[Bug]: swarm doesn’t kill off old containers

[hkd987]

Description

Using docker swarm old containers are still active and running after a deploy.

Minimal Reproduction (if possible, example repository)

Deploy a nodejs app using docker swarm

Exception or Error

Old containers still running after a deploy & logs not being output.

Version

Lastest, however this has been going on for some time.

[hkd987]

Currently I’m testing a less than ideal workaround where I stop the containers and then deploy again.

[hkd987]

Stopping all services before deploy resolves this issue although this isn't ideal and prevents rolling updates.

[garyhtou]

This is a shot in the dark, but maybe the solution is to add the --prune flag when calling docker stack deploy.

coolify/app/Jobs/ApplicationDeploymentJob.php

Line 1157 in 1c12cb9

executeInDocker($this->deployment_uuid, "docker stack deploy --detach=true --with-registry-auth -c {$this->workdir}{$this->docker_compose_location} {$this->application->uuid}"),

From what I'm observing, Coolify generates a new service name per deployment. Without the --prune flag, Docker will keep the old services running even if they are no longer referenced in the new stack/docker-compose file.

I've thrown together PR #5141 in case it is the solution. However, I haven't gotten around to testing it yet.

[djsisson]

Out of interest is this seperate from dockers rollback feature for swarm, as it will keep your old containers incase you need to rollback

[garyhtou]

I believe this is separate. From my understanding, the rollback feature keeps old images. When you decide to rollback, it creates a new deployment using the old image. It doesn't keep old containers running.

[djsisson]

If this is generating a diff uuid everytime then it will still create a new stack and leave old one running

Maybe switch to name?

Oh its the id of the service inside the app that changes, OK yeah this should be fine then