14 days ago
Since Friday I've been hitting the deployment Apache error "More than one MPM loaded.". Redeploying previous successfully deployed commits, resulted in the same error. Local builds and starts of the service worked fine.
I've applied following workaround to fix this by manually changing Apache to listen to port 8080 (it seems that $PORT is set to 8080 for some reason) and added an entrypoint script that disables mpmevent and mpmworker at container startup. I've also found another user pointing the same issue but with php:8.4: https://station.railway.com/questions/urgent-all-wordpress-sites-deployed-to-07dee2ef.
My project ID is: 0d1ad3d1-41ff-4fcd-b3a2-d3a82a1fa15d
Question is were there any changes that would justify the above from the Railway side or did you observe or receive any similar issues with other clients? Hopefully I've provided enough info, first time doing this.
Thanks in advance,
3 Replies
Additionally tried building and starting the container on Railway and locally with bare library/php:8.2-apache image. This resulted in the same MPM error on Railway and no issues locally.
13 days ago
hey @kozmo2854
yeah this sounds like something changed on railways end with how they handle ports or apache configs. the fact that your local builds work fine but railway keeps failing is a pretty clear sign
the $PORT being set to 8080 is normal for railway since they use that as the default internal port. but apache should be handling that automatically with the base php images so something definitely broke
couple things to try:
check if you have any PORT env variables set in your railway project settings that might be conflicting
try adding this to your dockerfile if you havent already:
RUN a2dismod mpm_event mpm_worker
RUN a2enmod mpm_prefork
or if youre using the base image directly without a custom dockerfile you might need to create one now just to handle the mpm stuff
the other post you linked with php:8.4 having the same issue is interesting. seems like maybe railway pushed some infrastructure changes that affected how apache initializes in containers
@Railway team any insight here? this looks like it started affecting multiple users around the same time
also @kozmo2854 what does your current entrypoint script look like? might be able to simplify it depending on what youre doing
i also have this issue - My PHP Apache Docker app won't deploy - container starts then immediately crashes with error: AH00534:
apache2: Configuration error: More than one MPM loaded
No code changes made
Old deployments that previously worked now also crash
Tried different regions - same result
Using php:8.2-apache base image
Anyone else seeing this? Is there a known issue with the php:8.2-apache Docker image? - need urgent immediate help !