Cloudron 9.0 (beta) bug reports

[Kubernetes]

I think there is something wrong with OID and Safari on iOS since upgrade to Cloudron 9. I cannot login to gitea instance or App Proxy with login requirement. I get error message: "You do not have access" - of course I do.

Error Message in Browser:
{"status":"Bad Request","message":"missing query argument "code""}

When I login on a macos computer there are no issues.

[SansGuidon]

Hey,
I'm worried that some API calls I made in the past are no longer working with V9, and that's both with my instance and with Cloudron demo instance.
I'm trying to simply use the API Token and get system disk usage via the route documented in the API.
e.g with the demo instance of cloudron and a token:

$ curl -H 'Content-Type: application/json' -H 'Authorization: Bearer <REDACTED>' https://my.demo.cloudron.io/api/v1/system/info
{
  "info": {
    "sysVendor": "FUJITSU",
    "productName": "CELSIUS-FTS",
    "uptimeSecs": 410396,
    "rebootRequired": false,
    "activationTime": "2025-01-02T18:04:55.000Z"
  }
}

$ curl -H 'Content-Type: application/json' -H 'Authorization: Bearer <REDACTED>' https://my.demo.cloudron.io/api/v1/system/disk_usage
{
  "status": "Not Found",
  "message": "No such route"
}

Did some routes change compared to the API Docs?
https://docs.cloudron.io/api.html#tag/System/operation/getDiskUsage

Thanks!

[jdaviescoates]

Have these setting be separated out in Cloudron 9?

I wanted to turn of Automatic Platform updates, but leave App updates on:

[d19dotca]

A small observation on the Event Log page (sorry again if this has been reported already, I didn't notice it yet):

There seems to be no gap between the Time & Source columns, at least not using the date format and such that I see. Notice the bleed-in between PM and the Source info like "mail" and "oidc", etc.

Screenshot below. I am using the Safari web browser on macOS.

[nebulon]

@Kubernetes said in Cloudron 9.0 (beta) bug reports:

I think there is something wrong with OID and Safari on iOS since upgrade to Cloudron 9. I cannot login to gitea instance or App Proxy with login requirement. I get error message: "You do not have access" - of course I do.

Is there some way to get console errors in this case? I don't have an iOS device to test this on unfortunately. Or can others reproduce this also?

[Kubernetes]

@nebulon are these the logs you are looking for?

[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:27 +0000] "GET /login?redirect=/ HTTP/2.0" 200 292170 "-" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:27 +0000] "GET /api/v1/apps/[REDACTED]/icon?1762765767843 HTTP/2.0" 200 14575 "-" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:27 +0000] "GET /api/v1/apps/[REDACTED]/icon HTTP/2.0" 304 0 "-" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:29 +0000] "GET /openid/auth?client_id=[REDACTED]-proxyauth&scope=openid%20profile%20email&response_type=code&redirect_uri=https://[REDACTED]/callback HTTP/2.0" 303 79 "-" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:29 +0000] "GET /openid/interaction/36heYBCBJOx4nnREDACTED3Cde1Ylinxnsxk HTTP/2.0" 200 2325 "-" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:29 +0000] "GET /api/v1/cloudron/avatar HTTP/2.0" 304 0 "https://my.[REDACTED].de/openid/interaction/36heYBCBJOx4nnREDACTED3Cde1Ylinxnsxk" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:29 +0000] "GET /api/v1/apps/[REDACTED]/icon?1762765769473 HTTP/2.0" 200 14575 "https://my.[REDACTED].de/openid/interaction/36heYBCBJOx4nnREDACTED3Cde1Ylinxnsxk" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:31 +0000] "GET /openid/interaction/36heYBCBJOx4nnREDACTED3Cde1Ylinxnsxk/abort HTTP/2.0" 303 0 "https://my.[REDACTED].de/openid/interaction/36heYBCBJOx4nnREDACTED3Cde1Ylinxnsxk" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:31 +0000] "GET /openid/auth/36heYBCBJOx4nnREDACTED3Cde1Ylinxnsxk HTTP/2.0" 303 169 "https://my.[REDACTED].de/openid/interaction/36heYBCBJOx4nnREDACTED3Cde1Ylinxnsxk" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:31 +0000] "GET /callback?error=access_denied&error_description=End-User+aborted+interaction&iss=https%3A%2F%2Fmy.[REDACTED].de%2Fopenid HTTP/2.0" 400 68 "https://my.[REDACTED].de/" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"
[no timestamp]  [REDACTED] - - [10/Nov/2025:09:09:33 +0000] "GET /callback?error=access_denied&error_description=End-User+aborted+interaction&iss=https%3A%2F%2Fmy.[REDACTED].de%2Fopenid HTTP/2.0" 400 68 "https://my.[REDACTED].de/" "Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/26.1 Mobile/15E148 Safari/604.1"

