Nextcloud AIO - temporarily connection problems

ℹ️ Support
, , , , , ,
1

Nextcloud version : Nextcloud Hub 8 (29.0.4)
Operating system and version : Proxmox 8.2.4 / Debian Container / Docker
System: Nextcloud AIO

Hi community,

I recently switched from a dedicated installation of Nextcloud on a Ubuntu Server to a Proxmox Server. While I made a new installation I decided to run Nextcloud AIO as the best maintained setup.

Now I have a Proxmox Server with a AMD Ryzen CPU. The Proxmox has a Debian Base. The Nextcloud runs in a container, so also in a Debian setup. The internet connection is handled by a NGX Reverse Proxy.

After a few minor problems during the installation the NC runs good and fast.

To the issue:
At first I got sometimes a message from my PC - the NC would not be available. At first I ignored the message, because ever test of the connection showed no problem.

But now I had some situations while the website was not available. So I started the AIO admin page and all containers show green. But in the log of the Nextcloud AIO container this log entries where shown:

2024-08-23T10:22:32.317342774Z ✓ Reset callback url autodetect
2024-08-23T10:22:32.317372881Z Checking configuration
2024-08-23T10:22:32.317377170Z 🛈 Configured WOPI URL: https://evergreen.dynv6.net
2024-08-23T10:22:32.317380846Z 🛈 Configured public WOPI URL: https://evergreen.dynv6.net
2024-08-23T10:22:32.317384333Z 🛈 Configured callback URL: 
2024-08-23T10:22:32.317387679Z 
2024-08-23T10:22:37.360671471Z Failed to fetch discovery endpoint from https://evergreen.dynv6.net
2024-08-23T10:22:37.360701718Z cURL error 28: Connection timed out after 5002 milliseconds (see https://curl.haxx.se/libcurl/c/libcurl-errors.html) for https://evergreen.dynv6.net/hosting/discovery

While I could not reach my NC I tried if the internet connection or the reverse proxy have a problem, but my other websites (also proxmox containers with photoprism an immich) worked well.

After a few minutes the problem disappeared and NC worked again without any issue. I didn´t change anything, I didn´t restart any container, nothing.

I cannot see where the problem might be, every part (internet connection, reverse proxy, dyndns service) are involved while using my other websites, there cannot be the problem.
Is my setup with a docker inside a container a problem?

I don´t know if this other issue is a point, but if I restart the proxmox container or the docker containers NC need up to 15 minutes to start. The apache, nextcloud and the push notification container stay yellow for a long time and after the 15 minutes the start of NC is over and everything work well.

I would appreciate every tip you have for me.

Thanks

2

I’t definietely not the best approach to do nested virtualization but in general it “should” work.

I saw similar issues in the past when connections between hosts in a docker setup (NC to Collabora - same as your issue - /hosting/discovery) failed/delayed as there was some issue using DNS resolution hosted on a container on the same host as well. The symptom was - DNS took long time ( up to 10-15 sec each request) the solution was to add public domain to docker “alias” of the reverse proxy see Probably DNS help with NC Docker + Collabora + Wireguard tunnel - #5 by wwe for details.

This might be little different in your case with separate reverse proxy but sounds similar.

3

Hi and thank you for the hint. But I think I don´t understand what to do. I use a NC AIO, so there is no option to insert lines in the docker files, and the proxy is a NGX Proxy Manager on an other proxmox container, I think there cannot be a docker network on two different machines with different IP addresses.

Or did I understand it wrong?

4

with native Docker definitely not (Swarm or Kubernetes). but maybe you can achieve something similar pointing the public DNS to the reverse proxy IP (e.g. extra hosts)

5

Hello, I suspect that you have a DNS propagation problem (there are 2 different IPs queried from different DNS providers)
you can test it here f.e.:
https://www.nslookup.io/domains/evergreen.dynv6.net/dns-propagation/a/

so anytime you get the correct ip at random it works and anytime you query the wrong ip it obviously doesn’t work. I currently have the same problem with my domain, also on dynv6…
I suspect this is a general issue with ipv6 atm, but I there is no way to report the bug on their site since I only get an error message when trying to post there. If this won’t be fixed I we will have to look for an alternative …

6

Thank you for the hint, it sounds quite plausible.
I suspected dynv6 since a few months. Even with my old server I had sometimes problems with the availability, but the old setup was connected to more than one URL. So I had another URL from selhost and this one never showed a problem.
But that problem is not easy to fix, only way is to change to another dyndns provider but that would change the URL for all of my users.

Thank you
Homer

7

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.