Had this exact thing a while back. docker-entrypoint.sh: command not found on the official mysql image is basically always

▎ one of 3 things on Railway:

▎

▎ 1. Volume mounted on top of a system path (90% of the time this is it)

▎ The docker-entrypoint.sh script lives in /usr/local/bin/ inside the mysql image. If your volume is mounted at

▎ /usr/local/bin, /bin, or / it shadows that script and the container has nothing to exec → crashloop.

▎

▎ Go to Service Settings → Volumes and check the mount path.

▎ - Should be: /var/lib/mysql

▎ - Anything else = thats the bug.

▎

▎ If its wrong you cant just change it in place without loosing data. You have two options:

▎ 1. Spin up a throwaway service (alpine or whatever) mount the same volume at /data, copy the raw files out / run mysqldump

▎ from a sidecar, then restore into a fresh volume mounted at /var/lib/mysql

▎ 2. If the data is disposable just delete the volume and re-create it with the right mount path

▎

▎ 2. Custom Start Command set to something weird

▎ Check Settings → Deploy → Custom Start Command. If theres something like docker-entrypoint.sh mysqld in there (without the

▎ full path) and PATH doesnt resolve it you get exactly this error.

▎ Fix: leave start command empty, let the image's default ENTRYPOINT run.

▎

▎ 3. Image got swapped

▎ If at some point the base image was changed from mysql:8 to a fork or custom build that doesnt ship the entrypoint

▎ script... thats it. Go back to mysql:8.0 or mysql:lts.

▎

▎ Easiest way to narrow it down — drop a screenshot of:

▎ - Service Settings → Source (the image)

▎ - Volumes (mount path)

▎ - Deploy → Custom Start Command

▎

▎ with those 3 its a one message fix.