podman

Minha bolha, bolha minha

E a bolha está de pé. Ou quase isso.

A primeira semana em operação foi erro 502 o tempo todo. Achei que o problema era como estava funcionando pelo systemd. Então criei um serviço novo só pra ela.

  
# /etc/systemd/user/podman-compose@.service

[Unit]
Description=GoToSocial as container service
StartLimitIntervalSec=0


[Service]
Type=simple
User=helio
Group=helio
#WorkingDirectory=/home/helio/gotosocial
ExecStart=/home/helio/gotosocial/entrypoint.sh start
ExecStop=/home/helio/gotosocial/entrypoint.sh stop
Restart=always
RestartSec=30

[Install]
WantedBy=default.target

Depois achei que era o enviroment. Comentei a parte de WorkingDirectory, como pode ser visto acima. Também troquei o podman-compose up por esse script entrypoint.sh.

  
#! /usr/bin/env bash

GOTOSOCIAL_DIR="/home/helio/gotosocial"

start_gotosocial() {
    echo "Starting gotosocial" 
    cd $GOTOSOCIAL_DIR
    /usr/bin/podman pull docker.io/superseriousbusiness/gotosocial:latest
    /usr/bin/podman pull docker.io/library/postgres:latest
    /usr/bin/podman-compose down
    sleep 5
    /usr/bin/podman-compose up
}

stop_gotosocial() {
    echo "Stopping GoToSocial"
    cd $GOTOSOCIAL_DIR
    /usr/bin/podman-compose down
}

case $1 in
        start) start_gotosocial ;;
        stop) stop_gotosocial ;;
        restart) $0 stop
             sleep 30
                     $0 start
                     ;;
        *) echo "Unknown option: $1" 
           exit 1
esac

Os podman pull estavam antes no serviço do systemd. Joguei tudo pra dentro do script. E o resultado foi: 502.

Então comecei a considerar que tinha feito algo errado no compose.yml.

  
services:
  gotosocial:
    image: docker.io/superseriousbusiness/gotosocial:latest
    container_name: gotosocial
    user: 1000:1000
    networks:
      - gotosocial
    environment:
      # Change this to your actual host value.
      GTS_HOST: bolha.linux-br.org
      GTS_DB_TYPE: postgres
      GTS_CONFIG_PATH: /gotosocial/config.yaml
      # Path in the GtS Docker container where the
      # Wazero compilation cache will be stored.
      GTS_WAZERO_COMPILATION_CACHE: /gotosocial/.cache
      ## For reverse proxy setups:
      GTS_TRUSTED_PROXIES: "127.0.0.1,::1,172.18.0.0/16"

      ## Set the timezone of your server:
      TZ: Europe/Stockholm
    ports:
      - "127.0.0.1:8080:8080"
    volumes:
      - data:/gotosocial/storage
      - cache:/gotosocial/.cache
      - ~/gotosocial/config.yaml:/gotosocial/config.yaml
    restart: unless-stopped
    healthcheck:
      test: wget --no-vebose --tries=1 --spider http://localhost:8080/readyz
      interval: 10s
      retries: 5
      start_period: 30s
    depends:
      - postgres
        
  postgres:
    image: docker.io/library/postgres:latest
    container_name: postgres
    networks:
      - gotosocial
    environment:
      POSTGRES_PASSWORD: *****
      POSTGRES_USER: gotosocial
      POSTGRES_DB: gotosocial
    restart: unless-stopped
    volumes:
      - ~/gotosocial/postgresql:/var/lib/postgresql
    ports:
      - "5432:5432"
    healthcheck:
      test: pg_isready
      interval: 10s 
      timeout: 5s
      retries: 5
      start_period: 120s

networks:
  gotosocial:
    ipam:
      driver: default
      config:
        - subnet: "172.18.0.0/16"
          gateway: "172.18.0.1"
volumes:
  data:
  cache:

Nada de muito fantástico. Um postgres rodando junto com um gotosocial. Algumas configurações de proxy, que é o nginx da máquina, e é isso. E continuava o 502.

Mas se eu entrava na máquina, e rodava uma sessão de tmux e dentro dela chamava o podman-compose up, daí tudo funcionava. Dei então uma olhada no erro.

