Docker containers are often inaccessible externally via Nginx after system updates. Only method of recovery is to reboot OMV.

Hi folks.

I've been experiencing an issue that I have so far been unable to troubleshoot. I have about 10 Docker containers that I access externally using Nginx Proxy Manager. Namelt Sonarr, Radarr, Plex etc. Each of these containers are all accessible internally, but often after updating a container, the external access would stop responding and I would eventually get a 502 Bad Gateway error. This has happened to multiple containers and I have not been able to identidy a pattern.

Example: I update Sonarr via Portainer. The container reboots and the update is successful. Accesing Sonarr via interal IP address is perfect however, accessing externally via Nginx suddently times out. All other containers I have set up via Nginx are working fine. I've tried restarting Portainer and even stopping it completely, but this makes no difference, so I'm fairly sure that has nothing to do with it. I've also tried restarting Nginx containter and recreating Nginx host entry. The only thing that's ever worked is rebooting OMV. When I do this, sometimes another container will start doing the same thing. So Radarr might then become inaccessible externally, but Sonarr starts working again.

I should emphasise that this does not happen every time I update a container. Only sometimes, and it seems to be a different containter for reasons that are so far unknown to me.

I'm asking here first because the only solution I've found so far is to reboot OMV completley, then either they all start working again, or another cointainer becomes unresponsive externally. I am not sure where to troubleshoot this. Really appreciate any suggestions. By the way I'm currently using the latest OMV 7.0.5-1 build. Thank you!

Zitat von chente

I use NPM and access several containers externally but I don't use Portainer, all containers are configured with the compose plugin. Containers are automatically updated every week by the compose plugin. I have never had access issues after updates on any of the containers.

Could this problem be related to Portainer only? Have you searched for similar problems on github in Portainer?

Thanks for the reply. Portainer was my first suspicion too however, I tried both restarting and stopping Portainer completely from the console and even when it was stopped, the issue was still there. I then thought that maybe theres an issue with Nginx, but restarting the Nginx container didn't fix it either. Pinging also doesn't help because all pings seem to complete ok. I've had a look at the logs in Sonarr in this latest issue and there's nothing there of note. It's been a real head scratcher for me to troubleshoot! Thanks.

Zitat von chente

I would try setting up the containers with the compose plugin, just to completely rule out that it's a Portainer issue. If you have the containers implemented through a stack, it is just copy and paste.

Thank you. I'll give this a try. It's such a stange and annoying problem

I think this is now sorted. It looks like it was Nginx Proxy Manager that was causing the issue. I deleted this entirely and set it up from scratch again. It's working great now, and if it still works after the next batch of Radarr/Sonarr updates, then that should be it sorted. I have no idea why this was happening but either way, it seems to be good now, fingers crossed!