Comparison of Different LiteSpeed PHP suEXEC Setups https://download.litespeedtech.com/images/logos/litespeed/litespeed-logo.png 2019-10-01 15:11:15 LiteSpeed Web Server offers a number of different PHP suEXEC setups for various shared hosters' needs and goals. The following table is offered as a simple guide to choosing the PHP setup right for you and your users.

Comparison of Different PHP Process Modes

Choose the process mode that is right for you and your users

LiteSpeed Web Server offers a number of different PHP process modes to meet the various needs and goals of shared hosting providers. The following table offers a simple guide to choosing the PHP process mode that is right for you and your users.

All of the following LiteSpeed PHP process modes completely support the suEXEC execution method. (each PHP process running as the owner of the virtual host's document root). More detailed information on these setups can be found in each setup's documentation.

Setup Description Recommended for
Worker mode Standard PHP with LSAPI. Creates new processes when needed. Shared hosting providers with limited resources that want custom php.ini files.
Daemon mode Forks all processes from a constantly running daemon process for improved speed and efficiency. All shared hosting providers who do not need custom php.ini files.
ProcessGroup mode Each user account gets a separate process group with a single parent process used to fork child processes off of when needed. Shared hosting providers who want to leverage extra RAM into more powerful opcode caching for improved PHP performance.

Detailed Feature Comparison

Worker mode Daemon mode ProcessGroup mode
Forks processes instead of spawning new processes No Yes Yes
Number of parent processes 0 1 One for each process group (each user)
Customize number of processes per account? Yes Yes Yes
Can enable only for certain virtual hosts? No No Yes
Custom php.ini supported?
(Including CageFS custom php.ini)
Yes No Yes
Opcode caching No Shared throughout server Dedicated per-user