  
Oct 15 10:16:56 mimir entrypoint.sh[1895291]: podman-compose version: 1.0.6
Oct 15 10:16:56 mimir entrypoint.sh[1895291]: ['podman', '--version', '']
Oct 15 10:16:56 mimir entrypoint.sh[1895291]: using podman version: 4.9.3
Oct 15 10:16:56 mimir entrypoint.sh[1895291]: ** excluding:  set()
Oct 15 10:16:56 mimir entrypoint.sh[1895291]: ['podman', 'ps', '--filter', 'label=io.podman.compose.project=gotosocial', '-a', '--format', '{{ index .Labels "io.podman.compose.config-hash"}}']
Oct 15 10:16:56 mimir entrypoint.sh[1895303]: time="2025-10-15T10:16:56+02:00" level=warning msg="RunRoot is pointing to a path (/run/user/1000/containers) which is not writable. Most likely podman will fail."
Oct 15 10:16:56 mimir entrypoint.sh[1895303]: Error: default OCI runtime "crun" not found: invalid argument
Oct 15 10:16:57 mimir entrypoint.sh[1895291]: Traceback (most recent call last):
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/bin/podman-compose", line 33, in 
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     sys.exit(load_entry_point('podman-compose==1.0.6', 'console_scripts', 'podman-compose')())
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/lib/python3/dist-packages/podman_compose.py", line 2941, in main
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     podman_compose.run()
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/lib/python3/dist-packages/podman_compose.py", line 1423, in run
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     cmd(self, args)
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/lib/python3/dist-packages/podman_compose.py", line 1754, in wrapped
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     return func(*args, **kw)
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:            ^^^^^^^^^^^^^^^^^
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/lib/python3/dist-packages/podman_compose.py", line 2038, in compose_up
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     compose.podman.output(
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/lib/python3/dist-packages/podman_compose.py", line 1098, in output
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     return subprocess.check_output(cmd_ls)
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/lib/python3.12/subprocess.py", line 466, in check_output
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:   File "/usr/lib/python3.12/subprocess.py", line 571, in run
Oct 15 10:16:57 mimir entrypoint.sh[1895291]:     raise CalledProcessError(retcode, process.args,
Oct 15 10:16:57 mimir entrypoint.sh[1895291]: subprocess.CalledProcessError: Command '['podman', 'ps', '--filter', 'label=io.podman.compose.project=gotosocial', '-a', '--format', '{{ index .Labels "io.podman.compose.config-hash"}}']' returned non-zero exit status 125.
Oct 15 10:16:57 mimir systemd[1]: gotosocial.service: Main process exited, code=exited, status=1/FAILURE
Oct 15 10:16:57 mimir systemd[1]: gotosocial.service: Failed with result 'exit-code'.
Oct 15 10:16:57 mimir systemd[1]: gotosocial.service: Consumed 1.481s CPU time.
Oct 15 10:17:27 mimir systemd[1]: gotosocial.service: Scheduled restart job, restart counter is at 1280.
Oct 15 10:17:27 mimir systemd[1]: Started gotosocial.service - GoToSocial as container service.
Oct 15 10:17:27 mimir entrypoint.sh[1895707]: Starting gotosocial
Oct 15 10:17:30 mimir entrypoint.sh[1895781]: podman-compose version: 1.0.6
Oct 15 10:17:30 mimir entrypoint.sh[1895781]: ['podman', '--version', '']
Oct 15 10:17:30 mimir entrypoint.sh[1895781]: using podman version: 4.9.3
Oct 15 10:17:30 mimir entrypoint.sh[1895781]: ** excluding:  set()
Oct 15 10:17:30 mimir entrypoint.sh[1895781]: podman stop -t 10 postgres
Oct 15 10:17:31 mimir entrypoint.sh[1895781]: exit code: 0
Oct 15 10:17:31 mimir entrypoint.sh[1895781]: podman stop -t 10 gotosocial
Oct 15 10:17:31 mimir entrypoint.sh[1895781]: exit code: 0
Oct 15 10:17:31 mimir entrypoint.sh[1895781]: podman rm postgres
Oct 15 10:17:31 mimir entrypoint.sh[1895781]: exit code: 0
Oct 15 10:17:31 mimir entrypoint.sh[1895781]: podman rm gotosocial
Oct 15 10:17:31 mimir entrypoint.sh[1895781]: exit code: 0
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: podman-compose version: 1.0.6
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: ['podman', '--version', '']
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: using podman version: 4.9.3
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: ** excluding:  set()
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: ['podman', 'ps', '--filter', 'label=io.podman.compose.project=gotosocial', '-a', '--format', '{{ index .Labels "io.podman.compose.config-hash"}}']
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: podman volume inspect gotosocial_data || podman volume create gotosocial_data
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: ['podman', 'volume', 'inspect', 'gotosocial_data']
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: podman volume inspect gotosocial_cache || podman volume create gotosocial_cache
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: ['podman', 'volume', 'inspect', 'gotosocial_cache']
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: ['podman', 'network', 'exists', 'gotosocial_gotosocial']
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: podman create --name=gotosocial --label io.podman.compose.config-hash=4f4b10e0c67c04b7b4f2392784b378735d4378d9d411f1405cf3819c6207bd1a --label io.podman.compose.project=gotosocial --label io.podman.compose.version=1.0.6 --label PODMAN_SYSTEMD_UNIT=This email address is being protected from spambots. You need JavaScript enabled to view it. --label com.docker.compose.project=gotosocial --label com.docker.compose.project.working_dir=/home/helio/gotosocial --label com.docker.compose.project.config_files=compose.yaml --label com.docker.compose.container-number=1 --label com.docker.compose.service=gotosocial -e GTS_HOST=bolha.linux-br.org -e GTS_DB_TYPE=postgres -e GTS_CONFIG_PATH=/gotosocial/config.yaml -e GTS_WAZERO_COMPILATION_CACHE=/gotosocial/.cache -e GTS_TRUSTED_PROXIES=127.0.0.1,::1,172.18.0.0/16 -e TZ=Europe/Stockholm -v gotosocial_data:/gotosocial/storage -v gotosocial_cache:/gotosocial/.cache -v /home/helio/gotosocial/config.yaml:/gotosocial/config.yaml --net gotosocial_gotosocial --network-alias gotosocial -p 127.0.0.1:8080:8080 -u 1000:1000 --restart unless-stopped --healthcheck-command /bin/sh -c 'wget --no-vebose --tries=1 --spider http://localhost:8080/readyz' --healthcheck-interval 10s --healthcheck-start-period 30s --healthcheck-retries 5 docker.io/superseriousbusiness/gotosocial:latest
Oct 15 10:17:36 mimir entrypoint.sh[1895920]: exit code: 0

