you posted.

Hi @girish Token authentication works great on our Cloudron JupyterHub. We can share notebooks fine but with no real time collaboration: we just get the usual overwrite problems.

Reading https://github.com/jupyterhub/jupyterhub/issues/3578 and https://discourse.jupyter.org/t/jupyterhub-collaboration-and-tags/11259
it seems that collaborative flag would also work with JupyterHub.
It tried adding c.LabApp.collaborative = True in customconfig.py with no luck.
I'll investigate more thouroughly later.

Thanks @girish for all the pointers!

@jdaviescoates Thanks.
It's not clear to me what parameters can be changed in /app/data/customconfig.py
besides c.Spawner.mem_limit and c.DockerSpawner.container_image.
I searched in https://jupyterhub.readthedocs.io/en/stable/search.html?q=collaborative but didn't find many clues.
I'll try c.Spawner.cmd = 'jupyter-lab --collaborative' but that'll be mostly guesswork

Hi,

How can I enable real time collaboration in JupyterHub?
The doc (https://jupyterlab.readthedocs.io/en/stable/user/rtc.html) says

Editors are not collaborative by default; to activate it, start JupyterLab with the --collaborative flag.

How can I do that in a Cloudron env?
Thanks,
Camille

Thanks @girish for the fix. I have reported the problem upstream : https://github.com/Mailtrain-org/mailtrain/issues/1030

Actually, I get the same problem with a fresh install of the app (Package v1.7.0
App v1.24.1).
Has anyone a working instance of this version?

Hi,
Since a few days (and it doesn't seem related to the last update : Mailtrain Package v1.7.0 App v1.24.1), I get this error "Unexpected error (Bad Gateway)" whenever I try to upload an image with Mosaico or GrapeJS.

With both editors, the logs tell me

Mar 17 14:47:53 fs.js:168
Mar 17 14:47:53 throw new ERR_INVALID_CALLBACK(cb);
Mar 17 14:47:53 ^
Mar 17 14:47:53
Mar 17 14:47:53 TypeError [ERR_INVALID_CALLBACK]: Callback must be a function. Received undefined
Mar 17 14:47:53 at makeCallback (fs.js:168:11)
Mar 17 14:47:53 at Object.unlink (fs.js:1199:14)
Mar 17 14:47:53 at ReadStream.<anonymous> (/app/code/node_modules/jquery-file-upload-middleware/lib/uploadhandler.js:149:48)
Mar 17 14:47:53 at ReadStream.emit (events.js:327:22)
Mar 17 14:47:53 at ReadStream.EventEmitter.emit (domain.js:467:12)
Mar 17 14:47:53 at endReadableNT (internal/streams/readable.js:1327:12)
Mar 17 14:47:53 at processTicksAndRejections (internal/process/task_queues.js:80:21) {
Mar 17 14:47:53 code: 'ERR_INVALID_CALLBACK'
Mar 17 14:47:53 }
Mar 17 14:47:53 npm ERR! code ELIFECYCLE
Mar 17 14:47:53 npm ERR! errno 1
Mar 17 14:47:53 npm ERR! mailtrain@1.24.0 start: `node index.js`
Mar 17 14:47:53 npm ERR! Exit status 1
Mar 17 14:47:53 npm ERR!

The images actually get uploaded to /tmp.
Anyone having the same problem or any idea on what to investigate ?

Thanks !

@Susanne said in Django - The Web Framework:

https://www.djangoproject.com/
maybe something similar like the Cloudron LAMP

Yes, a combo Django + Postgres would be great.

Hi,

When you launch a new intallation on Scaleway, your script gets stalled, because it's actually waiting for an input, that you can see only if you

tail -f /var/log/cloudron-setup.log

in another terminal. Tailing so is mentioned when you launch the script, but it might be a good thing to also mention it in the doc (https://cloudron.io/documentation/installation/) .
In this second terminal you are presented with the configuration window of libssl1.1. (see https://framadrop.org/r/zT1Hpb_Gpv#RLhN9KT0E1thH8/a/6UWftNLxAN+slsmEeY5bOgtssM= )
So you have to type in the first terminal (basically just "Enter") to answer what you see in the second.
It's easy to workaround, but it's a little disorientating at first, so it might help to change it.
Cheers,
Camille

Apparently, it was related to email notifications.

Hi @girish Yes I did. It's about 15 Go, and plenty of it is free, so I'd be surprised it was the problem.
We have 15 concurrent apache processes and that's the value in the mpm_prefork.conf file, but as it's read only, we cant' change this value ; and also, because those processes accumulate, increasing the max value would just postpone the problem, I'm afraid.

Hi,
Since a few days, our Kanboard instance ends up stopping responding a few hours after a reboot.
We have a few errors like

[mpm_prefork:error] [pid 1] AH00161: server reached MaxRequestWorkers setting, consider raising the MaxRequestWorkers setting

in the logs. Could that be related ?

Thanks,
Camille

@nebulon Yes I saw this pb with sudo, (but that can be solved if you tail -f in another terminal as advised: you can type the answer in the original terminal and see the result in the second). Tzdata has a text UI, but I don't know if it will create problems in the context of the install script :

HTH,
Camille

Hi,
Thanks for the fix.
Regarding the export DEBIAN_FRONTEND=noninteractive statement : I'm really not sure it's required, but apt-ing tzdata will prompt an interactive interface that might not be needed.
Best,
Camille

Hi,

The install script will fail on Scaleway / Ubuntu 18.04 Bionic as

export DEBIAN_FRONTEND=noninteractive
apt install tzdata

fixed it for me.
Maybe the install script could be updated to take this into account?

Thanks,
Camille