Ubuntu 20.04 --> Docker 20.10.2 issues

Support
#1

Openemr AWS Standard Cloud Installation: 6-0-0
Ubuntu 20.04.3 OS
Docker 20.10.2 cli

So, over the last few weeks I have had to recover and relaunch openemr instances when making changes (whether installation or other endeavors). I kept making assumptions that perhaps the security groups were wrong or something configuration wise had been changed (subnets, route tables, ENIs, etc.) This last time I realized that docker is the problem. It turns out that if there is an issue where the EC2 instance has to be restarted that the docker daemon for docker 20.10.2 does not automatically restart any running containers.
Sudo docker container ls will not recognize any container as being installed.
Sudo docker ps -a lists the container as exited. After a number of unsuccessful attempts to restart the container, looking through docker logs, and looking online, I used sudo systemctl status docker and found that docker enabled ipv6.
Any solutions to this problem?

#2

I hadn’t heard about this issue – but OpenEMR Standard runs on Ubuntu 16.04 and it pulls in Docker’s personal repo, not Ubuntu’s. If you upgraded the instance to 20 you may not have the version of Docker you’d prefer to have.

#3

Thank you for your response and this is exactly what happened because of functionalities/unintended by-products of docker along the way.

The work around for any version of docker that starts to throw errors because of auto-switching to ipv6, rendering your container momentarily useless, is the following:

  1. roll back to a docker version that was functioning as expected
    -or-
  2. Run a command that willl force ipv6 to ipv4 as demonstrated here:
    docker run --publish “0.0.0.0:80:80” --publish “0.0.0.0:443:443” --detach [container]
1 Like
#4

That’s very good to know – thanks! I really need to take a look at the launch script and get it 20-compatible. Given that this will probably bite me then, it’s nice to have one problem identified.