I wanted to setup a apache2 template to use a directory under public_html/public and I couldn’t get it to work with nginx or apache2 directly. htaccess method always rendered a 500 error. well in the midst of changing something via the templates and then selecting the template with the gui I ran into an ssl error so I tried to run v-add-letsencrypt-host again and then every site I have setup across 4 other users started showing 403 errors everywhere. trying to check the default and revert everything didn’t work so I started checking the logs and found this:
[autoindex:error] [pid 6566] [client XXX.XXX.XXX.XXX:39970] AH01276: Cannot serve directory /var/lib/roundcube: No matching DirectoryIndex (index.html,index.cgi,index.pl,index.php,index.xhtml,index.htm) found, and server-generated directory index forbidden by Options directive
google turned up nothing helpful. literally trying to keep my changes isolated I still messed it up… not sure what I did wrong but if you could either point me in a direction to maybe recovery from this mess quickly awesome otherwise maybe a guide on how to allow document root overrides?
Thanks again for responding to my email so quickly and pointing me back to the forums. maybe someone else will benefit from a public answer too.
the roundcube error is very specific and I doubt it’s directly related to your problem at all.
to recap: you created a custom .tpl and .stpl file in the template folder(s), then selected it from within the control panel and after that everything went to hell?
my guess: most likely there is/was something wrong in the template, could be as simple as a tiny typo. after changing the template either apache2 or nginx could not reload properly because it detected the faulty config. did you check if your apache is running at all ?
the individual log files per domain should give more hints, what’s failing, I’d expect you see warnings about the missing upstream. try to restart apache2 from your shell and if it fails use journalctl -xe to see the cause, this often gives a good idea, if it’s config related.
other than that, just set everything back to the standard/default templates (in which you hopefully did not mess around ;-)) and make sure that nginx and apache are restarted properly afterwards.
I think there was one other thing I tried to do which was change the ssl to support another alias but never could get it to accept it fully. Also I tried going into settings where it was server>configure>webmail and clear out the webmail alias and clicked save.
first thing I tried was deleting the templates and switching back the admins one site to the default. I also did immediately check the systemctl status apache2 and nginx which there was a configuration error at the time with the default during one of the changes that I immediately was able to resolve by changing the template and restarting apache2. nginx -t reported fine each time that it was working even right before the crash. the logs were outputting quite a bit of information just that error showed up on every individual log file per domain also but there were a batch of 10 or so logs I should have captured them in my snapshot I made after it went all downhill.
I have it running now with defaults via setting up a new server again and got the htaccess method to work ok for the document root change via apache.
at this point if you need me to look into the bug further let me know in the next few days or I will just scrap the backup of the broken instance.