A parte final, com podman create, é o systemd reiniciando o serviço. O problema está on início, onde há um crash de python: subprocess.CalledProcessError: Command '['podman', 'ps', '--filter', 'label=io.podman.compose.project=gotosocial', '-a', '--format', '{{ index .Labels "io.podman.compose.config-hash"}}']' returned non-zero exit status 125.

Eu entrava na máquina e rodava o comando pra ver o resultado:

  
❯ podman ps --filter 'label=io.podman.compose.project=gotosocial' -a --format '{{ index .Labels "io.podman.compose.config-hash"}}'
4f4b10e0c67c04b7b4f2392784b378735d4378d9d411f1405cf3819c6207bd1a
4f4b10e0c67c04b7b4f2392784b378735d4378d9d411f1405cf3819c6207bd1a

E mostrava os containers rodando (porque tinha sido reiniciados pelo systemd). Eu ficava com aquela cara de "ué!?".

via GIPHY

No início do erro, tem essa outra mensagem aqui: Error: default OCI runtime "crun" not found: invalid argument . Então fui olhar se era algum problema nesse crun. E está instalado (acho que veio como dependência do podman.

  
❯ which crun
/usr/bin/crun
❯ dpkg -S /usr/bin/crun
crun: /usr/bin/crun

Busquei sobre erros do GoToSocial mesmo. E nada.

Olhando pra todo lado tentando descobrir o que poderia ser, reparei em outro erro: msg="RunRoot is pointing to a path (/run/user/1000/containers) which is not writable. Most likely podman will fail." .

Isso soou promissor. Então de repente o pointing path não estava disponível pra escrita. Poderia ser... systemd? Com isso eu comecei a buscar algo relacionado com timeout ou user logout. Acabei encontrando o artigo abaixo:

https://www.reddit.com/r/podman/comments/192kqgf/podman_container_stops_after_user_logout/

Nesse artigo alguém comenta que pode ser uma opção de container linger. Segui a referência que tinha sobre isso.

ttps://www.freedesktop.org/software/systemd/man/latest/loginctl.html#:~:text=Enable%2Fdisable%20user%20lingering%20for,to%20run%20long%2Drunning%20services

loginctl? Faz até sentido isso. Mas o podman não deveria descrever isso na documentação? Então fui buscar e achei isso aqui:

https://docs.podman.io/en/latest/markdown/podman-system-service.1.html

Pra deixar bem ilustrado onde aparece a referência de linger na documentação:

Algo que é vital pra funcionar como serviço aparece como... exemplo??? Os caras tão de brincation uite me.

Mas no fim era isso mesmo. Bastou um sudo logictl enable-user helio pra ter o container rodando depois que eu saio da sessão.

Se eu tivesse decido rodar com docker compose, eu provavelmente não teria o mesmo problema uma vez que roda com o privilégio de root. Então fica mais essa lição aqui. E mesmo tendo lendo a documentação, sempre aparecem alguns pontos que a porra da documentação só dá um peteleco em cima e dentro dos exemplos ainda por cima.

Mas está funcionando. Minha bolha, bolha minha.

Os artigos mais lidos de 2024

podman

podman

Minha bolha, bolha minha

Estatísticas

Imagem aleatória