fix(docker): openssh s6 supervisor crash loop in new builds #372
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Problem
New Docker images built from
Dockerfilecrash loop due to openssh s6 supervisor failing with exit code 256. Thes6-openssh-noopvolume mount at/etc/s6/openssh/runis not preventing the crash.Current workaround
Production runs on an older image (
gc7cfcf894b) where this didn't occur.Root cause
The s6 supervisor restarts openssh infinitely when it fails to bind port 22 (rootless UID 1000 cannot bind privileged ports). The noop script should prevent this but the new image may have a different s6 init sequence that runs before the volume mount takes effect.
Proposed fixes
START_SSH_SERVER=falseto environment (may not be respected by s6)Dockerfile.rootlesswith adjusted volume paths in composeRUN chmod +x /etc/s6/openssh/run && echo "exec sleep infinity" > /etc/s6/openssh/runto DockerfileClaude Opus 4.6 (1M context) noreply@anthropic.com