Cloudron Forum

I have not seen it reported here and am also not able to reproduce it on the demo system, but when i search for an app and enter the second character the dashboard breaks with a javascript error.

utils-Cb1Rvih8.js:548 Already activated
index-D64usbhx.js:41926 Cloudron dashboard v9.0.13
index-D64usbhx.js:781 image loading failed, try to load /img/appicon_fallback.png
index-D64usbhx.js:781 image loading failed, try to load /img/appicon_fallback.png
style-D04uIdi2.js:2200 TypeError: Cannot read properties of undefined (reading 'toLocaleLowerCase')
    at index-D64usbhx.js:2339:295
    at wrappedFn (style-D04uIdi2.js:1051:19)
    at Array.filter (<anonymous>)
    at apply$1 (style-D04uIdi2.js:1059:27)
    at Proxy.filter (style-D04uIdi2.js:930:12)
    at ComputedRefImpl.fn (index-D64usbhx.js:2335:25)
    at refreshComputed (style-D04uIdi2.js:614:29)
    at get value (style-D04uIdi2.js:1740:5)
    at Proxy.<anonymous> (index-D64usbhx.js:2525:45)
    at renderComponentRoot (style-D04uIdi2.js:5499:17)
logError @ style-D04uIdi2.js:2200
handleError @ style-D04uIdi2.js:2194
renderComponentRoot @ style-D04uIdi2.js:5533
componentUpdateFn @ style-D04uIdi2.js:6728
run @ style-D04uIdi2.js:465
runIfDirty @ style-D04uIdi2.js:498
callWithErrorHandling @ style-D04uIdi2.js:2142
flushJobs @ style-D04uIdi2.js:2313
Promise.then
queueFlush @ style-D04uIdi2.js:2245
queueJob @ style-D04uIdi2.js:2240
effect2.scheduler @ style-D04uIdi2.js:6764
trigger @ style-D04uIdi2.js:488
endBatch @ style-D04uIdi2.js:546
notify @ style-D04uIdi2.js:779
trigger @ style-D04uIdi2.js:766
set value @ style-D04uIdi2.js:1589
createVNode.onUpdate:modelValue._cache.<computed>._cache.<computed> @ index-D64usbhx.js:2477
callWithErrorHandling @ style-D04uIdi2.js:2142
callWithAsyncErrorHandling @ style-D04uIdi2.js:2149
emit @ style-D04uIdi2.js:5388
set @ style-D04uIdi2.js:5341
set value @ style-D04uIdi2.js:1635
createElementBlock.onUpdate:modelValue._cache.<computed>._cache.<computed> @ style-D04uIdi2.js:16237
(anonymous) @ style-D04uIdi2.js:10083

@samuelhalff the official cloudron package for it is located at https://git.cloudron.io/packages/librechat-app

@joseph thank you. I can conform that after running the sql query the users do show up.

My observation with an empty database was because of a syntax error in my query.

I tried the upgrade today again. This is the log at the time of the upgrade:

