If you want to explore having a custom package, you have to fork https://git.cloudron.io/cloudron/mastodon-app/ and patch in your theme code accordingly .

@JOduMonT Both issues are fixed in the latest package.

@girish said in Vultr DNS default TTL:

The initial value set is really just a default and nothing else. It's not even a preferred value as such. Admins have perfectly valid reason to set up long TTLs (especially for long lived sites).

This would be a good reason to have TTLs be configurable 😉

But okay, I think I understand that it'd be a bigger backend change than the gain from it, then. I suppose I could find a way to use the Vultr API to modify all the TTLs for consistency-sake.

@girish That issue appears to have happened again two days ago... So originally on May 19th, but most recently I see occurrences of Temporary translation failure on May 20, 2022 around 11:50 PM Pacific Time. So it's definitely an infrequent issue, but appears to be recurring periodically.

@girish Yes, they certainly are all different, and Cloudron certain is my favoured tool for production instances management.

However, there seems to be a universally acknowledged issue in the bar for entry for apps being packaged is so high, that the wish list grows faster than apps are packaged. So what do we do?

Some more confident system admins can fire up new small VPS/VM instances and try installing those unpackaged apps using their official Docker Compose scripts, and anyone else just has to wait or find an alternative.

I'm trying to find a way to reduce the barrier to entry for that next-best alternative.

Maybe I can ask the question a different way...

How do you suggest a Cloudron subscriber gets to use any apps on the wish list that aren't packaged or don't have any available packagers in the immediate future?

My whole technology career is based on figuring out how to get out of technology traps, so what't the next-best alternative here?

Would having Cloudron able to fire up and manage KVM containers solve this? Kinda like a multi-cloudron but on a single Cloudron server? Kinda like Cockpit offers.

@plusone-nick Yes, the are limitations but imho nothing extremly low. E.g. Microsoft SQL Server Express is limited to 10GB user data compared to the 12GB of Oracle XE - and despite of this limitation both products have their value and are widely used in small settings.

@girish said in High Performance Back-end for Nextcloud Files:

@eganonoa thanks for the notes! I will test this out once we have get 7.2 out and see how we can fix up nextcloud.

Awesome. Thank you. Glad I've been able to help out somewhere. Cloudron is a very important business to us.

@borax We use the latest prosody package from https://download.jitsi.org/stable/ which is 1.0.6155-1 (package version and not prosody version).

Can you tell us how to test this and also where this prosody requirement comes from?

@girish damn, thanks.

I'll add something there too.

@jdaviescoates I put the docs in https://docs.cloudron.io/apps/roundcube/#upload-size . I think adjusting $config['max_message_size'] = '100M'; takes it over 75M . Note that roundcube uses 1/3 of the size here for attachments (the rest is for the message body itself). So, maybe make it like 150MB or something.

@girish yes, seems to be entirely self-hosted.
There is no landing page or app front page.
You get an error if you try to access the URL.
It just needs to be running, and then typically in the settings of a browser extension (under Advanced) you can specify your custom url (in form of https://app.domain.tld/v2)

LanguageTool is one of the built-in "services", as they call their packaged apps. Terminology a little strange, but hey ho.

So really it is just a question of :

That's all that's needed from memory.
I can reinstall the service if you need 💯 % certainty of steps.

Coolify is interesting, but quite sparse in terms of apps.
Not a patch on Cloudron.
Seems more angled to deploying own written apps.
Links to Github nicely to pull repo, but I can't get my Cloudron Gitlab linked.

@robi I seems to be built with only about 5% php, it's mainly built in JavaScript.

I doubt very much this would install and run on LAMP nor even on LEMP, the installation is done through CLI and requires to install Node, Npm, PM2, Nodemon, MySQL, and MongoDB.

I believe that should be ported to docker first, before it could be installed on Cloudron.

@Grienauer said in app does not respond. statping-ng replacement possible?:

As the cloudron was updated to statping-ng… what steps should be done to do this for my old version?

This has not been done yet, I am yet to find the time to switch it.

@girish Thank you for the instructions, it's really simple

This should be fixed with the next release. There were a couple of state issues which needed fixing.

This should now be fixed with next release. It will show 100+ in such a case.

I can only reproduce the LICENSE.md issue and tried to create a symlink for it into /app/data but apparently their health checking code is not realizing the writeability through the symlink. We have to investigate further upstream then to see what solutions may be possible.

Edit: I created an upstream issue to gather more information on this https://github.com/invoiceninja/invoiceninja/issues/7458

@girish I think the behavior has to be mime-type dependent then. Currently clicking on a file as such will open the side-bar and attempts to load the content simply in an iframe. This was done to keep the local browser set filetype handling. For everything the browser does not support directly, this will trigger a downloading. Unfortunately I don't see a way to determine this upfront client-side in surfer, without attempting to add built-in viewers for content types like pdf. Maybe the sidebar could simple be less animated to avoid some of that?

For the archive support, this will be added then.

Ah if your instance is behind a NAT in your local network, then you have to forward a few ports from your router to the internal Cloudron server. 443 and 80 are basically the bare minimum. You can get some overview at https://docs.cloudron.io/security/#cloud-firewall

Further as @girish recommended, disable the cloudflare proxying at first until you managed to get it working without it. Just to rule out one possible point of failure.
The DNS records values in your case would be the public IP of your router. Then with the portforwarding rules, the connections will reach your internal Cloudron server. In case you are on an ISP connection where the IP might change on reconnect, you can consider using the dyndns feature of Cloudron.

@girish yay ! 🍾
Despite the hour I couldn't resist trying it.
I changed the location of the api from minio-api.domain.tld to minioapi.domain.tld (just removed the hyphen) and saved the change.
Renewed certs and logs now show the api domain in there.
Tested with Minio mc CLI and Forklift : they both list buckets and contents.
Will check MountainDuck and others later.

Thanks for your patience and support.
Marking it solved ! 🍾

@jdaviescoates said in Miniflux - Miniflux is a minimalist RSS reader:

Have you got a link to (or could you write?) a guide to setting something like that up?

@ajtatum I created a section for miniflux - https://forum.cloudron.io/category/144/miniflux . Feel free to create new post for your write up there.

@MORGiON There shouldn't be an issue. Can you post a screenshot of what error you are seeing ? Is the nginx error on your Cloudron dashboard or is it on cloudron.io ?