The error is "start failed: no command specified, set cmd or entrypoint"
This was the working configuration before.
I am not sure if the issue is related to this specific release since i have skipped some releases.
perhaps because at least entrypoint.sh it exists at the root of the layer, but docker 7.23 does not see it. you must either overwrite it or rebuild it, explicitly specifying /entrypoint.sh .
I found I had to update my routers running 7.19.x through 7.20 and/or 7.21 to get to 7.22/7.23 because there were some subtle container configuration changes (my assumption; it worked, so I didn't dig into the cause).
If i set entrypoint=/usr/bin/start.sh
Container starts, but after ~10 seconds it ends with exit code 0
2026-06-11 23:42:59 container,info,debug pihole_2025080.tar: *** start
2026-06-11 23:42:59 container,info,debug pihole_2025080.tar: *** started <DNSMASQ_USER,root> <FTLCONF_database_maxDBdays,2> <FTLCONF_dns_upstreams,8.8.8.8;8.8.4.4> <FTLCONF_webserver_api_pass
word,mysecurepassword> <TZ,Europe/Amsterdam> /usr/bin/start.sh
2026-06-11 23:43:08 container,info,debug pihole_2025080.tar: *** exited with status 0
I am able to start pihole like this:
1 )set entrypoint = sleep 3600
2) once container is up and running, I manually start pihole by executing "./usr/bin/start.sh" ,
Of course, this is stupid, and it is not a solution. I will check if I can downgrade RouterOS back to 7.19.2.
FROM alpine:latest
RUN echo "http://dl-cdn.alpinelinux.org/alpine/edge/testing" >> /etc/apk/repositories
RUN apk add --no-cache pihole
CMD ["pihole-FTL", "-f"]
Recreating the images would be the least desirable solution for me.
Some of my containers contain manually inserted data that has accumulated over time.
If i have to rebuild contianers, ony restoring that data would be a time consuming process.
I would prefer to explore options that preserve the existing containers and data before start with full rebuild.
