Hello,
I am setting up a PHP-FPM–based system on Debian 12 using HestiaCP.
The architecture consists of Nginx (reverse proxy), Apache (backend), and PHP-FPM, with MultiPHP enabled.
During the setup, I tried creating a custom backend/template for PHP 7.4
(for example, something like PHP-7_4-video) to use on a per-domain basis.
However, in practice I encountered several issues, including:
PHP-FPM pools being removed or not properly regenerated after domain rebuilds
Socket conflicts (.sock files)
PHP-FPM services entering restart loops when MultiPHP was enabled
Occasional 500 Internal Server Errors
Because of these issues, I decided to abandon the custom backend approach and revert the domain back to the standard PHP-7_4 backend, after which the system became stable.
My Question ![]()
In systems running HestiaCP:
Is creating custom PHP-FPM backends/templates per PHP version a recommended approach?
Or should the PHP-FPM backend layer be kept as close as possible to the default configuration due to Hestia’s internal architecture?
More specifically:
In which scenarios do domain-level custom backends work reliably with HestiaCP?
When using MultiPHP, is there a recommended method to avoid backend/template and socket conflicts?
In this scenario, was there a different or more appropriate approach that should have been taken?
I would appreciate insights from experienced users and developers.