Cloudron Forum

@robi said: The question is what the experience is like in the terminal. in an app packaged with cloudron-minibase ? as far as I can tell, it is the same as with cloudron/base.

@necrevistonnezr said: Just out of curiosity (and little technical knowledge): Shouldn't a system that isolates apps via docker not be taken down completely by a malformed configuration file of one app? Yeah, absolutely. This is a good conversation to have Cloudron was written primarily for pre-packaged apps written by us and not as a sandboxing environment to run arbitrary applications. When packages were written by us, we could "trust" the input in the manifest file, description, post install etc. When we made custom apps possible, the worst case was that a user sabotages himself with a poorly written manifest. But now with community apps, there is a new vector that malicious apps can do unknown things. For example, what if the description has some clickjacking or scam link? So, you can hit problems even without installing or running an app (so we don't hit run time isolation). This is also why we don't want to integrate community listing straight into the main dashboard, this would give too many good feelings of trust. We are fixing issues (like this) as and when they come, but one should only run 3rd party packages they trust. Cloudron company's primary focus is still to runs the apps from the main app store and not to make a general trusted environment to run random stuff (FWIW, I don't think such a product exists out there).

TIP : Opencode auto installs git to a new empty folder. So it's worth creating an empty file in a new project. My time-saver workflow : create a bash script ~/new-proh=ject.sh and chmod +x ~/new-project.sh #!/bin/bash [ -z "$1" ] && { echo "Usage: $0 <project-name>"; exit 1; } echo "alias ll='ls -al'" >> ~/.bashrc source ~/.bashrc mkdir /app/data/projects/$1 touch /app/data/projects/$1/README.md from app UI, use any existing project session terminal to run it : ~/new-project.sh projectname then I can do git init and repo connection as I want it (custom values)

That is fixed now, needed an openid integration fix. Thanks for reporting!

Another thing if you activate codebase memmory on repos the search with graphql is instant and sometimes I thought models are hallucinating even Fable 5.1. But it wasn’t the case, whole repo loaded and schemas loaded in ram and results are at least 4 times faster then normal windows based same tooling and usage as no overhead of windows filesystem nor Microsoft defender

@necrevistonnezr my own approach at present is to mostly not use Community Apps yet. If/ when I do so, I like to check e.g. When the user reg'd on this forum (as a proxy of how long they've been using Cloudron) The users "reputation" score on the forum If the user has an "App Dev" or other (e.g. Partner, Translator) badge on the forum How often the Community App gets updated (which I'd then compare to how often the upstream up gets updated) Ideally, I'd like http://ca.cloudron.io/ to help surface this info in one place to inform my decision-making.

@nebulon thank you I’m not suggesting anything major, so don’t go down any hosting / domain issues. Obviously I’ve missed some settings, sorry. When I have found them, they may be enough. Will check again later.

@timconsidine Yay works!

Mac and Windows binaries now in releases if you want to try w/o cloning and building. https://forgejo.tcjc.uk/cca/cloudron-portal-manager/releases

@andreasdueren thanks, I didn't know about Nova

Package updated to upstrean v0.28.1

Resolved by restoring com.espocrm.cloudronapp@2.20.4 Then updating to Cloudron 2.20.5 Then updating to Cloudron 2.20.6 (EspoCRM 9.3.6) I had a lot of apps going into error when 9.1.6 got installed on a reboot. Upgrading to 9.1.7 (out of beta) fixed most of them, but EspoCRM needed this intervention. Just FWIW if others are encountering the same.

I have just run into the same issue. Isn't the issue that the build process forks into different processes per worker and the sum of it takes too much memory? I had to set my memory limit to 5 GB to get the app to start again. It looks to me like the workers are given 75 % of the memory assigned to umami each, which leads to the issue. Also, does the app have to recompile every single time it restarts? Maybe the default memory could also be raised above 2 GB considering that several people (also in another thread) faced memory issues before.

@James thank you, will check it out and report back

@James thank you I had not initially intended to use CloudronVersions in this way, but it is a real time saver. Yep, once an app is published, probably no trimming (unless very old revoked versions).

Upgraded to upstream v2.2