[bug] update broke setup again

[h4de5]

is there a way to find out which directive in a config is no longer supported?

I have about 40 proxy-hosts, some (probably many, because they are very alike) include directives that seem to be no longer working. now after startup it goes into an endless loop of writing/checking/deleting nginx configs. I can't access the UI, nginx is in emergency mode, and the logs are filled with useless information like:

2026/01/17 12:03:08 [alert] 459#459: sendmsg() failed (9: Bad file descriptor)
2026/01/17 12:03:08 [alert] 459#459: worker process 2435 exited with fatal code 2 and cannot be respawned

it constantly states that config files cannot be read, because it seems the whole directory is cleared and rewriten all the time.

I can't pinpoint the wrong configs, because for some reasons, they switch all the time.

so far all I can do is reset the whole database - but I would really like not to.

[devedse]

The configs are generated on startup, they don't load, so the application won't start. This last update broke things for a lot of people.

I think the most important thing for @Zoey2936 to focus on is fixing the code so that it'll at least start if there's errors in the generated .conf files. It currently doesn't.

[h4de5]

the backup shows the same problem, the old docker image seems not to be avaiable anymore. So I removed the database, got a working but completly empty UI, rethought some of my life decisions, put back the old database once more and for some reason the startup loop was now gone.

but: all proxy hosts delivered self-signed certificate, cert update says successful, but doesn't update any files on the disk, disabling/enabling proxy hosts delivers at least the old certificate correctly and thats where I leave it. I'll take that lesson and try to switch to a more stable setup..

[mageunic]

is there a way to find out which directive in a config is no longer supported?

I have about 40 proxy-hosts, some (probably many, because they are very alike) include directives that seem to be no longer working. now after startup it goes into an endless loop of writing/checking/deleting nginx configs. I can't access the UI, nginx is in emergency mode, and the logs are filled with useless information like:

2026/01/17 12:03:08 [alert] 459#459: sendmsg() failed (9: Bad file descriptor) 2026/01/17 12:03:08 [alert] 459#459: worker process 2435 exited with fatal code 2 and cannot be respawned

it constantly states that config files cannot be read, because it seems the whole directory is cleared and rewriten all the time.

I can't pinpoint the wrong configs, because for some reasons, they switch all the time.

so far all I can do is reset the whole database - but I would really like not to.

Hi, do you use Authentik or specificly this method forward_scheme?

[jacksonm36]

Hello i do use authentik and rn its completly busted for some proxied hosts. my amp website is bricked vaultwarden behaves weirdly . mailcow works nextcloud works as well "fix" was posted on discord which helped me to at least make my password manager work again.

discusion link: https://discord.com/channels/840285826020933662/1462050853652529334

[mageunic]

I created earlier an issue ( transformed into discusion) about proxy buffering directive ...
#2543

it seems there is a big issue about cpu usage and reponse time ( in my case)...

[h4de5]

I read about forward_scheme in #2548 - and no, as far as I know, I don't use forward_scheme - at least in the produced proxy_host files I can't find that directive.
what I see now, is that the responses are very unstable. every few minutes the services are not reachable, websockets are closed and return a few moments later. there is no CPU/memory spike, and no errors in the logs :\

[Zoey2936]

you can try again with the last release, now only broken hosts will be disabled and others stay enabled, sorry this was an issue by me. But why they go offline is something else