Skip to main content


!Friendica Support This is the Smarty thing, right? I find the error message (in the shell) ambiguous, is it complaining that it doesn't have access to the directory, or that it does but isn't supposed to?
in reply to Lorraine Lee

Now I seem to have the CSS (although in my other one there's a picture).

Now the problem is /register is 404
in reply to Lorraine Lee

`/.well-known/host-meta` is also 404, but there is no `.well-known` directory.
in reply to Lorraine Lee

@Lorraine Lee it does not exist but is only a "virtual directory" so to say. It is served by Friendica when one tries to access it (like the /profile/<username> page etc.).

Is the URL rewriting configured for your webserver (Apache, nginx, lighttps whatever you use) for the location Friendica is installed to?
in reply to Lorraine Lee

@Lorraine Lee Are you using Apache or nginx? In the first case, you have to copy the .htaccess-dist file in the root Friendica directory to .htaccess. In the second, you have to setup a vserver in the nginx configuration, you can see an example in the mods folder
in reply to Hypolite Petovan

Apache. `.htaccess` is copied from `.htaccess-dist`, as per instructions.
in reply to Lorraine Lee

is it on a subdomain? (like friendica.bla.blub ) just guessing..
then maybe a setting in .htaccess
in reply to xy..

@xy @Lorraine Lee I had a similar issue here. Check config/node.config.php, there are entries that may point to non-existing paths. Especially urlpath.
in reply to Lorraine Lee

OK that turned out to be this one: https://askubuntu.com/a/1046844/66051

Now to make sendmail sendmail
in reply to Lorraine Lee

It's alive! Barring yet-undiagnosed deficiencies.
This entry was edited (1 year ago)
in reply to Lorraine Lee

Oof! Seemingly had everything working (as is far from the case on the shared hosting astoundingteam dot com #friendica instance) but now "No space left on device." `df` and `df i` show plenty of space and inodes on all devices (less than 6% usage in all cases). Is there some counter somewhere in the system (Ubuntu 22.04/Apache 2.4.52/MariaDB 10.6.12/PHP 8.1.2-1ubuntu2.11) that maybe sometimes doesn't decrement when blocks or inodes free up?
in reply to Lorraine Lee

@Lorraine Lee @Lorraine Lee I have seen such a message like this. Maybe then your database has a limit made by your hosting provider?
in reply to tom s

This is way, way under all the limits. Perhaps traffic spikes lead to spikes in inode consumption somehow. I'm aware of an issue that apparently has to do with semaphores. https://github.com/friendica/friendica/issues/8200 Right now the instance seems to be operating normally, seems to have righted itself somehow.
in reply to Lorraine Lee

@Lorraine Lee Yes, i know this situation, there is nothing to do about this. But, maybe check once again, if your temp-directory in the Admin-Account is ok, because if this is wrong, youre restricting yourself.
in reply to tom s

@tom s @Lorraine Lee If you need some setting advises, I use these in #FPM pool's php.ini file:
[f.haeder.net]
chdir = /
group = www-data
listen.allowed_clients = 127.0.0.1
listen.group = www-data
listen.mode = 0660
listen.owner = www-data
listen = /var/run/apache2/php7.4-fpm_friendica.sock
php_admin_flag[log_errors] = on
php_admin_value[error_log] = /var/www/.../phptmp/fpm-php.friendica.log
php_admin_value[memory_limit] = 128M
php_admin_value[session.save_path] = /var/www/.../phptmp
pm = dynamic
pm.max_children = 10000
pm.max_requests = 10000
pm.max_spare_servers = 15
pm.min_spare_servers = 5
pm.status_path = /pool-status
request_slowlog_timeout = 20s
slowlog = /var/log/php-fpm/$pool-slow.log
user = vuXXXX

So for my instance 128M is totally enough. 😀
in reply to Lorraine Lee

@Lorraine Lee Friendica itself may generate temporary files starting with http- in your PHP's temporary directory (here outside htdocs, alsways and highly recommended). It also can write there the friendica.log file which with a high debug level can easily produce a very large file. On productive/life systems I would recommend maximum WARNING level (and errors included).
in reply to Lorraine Lee

@Lorraine Lee So df is not showing any mount point with 100% usage? Here it was the case with /tmp (tmpfs) which ran full by Proton (a WINE adoption for Steam gaming platform).
in reply to Roland Häder

@Roland Häder @Lorraine Lee /tmp is also a good reason. She could check, if the /tmp-path in the admin-account is really correct and writeable.
in reply to Lorraine Lee

The ENOSPC problem seems to have resolved itself, but now, new -dev version, and I get

PHP Fatal error: Uncaught Error: Class "Friendica\Core\Logger\Util\Introspection" not found in /var/www/html/static/dependencies.config.php:192
Stack trace:
#0 /var/www/html/index.php(32): include()
#1 {main}
thrown in /var/www/html/static/dependencies.config.php on line 192
#1
in reply to Hypolite Petovan

in reply to Lorraine Lee

@Lorraine Lee @Josephine Michelle Draus @Lorraine Lee Thanks, please run the following command to reset your local changes back to the develop branch:
git checkout -- .

And run the Composer command again afterwards.
in reply to Lorraine Lee

@Lorraine Lee Run a git checkout -- . each time in your deployment process to be sure that you have unmodified (maybe only access rights have changed?) and write for yourself a deployment script like I have done.
in reply to Roland Häder

@Lorraine Lee And please show us a git diff where you ran the git status command. Your local configuration file is not shown to us in diff, don't worry. Please see the .gitignore file.
in reply to Hypolite Petovan

Just nuked and reinstalled. Seems the database is where the instance lives.
in reply to Lorraine Lee

@Lorraine Lee Careful with reinstalling at the same sub-domain, you need to backup your database because of the private keys in there. Also watch out for files in config directory.

Lo, thar be cookies on this site to keep track of your login. By clicking 'okay', you are CONSENTING to this.