• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Issue Enabling of HTTP/2 whilst LiteSpeed was in use has broken things

Daveo

Basic Pleskian
Server operating system version
Debian 12.4
Plesk version and microupdate number
18.0.57 #4
Our other server admin clicked 'Enable HTTP/2' in Advisor whilst we had LiteSpeed enabled and in use.

Since then, we have removed LiteSpeed, but somethings are not working properly.

ie:

1. xenForo forums, clicking or browsing to their admin.php file downloads the file instead of executing it
2. Invision forums, hoovering over a user's name displays 'There is a problem loading this content' and clicking their name generates a 500 error with the following error in the server logs files...
352868#0: *739 FastCGI sent in stderr: "PHP message: PHP Fatal error: Cannot declare class IPS\core\modules\front\members\hook128, because the name is already in use in /var/www/vhosts/*****/httpdocs/init.php(931) : eval()'d code on line 3" while reading response header from upstream

We do think it's the enabling of HTTP/2 whilst LiteSpeed was on the server that has caused this issue, but we have been unable to fix or resolve the issue ourselves. Any ideas on how to fix this, preferably without having to reinstall the system.
 
Please try,

If you have completely removed the LiteSpeed, have you re-enabled the NGINX?

Once re-enabled the NGINX, Please restart both NGINX, Apache and any FPM services in use under Tools & Setting > Services Management.

Then,
# plesk repair web domainname.tld
 
I have re-enabled NGINX and all services have been restarted, as we rebooted the server, but the issue is still happening.
 
1. xenForo forums, clicking or browsing to their admin.php file downloads the file instead of executing it

# plesk repair web domainname.tld should have fixed this issue as this command re-writes the default configuration files for the vhost using Apache and NGINX.

Do you have any additional settings in Additional Apache directives or Additional nginx directives?

Also please check the .htacess file.
 
(1): Maybe "short open tag" is set to "off", while it should be on as the script seems to be using <? instead of <?php. If it is set to "off", scripts won't be interpreted but downloaded as text.

(2): Maybe a PHP version issue. It's not a web server issue, but rather code cannot be interpreted correctly, so maybe the PHP version is too new or the script version is too old.
 
1. Short open tag is set to 'on'.

2. PHP version is 8.1 which is required for the script.

We've tried switching back from NGINX to Apache and removing any NGINX directives and the issue still persists.

.htaccess file isn't used when using NGINX, but we did check it and it's the default .htaccess file for either of the affected scripts.
 
You could go through extra configurations that may have been done to the /etc/httpd and /etc/nginx folders. Normally users who experiment with other server software like Litespeed tend to also experiment with settings, because they are hoping to increase the server's speed. Then all kinds of tweaks are done and forgotten, which results in all kinds of problems. Fix: Reinstall the whole server and restore a full Plesk backup.
 
Is the only way to fix any possible changes in /etc/apache2 & /etc/nginx is to reinstall server ?

I know that the other admin tried to enable HTTP/2 whilst LiteSpeed was in use and I fear this is what's caused the problem, as you can't use (and shouldn't be able to enable) HTTP/2 whilst using LiteSpeed. He said he clicked the Enable HTTP/2 button in Advisor.
 
Back
Top