Issue Disk Space

[nate]

Hi,

I'm having some problems with disk space, the dev/root was 99% earlier, and was unable to even login to plesk control panel. I've removed some files to clear some space/restarted the server and things are working now.

At the moment the disk space looks like this:

/dev/root 20G 14G 4.4G 77% /
/dev/md2 201G 16G 175G 9% /home

var/www is all in dev/root, would it just be a case of moving it to /home and setting up a symlink?

Thanks

 

[Bitpalast]

I discourage using a symbolic link, because Plesk might expect a real link for some operations. Instead, follow this guide to move the virtual host location:
How to change Virtual Hosts location in Plesk for Unix

 

[nate]

Hi Peter,

I've setup a test Plesk Onyx server and run the commands in the link above, all the files have moved from /var/www/ to /home, and you can edit the files in the File Manager within plesk GUI fine however I'm getting the error message:

Error (111)Connection refused: AH02454: FCGI: attempt to connect to Unix domain socket /home/system/subdomaion.domain.com/php-fpm.sock (*) failed Apache error
Error 104.238.169.141 AH01079: failed to make connection to backend: httpd-UDS, referer: subdomain.domain.com

I've restarted the server, and run the 'plesk repair web' command, and it now seems to be serving the servers default domain holding page for that subdomain instead of the website, for another subdomain it's serving a 503 with the same error as above

Also looks like PHP is failing to start:
Unable to start service: Unable to manage service by phpinimng: ('start', 'php7.0-fpm'). Error: Job for php7.0-fpm.service failed because the control process exited with error code. See "systemctl status php7.0-fpm.service" and "journalctl -xe" for details. Failed to start php7.0-fpm service

Any suggestions?

Thanks

 

[Bitpalast]

The PHP start failure can be linked to missing configuration files. Check the journal entries as advised in the error messages for details. They normally explain what is causing it.

The other part:
Make sure that the path to the new virtual hosts does not contain uppercase letters or special characters.

You can try
# plesk repair fs <domain.tld>
or
# /usr/local/psa/bin/repair --restore-vhosts-permissions

Also try to reconfigure all domains with
# /usr/local/psa/admin/sbin/httpdmng --reconfigure-all

You should also compare file and directory permissions to this table:
Virtual Host Structure (Linux)

Your /vhost directory must have permissions 755, too, and ought to be owned by root:root. The partition and directory this resides under must be fully accessible by root, psaserv and psacln.

 

[nate]

Thanks for the info, we've managed to get it working. For some reason there was no .conf file for the subdomain in /etc/php/7.0/fpm/pool.d after changing the virtual hosts directory, by changing the name in hosting settings it generated a new one.

Also would you know why the hosting settings set as 'FPM application served by nginx' is different to the PHP settings 'FPM application served by Apache'? Changing PHP settings to nginx results in 'The settings were successfully updated.' but the setting reverts back to Apache (see attached image).

Thanks

 

[Bitpalast]

Please see this article to resolve the FPM Nginx issue:
Cannot switch PHP handler to FPM application served by nginx