Cloudron Forum

@potemkin_ai Indeed, Jitsi itself, running alone, in its own comfy space, installed per the Jitsi guide does run smooth as silk (albeit with some occasional drops). It is when you get into other implementations that are not Debian/Ubuntu that it seems to have issues. I've tried the OpenSUSE and Docker deployments and eventually always encounter an issue with colibri/websockets.

As for Element Call, they seem really focused on their new Element Starter Suite and Element Server Suite which is a lot like Element Matrix Services (EMS) but self hosted. Element Call is part of some deployments and really sparkles there. What is really exciting is that the Integrations options have really revved up.

I am testing both suites now and can say this is the most complete Element has felt when self hosting and I am not just saying that because you can default your deployment to use Hydrogen instead if you want to cut some bloat. Ha-ha!

Element Server Suite can integrate with everything and the kitchen sink and is apparently the only one who Element Call wants to be friends with.

@nebulon @girish and @nebulon I know you two are so busy and so I apologise for not following up on this. I eventually was able to resolve this matter, though how remains a bit of a mystery but I think I just had to wait for the state to reset on its own.

Anywho, sorry for not getting back to you on this.

@nebulon hey hey! This issue is certainly not isolated @avatar1024 but as @nebulon said, you really do need to examine the browser, console and WebRTC logs to get a handle on what may be going on. That being said, it is talked about on the Git by the creator:

i am getting Producer Transport Failed and consumer Transport Failed when i open my video or mic then this issue arise if my video and audio remains turn off then i don't get Producer Transport or consumer Transport issue. is this a bug or issue on my side? and how do i resolve it?

If you not have this issue on the official demo: https://sfu.mirotalk.com It's definitely some of your side misconfiguration.
You can follow 👉 [this documentation](https://github.com/miroslavpejic85/mirotalksfu/blob/main/docs/self-hosting.md) on how to self-host it correctly.
Thank you for the understanding.

Follow the link to the self-hosting guide where they've detailed the specifics all of which it appears Cloudron has implemented. @nebulon, correct me if I am wrong but the config.js located at /app/pkg that we edit in the Files is "tied" to the one located at /app/code/app/src, its config.js and config.template.js, yes? The intention being to have it allow for customisation without breaking what makes the magic work, so to speak? Ha-ha!

All this to say @avatar1024 that in my tests, it usually had to do with browser configurations, VPNs, extensions, blockers, etc. but in a properly configured production environment, you'd want to pull the browser logs to see where and why it fails.

Hope that helps.

Title says it all. I lost access to a server that went bye-bye and I reinstalled Cloudron and tied it to my account before even thinking to delete the apps that were on it and so now it won't let me install new apps on the new test server. Anything we can do?

@potemkin_ai in case you didn’t get a chance to find it (regarding the homeserver caveats):

By default, the app expects you to have a Matrix homeserver (such as Synapse) installed locally and running on port 8008…

Element Call requires a homeserver with registration enabled without any 3pid or token requirements, if you want it to be used by unregistered users…

Therefore, to use a self-hosted homeserver, this is recommended to be a new server where any user account created has not joined any normal rooms anywhere in the Matrix federated network.

It is for this reason that Element Call, while a robust and impressive solution the Matrix ecosystem, isn’t really a replacement (yet) for similar services. That being said, when deployed as part of an integrated Element Matrix solution such as Element Server Suite alongside Element X, it is incredible.

Jitsi Meet VideoBridge is crashing anytime more than two people join on any service that used it. It’s good that Element was working on Element Call before that. @girish and @nebulon, how was Cloudron able to run Jitsi so smoothly before when no other service could? Is that why the Cloudron version remained an older version?

@girish and @nebulon I really want to thank you for once again, as you always do, taking the time to individually address concerns and providing insight and details as you do. I’ve will also begin some testing of the SFU version of MiroTalk in its unstable form. Keep it up!

@nebulon

I've been tracking MiroTalk's rise since Jitsi’s decline— given Jitsi's frequent crashes across various servers, wow what a decline. My focus is on the SFU version amidst its alternatives: P2P, C2C, and BRO within the MiroTalk Web Admin suite to test all of them.

1. What held back SFU's initial adoption on Cloudron? Despite P2P and SFU's apparent similarity, differences in security features, like meeting locks available in SFU but not P2P, and some other custom settings.

Is P2P's initial preference due to its simpler setup, and if so, what privacy and customization aspects are compromised by it being pure WebRTC. It can only do one and one then until SFU becomes fully viable?

@robi ahhh that makes more sense. I figured that that was what your post was referring to and I recognised the Linux CLI look and feel in your post but I am still not as well versed in things to have picked up on the nuances of what your post meant. My bad.

