Local deployment doesn’t replace old container

[VlexWest]

To Reproduce

Pushing changes builds a new image and container, but the old container keeps running, so updates never take effect.

1. Push code changes.
2. A new image/container is built, but the old container remains running.
3. On portainer the image of the old container is not being tagged as :latest anymore, but still running

Current vs. Expected behavior

Current
Old container stays running, new build isn’t deployed.
I need to:

1. Stop the application
2. docker system prune -af
3. Redeploy the application

clean cache toggle is not doing anything except the deployment takes longer, but in the end it's not replacing the old container.

Update: If I remove the old container manually (e.g. via Portainer) then the new container spawns out of nowhere..

Expected
Old container stops and is replaced by the newly built container.

Provide environment information

OS: Ubuntu 24.04 LTS
Dokploy: 0.24.4

Which area(s) are affected? (Select all that apply)

Local Development

Are you deploying the applications where Dokploy is installed or on a remote server?

Same server where Dokploy is installed

Additional context

I'm not sure, but I freshly added a remote server & I think since then this issue is occuring.
It's failing for remote & local deployments tho. :(

Will you send a PR to fix it?

No

[waltailji]

I'm experiencing this too. If I manually deploy, new container is spun up after build, but if it's auto deployed through a github webhook push, old container stays running after the build is complete.

I have dokploy running in it's own VM and my deployed services in another (remote) VM.

Dokploy v0.24.4 on Debian 12.

[scarruccio94]

Same problem here, pausing and starting the container solves it in my case

[VlexWest]

Maybe this could also be related to the issue somehow?

[waltailji]

Maybe this could also be related to the issue somehow?

I wouldn't think so. a) it's just a warning (probably about ENV vars being passed through into the container image), and b) the build completed successfully.

On my end, I tried Dockerfile and nixpacks build types and got the same results: auto deploy builds complete successfully on the deployment logs but the container doesn't deploy. If I click from Native to Swarm on the Logs tab, I can see a new container show up but with no entries in the log, and the old container is still running. So it seems like it is stuck in some half-deployed state in the swarm.

[paracha3]

I am seeing same thing.

Ideally i would like functionality, the likes of watchtower should be present in dokploy, where if i pushed new docker image to (any) registry, that dokploy would periodically check and update my running docker container with latest image

[Siumauricio]

I need more details

1. What application are you building? if you have a repository that would be great.
2. Show your general/builder section of your application or compose.
3. If you are using any environment variables, please share them.

[VlexWest]

1. It's happening with any application right now, I've tested it on:

• Docker compose
• Application
  • e.g. React.JS/Vue.JS site
  • or FastAPI instance

2. Sure, here's for the frontend (React.js)

• General:

3. Environment variables

If there are more details you need, please let me know, will share it with you.

Important note: the container replacement worked before & just suddenly stopped working across all applications that are running via dokploy, but I don't know what it could be related with.

[AriDeve]

Same issue here at the moment.
Both of my current applications create deployments successfully, but as described, don't stop the previous container, and swarm contains a new container without any logs.

To answer the above question for both of my applications:

API:

1. TypeScript with ElysiaJS

3. Nothing that would be relevant, only connection URLs

Frontend (Using github actions and a webhook):

1. SvelteKit instance

3. Also nothing relevant, just a connection URI for the API above

[josias-r]

As I am new to dokploy I was confused and thought I'm not understanding how its supposed to work :P.

In my case I have the same behaviour with a simple Dockerfile via ZIP Upload on the same Server.

I can Press Deploy, Rebuild, Reload and the deployments/building always happens successfully, but until I press Stop and then Start again, the actual running server does get replaced.

[actcian]

I'm experiencing the same issue. The auto-deployment via GitHub webhook isn't replacing the container after the build completes.

As a workaround, I have to manually stop the container and redeploy it.

[docloulou]

same here

[Siumauricio]

I cannot reproduce this issue

Simple express app: https://github.com/Dokploy/swarm-test

Click on deploy:

When you enter to the domain you will see...:

After I modify something in the code...

The new deployment after I push some code and the autodeploy start..

Check again the service and the container have the latest changes.....