Sep 29 10:35:34 box:tasks update 21502: {"percent":70,"message":"Creating container"}
Sep 29 10:35:34 box:apptask createContainer: creating container
Sep 29 10:35:36 box:shell apptask /usr/bin/sudo -S /home/yellowtent/box/src/scripts/configurelogrotate.sh add 4f552ab3-85b8-45de-b46a-05dc384d0c10 /tmp/4f552ab3-85b8-45de-b46a-05dc384d0c10.logrotate
Sep 29 10:35:36 box:apptask startApp: starting container
Sep 29 10:35:36 box:tasks update 21502: {"percent":90,"message":"Configuring reverse proxy"}
Sep 29 10:35:36 => Linkwarden configs
Sep 29 10:35:36 box:tasks update 21502: {"percent":100,"message":"Done"}
Sep 29 10:35:36 => Update configuration
Sep 29 10:35:36 ==> Setting up OIDC
Sep 29 10:35:36 box:taskworker Task took 103.25 seconds
Sep 29 10:35:36 box:tasks setCompleted - 21502: {"result":null,"error":null}
Sep 29 10:35:36 box:tasks update 21502: {"percent":100,"result":null,"error":null}
Sep 29 10:35:36 => Run database migration
Sep 29 10:35:37 yarn run v1.22.22
Sep 29 10:35:37 warning Skipping preferred cache folder "/usr/local/share/.cache/yarn" because it is not writable.
Sep 29 10:35:37 warning Selected the next writable cache folder in the list, will be "/tmp/.yarn-cache-0".
Sep 29 10:35:37 $ dotenv -- yarn workspace @linkwarden/prisma deploy
Sep 29 10:35:37 warning Cannot find a suitable global folder. Tried these: "/usr/local, /usr/local/share/.yarn"
Sep 29 10:35:37 warning Skipping preferred cache folder "/usr/local/share/.cache/yarn" because it is not writable.
Sep 29 10:35:37 warning Selected the next writable cache folder in the list, will be "/tmp/.yarn-cache-0".
Sep 29 10:35:38 warning package.json: "dependencies" has dependency "prisma" with range "^6.10.1" that collides with a dependency in "devDependencies" of the same name with version "^5.21.1"
Sep 29 10:35:38 warning Skipping preferred cache folder "/usr/local/share/.cache/yarn" because it is not writable.
Sep 29 10:35:38 warning Selected the next writable cache folder in the list, will be "/tmp/.yarn-cache-0".
Sep 29 10:35:38 $ prisma migrate deploy
Sep 29 10:35:38 warning Cannot find a suitable global folder. Tried these: "/usr/local, /usr/local/share/.yarn"
Sep 29 10:35:39 Prisma schema loaded from schema.prisma
Sep 29 10:35:39 Datasource "db": PostgreSQL database "db4f552ab385b845deb46a05dc384d0c10", schema "public" at "postgresql:5432"
Sep 29 10:35:39 2025-09-29T08:35:39Z
Sep 29 10:35:39 89 migrations found in prisma/migrations
Sep 29 10:35:39 2025-09-29T08:35:39Z
Sep 29 10:35:39 Applying migration `20250826181008_add_field_to_tags`
Sep 29 10:35:39 Applying migration `20250922133423_add_field_to_user`
Sep 29 10:35:39 2025-09-29T08:35:39Z
Sep 29 10:35:39 The following migration(s) have been applied:
Sep 29 10:35:39 2025-09-29T08:35:39Z
Sep 29 10:35:39 migrations/
Sep 29 10:35:39 └─ 20250826181008_add_field_to_tags/
Sep 29 10:35:39 └─ migration.sql
Sep 29 10:35:39 └─ 20250922133423_add_field_to_user/
Sep 29 10:35:39 └─ migration.sql
Sep 29 10:35:39 2025-09-29T08:35:39Z
Sep 29 10:35:39 All migrations have been successfully applied.
Sep 29 10:35:39 Done in 2.36s.
Sep 29 10:35:39 => Copying cache
Sep 29 10:35:40 => Healtheck error: Error: connect ECONNREFUSED 172.18.19.54:3000
Sep 29 10:35:41 => Ensure permissions
Sep 29 10:35:41 => Starting Linkwarden
Sep 29 10:35:41 yarn run v1.22.22
Sep 29 10:35:41 $ concurrently "dotenv -- yarn workspace @linkwarden/web start" "dotenv -- yarn workspace @linkwarden/worker start"
Sep 29 10:35:42 [1] $ tsx index.ts
Sep 29 10:35:42 [0] $ next start
Sep 29 10:35:43 [0] - ready started server on 0.0.0.0:3000, url: http://localhost:3000
Sep 29 10:35:47 [1] Initializing the worker...
Sep 29 10:35:47 [1] Starting RSS polling...
Sep 29 10:35:47 [1] Starting link processing...
Sep 29 10:35:48 [1] Processing new RSS feed items for This Week in Self-Hosted
Sep 29 10:35:50 [0] (node:192) [DEP0060] DeprecationWarning: The `util._extend` API is deprecated. Please use Object.assign() instead.
Sep 29 10:35:50 [0] (Use `node --trace-deprecation ...` to show where the warning was created)
Sep 29 10:36:02 [0] Warning: For production Image Optimization with Next.js, the optional 'sharp' package is strongly recommended. Run 'yarn add sharp', and Next.js will use it automatically for Image Optimization.
Sep 29 10:36:02 [0] Read more: https://nextjs.org/docs/messages/sharp-missing-in-production

And after logging in again the app was empty like with the previous update.