@robi I am a huge fan of this app and I am so pleased to be able to play around in it. I know that you mentioned that there is no desktop app proper but I was curious if you had a chance to play around in the Linux CLI version? Seems to me that they are focusing on it to allow a level of functionality that can be embedded but I'll admit this is not my area of expertise.

@froodle So great to see the developer on here. I will certainly toss my support your way. This is a project that needs all the support it can get. I also followed your GitHub account.

It had some bugs last week but I was pleasantly surprised to see how quickly they were resolved. Great work!

I def think you have to strike a balance for sure but remember that the ease by which you access these services and sites is dependant on the sacrifice of one or more of your privacy, security or anonymity. You have to make that call for your own needs. What I will say is that the SPN provides the granular control you are looking for and I often hear people getting rid of Portmaster for the reasons you mention above. With SPN, I can set it so that Brave always opens with a connection in my home country but private windows in brave open in another. I have it set so that LibreWolf opens in the country I want to be streaming from and WaterFox is two hops obfuscated. The SPN then lets you determine the traffic even further. It’s worth diving into and maybe I should do a video l cause it is complex but their white paper does a good job explaining the architecture.

I can confirm that the SPN is a pretty strong tool and should not be discounted based on the reviews. It is always improving and provides you granular level control over where your connections originate and end at. If you guys want I can provide a detailed review.

Techlore did a review way back but I find him to be rather amateur and the features have improved since: https://safing.io/video/2021/05/27/techlore-reviews-portmaster-safing/ and the general idea is that you get port control, VPN, and Tor type power all in one.

I've been diving into this project called Libervia and I think it would be great for the Cloudron AppStore. Libervia isn't just another app, it's an XMPP-powered beast that could fill a niche we've been missing. We have some custom XMPP solutions using existing platforms but we've been lacking some solid XMPP action. Libervia is decentralised and federated web life, mixing XMPP and ActivityPub protocols. It's a universal communication tool that respects user freedom and privacy, like a boss.

You can check it out here: https://libervia.org

The source code for this bad boy's over at https://repos.goffi.org/libervia-backend/file/tip. In the docker-compose.yml file, you'll find the Docker image tagged as libervia/backend. See: Q2 2023

Here's the kicker, that got me thinking of making my first app suggestion (show mercy, ha-ha). Libervia is a bridge. It's got the ability to connect XMPP and ActivityPub, making it an amazing hub for users to roll with other ActivityPub apps. Mastodon, PeerTube, Pixelfed - all the big players we've got on Cloudron, even OwnCast recently, Libervia is like the perfect homebase for users to hub out into these other ActivityPub apps. I have also included a review of their Social Contract as I find it stands a bit taller than Blue Sky Social's which was recently discussed, link below.

By Q2 2023, they're planning to stabilize their ActivityPub ⬌ XMPP Gateway and API/Plugin for the bridge. They're also working to update and fix installation images for Flatpak and Docker, which makes it really easy to package from Cloudron, right? I've been getting my hands dirty with packaging apps lately, self taught and learning as I go and, while I'm not quite ready to take a big bite like Libervia, I'm all in to keep an eye on it and keep you guys posted.

Hit me back with your thoughts.

--A

• https://mashable.com/article/bluesky-twitter-terms-of-service
• https://forum.cloudron.io/topic/9020/blue-sky-social-at-protocol-authenticated-transfer-protocol-alternative-to-mastodon-activepub?_=1685653918908

@girish you hit the nail on the head. that was 100% what was going on but odd that it only started doing it recently. honestly, if I could use anybody else but Cloudflare, I would. #takingsuggestions

@nebulon I wanted to post my response in here so that others might find it helpful. If you run into this issue, turn off your proxying of Cloudflare, or whatever you use and also turn off any beta features. Then examine the console of the webbrowser you are in on the site from both a normal and private browsing window. While this didn't solve my issue, it helped me isolate where it happens and where it doesn't Hopefully it helps others. According to recent updates shown in my distro's logs, I think something fundamental changed in one of the main components of how these sites resolve but I can't tell if it is my end, the something itself, or Cloudron.

@nebulon sent and SSH support access granted.

@Lostuser this is exactly what I was seeing in the logs and some weird stuff with the Cloudron CLI as well. Glad I’m not the only one. This all started with the new update and I’ve changed zero. @nebulon If a fresh install of Cloudron didn’t fix it, that’s gotta mean something yeah? Can I rollback to the previous version maybe and see?

@nebulon I do believe this user and I have similar a issue being described different ways. I can confirm that my AppStore would not link to my Cloudron.io account during all this as well.

@jdaviescoates Thanks for the heads up. Yes, I have turned off everything. Still testing different browsers though and different NextDNS setups. What is odd is that this was never an issue at all until 7.4.1 and nothing has changed on my end.