Containers still running, but not available through localhost port

[TheFox]

Please see this comment.

[fcrisciani]

@TheFox, did you try to force the use of IPv4 with curl -4 <url>

[TheFox]

I didn't try curl -4 127.0.0.1:32000 but telnet 127.0.0.1 32000. But I'll try. Do you guess this is an issue with IPv6?

[fcrisciani]

@TheFox sometimes depending on the OS config, ipv6 is prioritized on ipv4 so curl tries ipv6 but timeout. just an idea to avoid chasing ghosts :)

[TheFox]

@fcrisciani I'm sure that IPv4 is used. Since I use the IP address 127.0.0.1 in the Apache reverse proxy no IPv6 will be used.

[fcrisciani]

@TheFox does the container ever restart and the failure is maybe happening after that?

[TheFox]

@fcrisciani Nope. The container does not restart ever. But when I manually restart the container it works fine again. Restarting the container or even the Docker daemon is no acceptable solution.

I commented to after ±20 hours: containers still running, but not available through port in the first place because the description is very similar to my problem. The container is running fine, after ~2 to 3 days (or so, I don't know exactly) the container isn't accessable via the --publish port but through the internal Docker Network ip address (from docker inspect -f '{{ .NetworkSettings.IPAddress }}' CONTAINER_ID).

What's even the main purpose (or advantage) to have an internal IP address and a public when I can use the internal anytime anyway?