Closed qchn closed 1 year ago
Hi there,
we've exactly the same issue, but using https://github.com/hoellen/docker-nextcloud/pkgs/container/nextcloud (behind a Traefik). The issue was identified in cases, where the system hosting the nextcloud stack is running in DMZ and traffic is directed with DNAT from external to it. We've two setups, where we spotted this. We are also running several other instances with the same stack on other places without DNAT and it's flawlessly working there. As we do NOT use this helm chat and even other docker images, I would guess it's some kinda general issue?
Okay ... forget about DNAT. It's working on some Nodes and on some not.
Okay ... I looked a bit deeper into it. On the setups, where it's broken, it works with nextcloud 21 and below. With nextcloud 22 and higher its broken. There was some UI (and maybe also other code) changes to the interface of creating users. So it is most probably, that there got something broken with those changes.
Have just encountered this issue on a fresh instance of 24.0.4-apache running on ECS
@ja573 have a look into https://github.com/nextcloud/server/issues/32432#issuecomment-1222143890! I would guess the circles app is activated in your setup. Try to disable it!
@ja573 have a look into nextcloud/server#32432 (comment)! I would guess the circles app is activated in your setup. Try to disable it!
Thank you, it worked for me. I had the same issue.
@ja573 have a look into nextcloud/server#32432 (comment)! I would guess the circles app is activated in your setup. Try to disable it!
Thank you so much for the hint, @waja! Indeed, disabling the circles app did help in my case.
Hello, I experience the following bug with a fresh NextCloud deployment on k8s using helm-chart.
Infos about my test environment:
Kubernetes setup: vanilla k8s with containerd runtime (3 master, 3 worker nodes) Kubernetes version: v1.23.5
Bug can be triggered with at least the following NextCloud: 23.03, 23.02 & 22.3.2
Steps to reproduce:
Note: Creating new users without adding them to a group works flawlessly!
Log from pod when bug is triggered:
Any hints on how to fix this?
Regards qchn