Dokploy using more disk space each day, not cleaning, how to fix

[bakirgdev]

Hello. Since I started using Dokploy, my disk usage went from 15 to 25+ GB in a few days for approx. 50 deployments done/built.

• I tried pruning via CLI using docker system prune -a -f --volumes but almost every time I skimmed a few MB or none
• I have turned on 'Daily Docker cleanup' funciton in Dokploy
• I have 1 Railpack and 2 plain compose apps
• I checked Docker volumes-images-containers-networks and everything is in order, there are no unnecessary resources...

I am not an expert wih Linux/Docker/Railpack caches and files, but I checked disk space usage and posted results below

I suppose this cache is from Railpack, but before exploring that I wanted to ask the community for some thoughts?

[bakirgdev]

WARNING: do not delete /var/lib/docker/overlay2/* directories, you will mess up Dokploy (like me :D)

[djsisson]

OK, but still run the command to see which current container is using what space.

Since the space is being used by containers that are in use. Doesn't matter if they are stopped or not.

[djsisson]

oh one other thing it might be, if you haven't altered the default logging options, and you have verbose services, then it could just be the logs from stdout

cd /var/lib/docker
find ./ -name *.json -size +1M -printf "%s %p\n"

this will list any json files over 1Mb

to fix you can switch to rotating log files

alter /etc/docker/daemon.json and then restart docker for changes to take effect

{
  "log-driver": "json-file",
  "log-opts": {
    "max-file": "3",
    "max-size": "10m"
  }
}

This will alter any new containers to use rotating log files, it won't alter any containers that have already been created (they would need to be recreated), you can of course overrule this in any compose file

[bakirgdev]

interesting! thanks for the info, but i highly doubt it's the logs...

because cca. 300 lines of logs are printed for each build across 50 of them. that doesn't add up to 10GB+ of mystery space used

now i migrated to Nixpacks (from Railpack), and get additional ~200Mb for ~20 deployments, thats okay for me

i believe the space problem is with Railpack, since the Dokploy config before and now is the same (except migrating to Nixpacks) and knowing Railpack is still v0

haven't investigated deeper since my concern is resolved

but again, thank you!

[djsisson]

its not build logs, but logs while the service is running,
so if you have a long running container, that logs everything to the terminal, it will add up over time

[bakirgdev]

oh, i get it know

but again, probably not, both of my services dont output logs. anyways this problem was observed for a few days of running time with max 10 visitors of the app, so there is no way the logs are the problem...

the command you provided doesnt give any results, so no files bigger than 1mb

[izniburak]

I have same issue.

I create some service to try something, then I remove them but disk space size not decrease. I tried Clean All option on Webserver->Space area but not change anthing.

[bakirgdev]

Are you using Nixpack/Railpack?

[izniburak]

for a few projects, yes. I tried them. But usually I use docker-compose or Dockerfile. Sometimes, I use a template on Dokploy as well.

[Paulitos]

Did you solve it? I had a heart attack today because of it. Everything went down because of ts.