Command history cleans up all the time

[ihor-sviziev]

Hi,

I found that when I'm rebooting my computer and then running warden env up - it all the time re-creates php-fpm container. As a result, the entire command history in the warden shell is getting lost.

It would be great if we could persist command history OR do not re-create the php-fpm container when it's not really needed (for instance new image was pulled, or some config has changed, etc.)

[Den4ik]

@ihor-sviziev if you will use warden env start container will not be recreated

[sprankhub]

+1 for persisting the command history.

[ihor-sviziev]

@Den4ik looks like warden env up re-creates php-fpm container because of changing the SSH_AUTH_SOCK variable after reboot.

I see this behavior on WSL2 + the following solution for ssh agent.
It seems if I'll switch to the solution with single static ssh auth sock - it shouldn't re-create containers.

if ! pgrep ssh-agent > /dev/null; then
  rm -f /tmp/ssh-auth-sock
  eval "$(ssh-agent -s -a /tmp/ssh-auth-sock)"
  ssh-add
else
  export SSH_AUTH_SOCK=/tmp/ssh-auth-sock
fi

Seems like on Mac OS we have static SSH_AUTH_SOCK, that's why we don't have such an issue, but we do have it on WSL2 and Ubuntu 20.04 (didn't look for a solution for clean Ubuntu)

[davidalger]

@ihor-sviziev Interesting workaround. That would explain why I hadn't experienced the behavior myself, primary driver being Mac OS only occasionally working on my Linux box. Using start vs up would prevent the re-create as well then, since it docker-compose doesn't look for changes in the container spec when only start has been requested. Sounds like we've got the best of both worlds: workaround to prevent re-create and persisted bash history thanks to the PR from @Den4ik making it happen.

[Den4ik]

@ihor-sviziev interesting. I didn't work on windows 10 years, but for now I really want to check it ))

[ihor-sviziev]

One interesting thing - today I had the same issue on Ubuntu 20.04, but with known hosts. After reboot - the php-fpm container was re-created and the ~/.ssh/known_hosts file was removed, so I wanted to check why that container was re-created.

I checked what was the output of warden env config before and after reboot - diff is interesting - that's IP of the traefik in the extra_hosts section:
https://github.com/davidalger/warden/blob/3cc14a24867191342564748962375a3d5d3c879b/environments/includes/php-fpm.base.yml#L9-L11

The last digit of IP was changed from 3 to 4 for some reason.
@Den4ik can we do the same thing for ssh known hosts as you did for bash history? AFAIK there might be some issues with permissions, right?

[ihor-sviziev]

I created a PR for fixing similar issue with the ~/.ssh/known_hosts file. #315