Portainer Stacks - stopping and starting again causes the application to start from scratch

[eduoliveira13]

Ask a Question!

I was starting to play with passbolt the other day and creating its stack on portainer and everything worked fine. However, i noticed that if I stop the stack and start it again it will cause passbolt to start from scratch again even though all volumes for persistent data are there and intact.
Has anyone else experienced the same? Not quite sure at this point if it is a portainet issue and a passbolt issue.

I am using the same docker-compose-ce.yaml file that passbolt provides.

Thanks

[Nick-Portainer]

Hello, can you share the compose? Is it using named volumes or anonymous volumes? If they are not named then you will get a new volume every time you re-deploy it

[eduoliveira13]

Hi Nick,

thanks for getting back to me.
This happens with either passbolt and the docker compose indicated here https://www.passbolt.com/docs/hosting/install/ce/docker/

I have tried also with a few nginx oficial docker compose and every time I atop a stack on portiainer 2.21 it will also remove the containers. The all use named volumes and they are not removed, so the data is there. However, with passbot there is an aggravated situation since it will also cause the application to start from scratch as if the database was not here and all users, etc will be gone.

It first I thought that it would be passbolt only but according to portianer dock stopping a stack should not remove the containers which in my case is happening with anything in a stack created using portainer and pasting the docker compose there (even with named volumes)

Again, thanks for your insight and suggestions

[Nick-Portainer]

Could you walk me through the exact steps you follow from deploying the stack to how you’re stopping it? Please also provide as much detail as possible about your environment, including server OS, Portainer Server version, Docker version and whether this is running on an agent environment or directly on the Portainer server. This information will help me replicate your setup accurately.

[eduoliveira13]

Absolutely,

In the meantime, I got the issue with the data not being retained! I found a typo (uppercase) on my code which caused 2 volumes for the mariadb to be created. everytime the stack was stopped it created another one which an empty db. After fixing that the issue with the persistent data was resolved.

As for the containers being removed when the stack is stopped, that is still unclear to me as the documentation says that it was not supposed to be removed.

Anyway, here is additional information you have requested:

HOST OS: Ubuntu 24.04.1 LTS
Engine Version: 27.3.1 (Installed manually on the Host.. NOT using the one offered with the Host OS install)
Portainer Version: 21.1.4 BE

Docker compose file for the for Passbolt:

name: passbolt
services:
  db:
    image: mariadb:10.11
    restart: unless-stopped
    environment:
      MYSQL_RANDOM_ROOT_PASSWORD: "true"
      MYSQL_DATABASE: "passbolt"
      MYSQL_USER: "passbolt"
      MYSQL_PASSWORD: "PASSHIDDEN"
    volumes:
      - database_volume:/var/lib/mysql
 
 
  passbolt:
    image: passbolt/passbolt:latest-ce
    restart: unless-stopped
    depends_on:
      - db
    environment:
      VIRTUAL_HOST: passboltdocker.domain.local
      VIRTUAL_PORT: 443
      VIRTUAL_PROTO: https
      CERT_NAME: passbolt.domain.local
      APP_FULL_BASE_URL: https://passbolt.domain.local/
      DATASOURCES_DEFAULT_HOST: "db"
      DATASOURCES_DEFAULT_USERNAME: "passbolt"
      DATASOURCES_DEFAULT_PASSWORD: "PASSHIDDEN"
      DATASOURCES_DEFAULT_DATABASE: "passbolt"
      EMAIL_DEFAULT_FROM_NAME: "COMPANY-Passbolt"
      EMAIL_DEFAULT_FROM: "[email protected]"
      EMAIL_TRANSPORT_DEFAULT_HOST: "SMTP-IP"
      EMAIL_TRANSPORT_DEFAULT_PORT: "25"
 
    volumes:
      - gpg_volume:/etc/passbolt/gpg
      - jwt_volume:/etc/passbolt/jwt
      
 
    command:
      [
        "/usr/bin/wait-for.sh",
        "-t",
        "0",
        "db:3306",
        "--",
        "/docker-entrypoint.sh",
      ]
    ports:
      - 80
      - 443
 
volumes:
  database_volume:
  gpg_volume:
  jwt_volume:

====================

Steps to stop the stack:
Login to portainer, Select the node, Select Stacks, Click on the name of the Stack to open its details, Click on "Stop this Stack"

See some print screens below
Stack Running originally

Stack Stopped - containers are not listed anymore and were removed without warning

Starting the container again recreates the containers (see create date that is different from the first print screen

[Nick-Portainer]

Can you check that when you stop the stack that the volumes are still there that were created for this deployment?