This happens when your Apache image has more than one MPM module (prefork/event/worker) enabled at the same time — Apache allows exactly one. It's almost always caused by a Dockerfile step that a2enmods a second MPM without first disabling the one already active in the base image.

If you're using php:apache (mod_php, non-thread-safe) with CodeIgniter, you need mpm_prefork and nothing else. Add this near the top of your Dockerfile, before installing PHP extensions:

RUN a2dismod mpm_event mpm_worker 2>/dev/null; a2enmod mpm_prefork

If instead you're on PHP-FPM (php:fpm + a separate Apache/nginx layer), you want the opposite — mpm_event, not prefork:

RUN a2dismod mpm_prefork 2>/dev/null; a2enmod mpm_event

Check whether your Dockerfile (or a Nixpacks-generated one, if you're not using a custom Dockerfile) has an a2enmod line for a second MPM — that's the one to remove or reorder.

AH00534: More than one MPM loaded is an Apache configuration issue rather than a Railway deployment issue.

Check your Apache config and make sure only one MPM is enabled (mpm_event, mpm_prefork, or mpm_worker). This usually happens when multiple MPM modules are loaded from different config files or when a custom image enables another MPM on top of the default.

A couple of questions that would help narrow it down:

Are you deploying with a Dockerfile or using Nixpacks?

Could you share your Dockerfile (or Apache config), especially anything that enables modules with a2enmod or modifies /etc/apache2?

The build succeeding but the container failing at startup is consistent with Apache refusing to start because of the duplicate MPM configuration.

gdg