[nebulon]

hm so looks like safari on iOS aborts the user interaction from that error. Not sure why, I guess we have to find some device to properly debug this then.

[girish]

@hakunamatata thanks, this is fixed in https://git.cloudron.io/platform/box/-/commit/d894de0784ffcb147226c38967311360293e704a .

You can temporarily ignore the notification but if it bothers you:

• Domains -> Edit that Cloudflare domain.
• Change the provider to something else. Then, change it back to Cloudflare. You have to re-enter the API key. Then, Save.

[girish]

@SansGuidon yeah, some routes have changed to adapt to the UI. We haven't updated the API and docs with Cloudron 9 API and screenshots yet (sorry, release and everything all going on in parallel) !

Currently, it's roughly this:

• /api/v1/system/block_devices - this lists block devices . currently, this does not return used space. this was an oversight, i have added this now.
• /api/v1/system/filesystems - this gives filesystems and size/used information . probably, what you want?
• /api/v1/system/filesystem_usage?filesystem=/dev/sda1 - this does a further "du" on one of the filesystems listed in the above call. Since this takes some time, the output is an event stream. You have to pass -H 'Accept: text/event-stream' .

[girish]

@d19dotca said in Cloudron 9.0 (beta) bug reports:

Before, I could search for "WordPress" for example and all the apps would filter to the WordPress installs.

Fixed in https://git.cloudron.io/platform/box/-/commit/a8513cc0fa8e8014da6d8aaa7a963c0ae168b4e5

[girish]

@WiseMetalhead said in Cloudron 9.0 (beta) bug reports:

It looks like the multipart upload settings for S3-compatible storage aren’t working as expected.

This works for me on DO Spaces and AWS S3. Which storage provider are you using? I can test with that. Amazon created a new module for S3 uploads and we had to switch over the node module, maybe there is some incompatibility.

[SansGuidon]

Thanks @girish, this will work

[WiseMetalhead]

@girish said in Cloudron 9.0 (beta) bug reports:

Which storage provider are you using?

I’m trying to use an S3 API Compatible (v4) provider.

[girish]

@WiseMetalhead yes, but which service or set up specifically? (or is this confidential?). How can we reproduce this?

[WiseMetalhead]

@girish said in Cloudron 9.0 (beta) bug reports:

which service or set up specifically

It’s a local provider for me, and there’s no specific configuration for it in Cloudron.
For convenience, I can create a bucket and send you the access details by email so that you can try to reproduce the issue.
However, if the problem lies in how this provider implements S3 compatibility, I don’t think it’s worth spending too much time on it.
Initially, I thought the issue was related to the generic S3 configuration itself.

[girish]

@WiseMetalhead yeah, our CI tests object storages of DO, AWS, Vultr, Hetzner, minio .

For convenience, I can create a bucket and send you the access details by email so that you can try to reproduce the issue.

Yes please, if that's possible, would be great to understand where the problem is. I suspect this issue will get eventually hit by someone.

[WiseMetalhead]

@girish said in Cloudron 9.0 (beta) bug reports:

Yes please, if that's possible

I’ve sent the details from my email address vlishchenko****@gmail.com

[girish]

@jdaviescoates said in Cloudron 9.0 (beta) bug reports:

I wanted to turn of Automatic Platform updates, but leave App updates on:

Not yet, have to look into this in coming releases.

[robi]

@girish Can you simply add another column to the backups page that keeps track of how long each backup ran? Would be very useful!

[girish]

@Kubernetes said in Cloudron 9.0 (beta) bug reports:

After upgrading to Cloudron 9 I cannot send e-mails with Masquerading-Feature

Thanks for reporting! Indeed that flag is inverted! I have fixed the UI text.

Workaround: If you disable it there, I think you will see that masquerading works.