I then wanted to check the postgres database and this looks like it was wiped:

db4f552ab385b845deb46a05dc384d0c10=> /dt
db4f552ab385b845deb46a05dc384d0c10-> /d

So even though it says that it was able to apply the migrations to the database the database itself is empty afterwards. That was because of a syntax error on my end. I'm more of a MariaDB guy.

PS: also using sso and basically only using it with my first user. next to a few links I have setup a few rss sources in the app. the rss feeds are probably the most source for links since otherwise i hardly use linkwarden.

I can confirm the same. All data of the app is no longer available after the last update. A restore to be state before the update brought the data back.

@james said in Update Documentation to warn of Cloudflare Proxy issues:

should this be done by default when Cloudflare is used

Does this refer to the Allow list for WOPI requests setting in Nextcloud? Automatically allowing all requests that come from the Cloudflare infrastructure seems like a bad idea for me.

@qwinter said in Outline business self hosted cost 4$?:

SAML is only in their biz editon

Cloudrons SSO is based on OIDC https://docs.getoutline.com/s/hosting/doc/oidc-8CPBm6uC0I

Sounds like a fun trip. How fixed is your route? A college once recommended to me to visit https://titosbunker.com/ if I am ever in Bosnia & Herzegovina.

@crazybrad said in Help Desk with Chat Behind User Login:

host it only for authenticated users, and embed the ticket form and the chat widget.

Yes, that sounds like a good approach to me.

@BrutalBirdie reading linked posts?

@james said in Installing custom Apps on Cloudron:

upload a CloudronManifest.json

I would absolutely welcome the idea to make custom app installations more easy for third party installations and less technical users. I am not sure however is uploading a json file will really make it easier for the majority of users, as these will still need to rely on someone else producing a properly formatting manifest file (and don't mess up formatting on the way).

Including the container image name in the manifest does sound like a good idea, but how about instead of uploading the manifest through the store the user could provide the url for where multiple such manifests are stored? Then app developers could publish their own stores.

My first thought: who has created this website and what gives them credibility? I did not watch the embedded video so maybe its answered there. but the only way to identify the creators is by their github link.

I was playing around with Gatus a bit today. While its nice that you can configure all your monitors through yaml, that also makes it a bit more complex to get started since you first need to learn the exact syntax. I guess a config generator tool would be nice.

What is is missing compared to uptime kuma is a public status page. You either have to expose your whole dashboard or put it behind auth.

My minimal Cloudron app can be found at https://git.9wd.eu/apps/cloudron-gatus. It supports OIDC Login and stores data in the postgres database.

@jdaviescoates said in Roundcube Development continues with actual features:

Nextcloud Mail is getting pretty decent

I had the mail app in Nextcloud installed for a while, but never really used it so I played around with it just now. It does have some weird glitches when moving mails and when you first click on a folder it can take quite a long time for mails to be loaded. Selecting individual mails takes longer to display them compared to Roundcube.

I think one of the main tasks for a mail client is for it to be quick and easy to use.

I do see two strategies for you mail migration.

1. do the migration when you think that no normal person would be sending mail at that time.
2. shortly before starting the migration close port 25 in your filewall. This will lead to a temporary sending failure for everybody that would try and if that person is not a badly configured spam bot it will try to deliver the same mail at a later time again.

Here is a how-to https://earthkhan.medium.com/turn-your-openapi-in-mcp-server-in-5-minutes-e2859383d0dc

Isn't the better place for a mcp server in Cloudron directly at the Cloudron Api? Via mcp your ai could talk to the Cloudron Api, and through this access files, logs, execute commands, install/update.

The api is described here: https://docs.cloudron.io/api.html

I would think since there is an openapi spec there is probably already some mcp server that could import it.

@jorrg said in Expose reverse-SSH tunnel to Cloudron app?:

:localhost:11434

By binding the port to localhost it is available to the Cloudron host, but the apps which are running in containers on a separate network cannot reach it.

You need to bin them to a global IP (and then firewall it so no others can access it).

@james even with DNS verification that wouldn't work for Domains such as .lan though.

@Shai said in Is it worth it to upgrade to 24.04 from 22.04?:

I might as well get a couple extra years before having to remember how to do this again.

Shouldn't that be reason enough? 24.04 is already a year old, time to upgrade from the previous LTS.