Cloudron Forum

@girish thanks, appreciate it! already tested, templates are back

@girish I'm afraid this update removed the custom invoices templates, I can't create new invoices because the templates are now "missing". Could you symlink that path again? /app/data/invoices was used for templates, /app/data/data/invoices/ for exported pdfs (https://www.kimai.org/documentation/invoices.html#create-your-own-invoice-document for reference)

@simon I had that as well a few years (!) ago, when I was using a non-SSD VPS with little CPU and just too many records in the sqlite db. The quick fix was using a lower interval and regularly hit the "Shrink Database" button in the settings, the real fix was migrating to a server with a faster storage

I had a question about that a while ago: https://forum.cloudron.io/topic/12503/what-will-happen-when-i-enable-ipv6-on-cloudron-when-some-domains-don-t-support-it/2?_=1733313066534

@girish suggested this:

If you change the provider of the things you don't control to no-op, and make sure A/IPv4 is correct, then it won't crash and just ignore them. Certs should also be renewed (as long as there is no incorrect AAAA record).

Thanks for clarifying. Yeah I realized that a while ago, planning to do it differently, maybe even before the migration!

Sorry for semi-hijacking this topic, but since it's somewhat connected, I think it's okay

I'm also about to migrate a VPS with a bigger setup. It uses a number of custom apps and the Docker registry is also part of the Cloudron apps. The migration can only work if I do a dry-run, right? If the Docker registry isn't up during the restore process, the apps can't pull from the live-registry, or am I overthinking this and it will work because it is either still online on the old server or will be on the new one? Are there "priority apps" during the restoration which would make it easier? I think I saw multiple apps being restore at once, so that wouldn't work.

Ah yeah, that's what I thought. Moved the build app to somewhere else. Would be nice to add --cpu-quota to the build command to mirror the apps settings (from https://dev.to/code42cate/dont-build-docker-apps-without-this-flag-48kg)

I'm using the build service to build a custom image. The last step, when it starts to compile everything, almost brings down the whole server, meaning all Couldron apps and system view become unresponsive, because it uses all available CPU power. I've tried setting it to 50%, which would mean it's only allowed to use 4 cores of the available 8, but as soon as it's running, it will max out all cores. The build-app itself shows almost no CPU usage (0.something%), but the system view shows this:

Any suggestions on how to limit the resources so this won't happen?

Sidenote: the link with the questionmark in Cloudron in the apps/resources view points to https://docs.cloudron.io/apps/#cpu-quota which is not working anymore, the current one is https://docs.cloudron.io/apps/#cpu-limit

https://forum.cloudron.io/topic/12817/oidc-login-broken-with-8-1-0-update/7?_=1732142838325

thanks @nebulon & @girish, this does indeed fix everything, appreciate it

Quick update: apparently removing that snippet and fresh install is working, but for various reasons, reinstalling all my Directus instances is not a great option. Thing I've tried and didn't work: restarting app/box/server, changing location, permissions, cloning, using a fresh browser.

@girish thanks for the reply, but still getting the same error even after restarting box and Directus. Any other ideas? Some cache that I might need to empty?

Hi there! I already sent an email to support, but since it's quite urgent, trying here as well, maybe someone has a solution for this. After updating to 8.1.0, OIDC authentication doesn't work anymore. It's easily reproducable, fresh install on 8.0.6 works, 8.1.0 doesn't.

It says Service Unavailable and refuses to login and logs show this:

WARN: [OpenID] Unknown OP error
err: {
"type": "OPError",
"message": "unauthorized_client (requested grant type is not allowed for this client)",
"stack":
OPError: unauthorized_client (requested grant type is not allowed for this client)

Any ideas/workarounds @nebulon @girish @joseph?

thanks for the pointer @humptydumpty, but I would want to show all the rows in a shared view by default (which doesn't seem to be an option anyways, maxed out at 100 it seems). It doesn't seem to be very robust, do you have a recommendation for an alternative? Basically I have tabular data and need to show each tab (~300 rows each) to a individual users (readonly) with different authentication, filtering is a bonus.

Hi there! Currently trying out NocoDB and at a first glance it's quite nice, but I can't seem to be able to get the env vars (new link is https://docs.nocodb.com/getting-started/self-hosted/environment-variables/ btw, the one from the docs and env link goes to an older version) to work...

I tried showing more rows than 25 and also disabling registration (via https://forum.cloudron.io/topic/6552/restricting-registration/), but those changes to env aren't reflected after restarting the app:

export NC_DISABLE_TELE=true
export NC_INVITE_ONLY_SIGNUP=yes
export DB_QUERY_LIMIT_MAX=100
export DB_QUERY_LIMIT_DEFAULT=1000

Any idea what's going on there (other than mixing true and yes, wth!)?

EDIT: Registration can be disabled via admin panel: https://docs.nocodb.com/account-settings/oss-specific-details#enable--disable-signup

thanks, made a snapshot and got rid of the most important/obvious logs, cheers

Unsure about the 90 days though, many of my servers have logs until the day they were created.

@nebulon sorry for digging up another corpse, but can you confirm that the event log is supposed to be cleared every two weeks? I'm about to hand over a Cloudron to a customer and wanted to see that all the try and error attempts during setup are not in there, but the Event Log shows pretty much everything, dating back almost two years. How can I manually clear the event/mail/cert logs?

Ran into the same issue, apparently Ubuntu 24 requires an edit in /lib/systemd/system/ssh.socket as well, you need to adjust the ListenStream to 202 and trigger a systemctl daemon-reload - after that it should be listening on port 202.

thanks @nebulon, I thought as much, but what's the actual syntax? The same as in the suggestions in the UI? Should that also update the xml file? Does the app require a restart afterwards? I just couldn't get two hosts to work...

Sorry to re-use this thread, but I tried adding two domains on a collabora instance and couldn't get it to work with both of them - it either gave me this 502 or some other WOPI error. Until I found this snippet and added it manually into the coolwsd.xml file.

The gui suggests you can add multiple domains via regex, split with pipes, but I tried all variations and just couldn't get it to work properly. Is there something that I've missed? One domain includes a dash, but with or without escaping, it just wouldn't work (only with the method above). Is there a trick to it to make collabora work with nextcloud domains?