Hello @makemrproper The following log line made me think this might be temporary MySQL issue: Apr 13 11:02:47 SQLSTATE[HY000] [1129] Host 'fd00:c107:d509::1' is blocked because of many Apr 13 11:02:47 connection errors; unblock with 'mysqladmin flush-hosts' If you see such a log entry, it might be a pointer that it might be just the MySQL service. @makemrproper said: I should have tried that before contacting you Please never hesitate to ask in the forum when you have any issues, like you have done here. We have seen many issues and logs and can use that knowledge to quickly get a picture of a situation. The good old "have you tried turning it off and on again" is something that can be attempted almost always. We are always happy to help since there should be no issue in the first place.

Well, forget this, it actually works. - It works for custom icons, text and others. However it simply refuses to pick up a different stylesheets. I am probably not looking in the right place.

@joseph Thanks for checking - This all sounds good. It is indeed an old instance which has been running, so it could be something that did not work during the migration.

Also here, thanks for sharing

Ok, after same more debugging with @girish it turned out that this solution works as intended. My Email-Services stopped working which was due to a change in local name resolution. Restarting the Email-Server and updating some configurations in my Firewall solved the issue.

@warg thanks for the report. fixed now in latest package.

@luckow I think the feature is still "experimental" per the docs - This solution is currently still in testing you could experience performance issues. The second solution listed in the section suggests enabling "Enable higher security image uploads". I can confirm that option works. Since, we enable .htaccess in the apache configs, directory indexes are disabled as well. I think if people still want local_secure, they can always fix up env as required but there's probably a reason that the upstream does not use this as the default.

BTW: kind of same behaviour for nextcloud. There is a UID in the user settings, which you can only see with /app/code/occ user:setting in the terminal. And because I switched from internal Cloudron LDAP to an external LDAP I got a new UID for all my users. The strange behavior from nextcloud is: Login with user:pass works. (the schema is firstname.lastname as username). But internally nextcloud adds a _RANDOMNUMBER to it. In my case a stephan.luckow is internally changed to stephan.luckow_3096 Problem is: all my calendar entries are based on stephan.luckow_3096 and not on stephan.luckow.

@girish Got it. So as long as I stick with the Cloudron LDAP, all is well. Thanks for the explanation, and also for the update to the docs.

@girish Only Bookstack. Yourls installed just fine.

I have updated bookstack package to support optional cloudron authentication. If you do a fresh install, you will get the option to install without cloudron auth. After that, you can enable external registration.

@subven thank you I will give it a try.