Cannot access site after setting up XMPP; docs don't help. :(

GiantFrog

Hey, everyone. I’m in a bit of a panic. I did some upgrades and tried to get XMPP rolling today, but now Firefox just gives me a “We can’t connect to the server at mail.my.domain,” when I try to load it up. No such luck sending or receiving from mail clients, either.

I’ve tried what the docs said about adding ADDITIONAL_SERVER_NAMES to the conf file, and while it looks like that’s helped many people, no such luck for me. I also translated a German post on here from February and tried to follow their workaround. I deleted “data/conf/nginx/ZZZ-ejabberd.conf” and restarted mailcow’s nginx container, but nothing has changed. I keep rotating between “$http_host,” “MAILCOW_HOSTNAME,” and “mail.my.domain” in the mailcow site config file for my server’s main nginx install.

I can’t seem to find any logs to point me in the right direction, either. Visiting the site causes no entries in /var/log/nginx/error.log nor access.log. I also can’t find anything that looks useful in docker-compose logs.

Any ideas on how to get my email back online, somehow? I don’t have a clue where to go from here. I hate letting my friends down, and of course I bring all my important emails through my own server…

GiantFrog

I’ve installed a desktop environment. When I visit mailcow’s port at 127.0.0.1, I am presented with a 404 Not Found. Was hoping I could disable XMPP from there, but I’m still making an assumption that it’s even the problem.

GiantFrog

Deleting ZZZ-ejabberd.conf and restarting the nginx container let my log in at 127.0.0.1 and disable XMPP. I still cannot load anything using my actual domain from any device. This leads me to believe I made myself two issues in one! But at least one is dealt with.

I did modify my nginx config to better match the one here yesterday, so maybe something got botched there. I didn’t keep my old config file around because I am the April Fool. Just a shot in the dark. I once again am out of ideas and would graciously accept input from anyone who bothers to read through my tale here.

GiantFrog

For anyone arriving from Google, looks like it’s finally fixed. I’ve been changing things wildly, but I think what did it was my DNS records, not my nginx config. When adding the records for XMPP, I also noticed mailcow gave me some that I didn’t have, so I brought them all in. I’d bet it was because of this TSLA record I added even though I don’t have DNSSEC (does that cause issues?), but maybe undoing all the XMPP records did it for me.

So, that let me load up from my domain, and deleting the ejabberd config & running docker-compose restart nginx-mailcow let me get in and disable XMPP through the admin panel. Definitely separate issues, just created at the same time, so it threw me for a loop.

rsembera

The XMPP was removed from mailcow maybe a month ago - and yes this was an unpleasant surprise to me. Otherwise I had a similiar problem, you can solve it by defining ADDITIONAL_SERVER_NAMES in mailcow.conf.