Auto start container

方式: Add the container to systemd

原因" Podman uses the standard fork/exec model

Step1. Start CT

./start-ct.sh

podman ps         # 查看 container 是否成功運行

Step2. 建立 .service file

podman generate systemd \
    --new --name CT_NAME > CT_NAME.service

"--new"

Using this flag will yield unit files that do not expect containers and pods to exist.
Instead, new containers are created based on their configuration files.

"--new" only works on containers and pods created directly via Podman.
It does not work on containers or pods created via the REST API or via podman kube play.

i.e.

/etc/systemd/system/nginx.service

[Service]
Restart=on-failure
ExecStart=/usr/bin/podman run \
        --cidfile=%t/%n.ctr-id \
        --cgroups=no-conmon \
        --rm \
        --sdnotify=conmon \
        --replace \
        -dit \
        --privileged \
        --volume /sys/fs/cgroup:/sys/fs/cgroup:ro \
        ...
ExecStop=/usr/bin/podman stop --ignore --cidfile=%t/%n.ctr-id
ExecStopPost=/usr/bin/podman rm -f --ignore --cidfile=%t/%n.ctr-id
Type=notify
NotifyAccess=all

no "--new"

ExecStart=/usr/bin/podman start nginx
ExecStop=/usr/bin/podman stop -t 10 nginx
ExecStopPost=/usr/bin/podman stop -t 10 nginx

Step3. Enable autostart

mv nginx.service /etc/systemd/system/

Check

systemctl list-unit-files | grep nginx

nginx.service                              disabled

systemctl enable nginx --now

systemctl status nginx

● nginx.service - Podman container-nginx.service
   Loaded: loaded (/etc/systemd/system/nginx.service; enabled; vendor preset: disabled)
   Active: active (running) since Wed 2022-10-26 13:27:16 HKT; 5s ago

Stop Container

If you try to run "podman stop nginx",

  => the container will be restarted by systemd because of to the “Restart=on-failure” policy.

More info: systemd restart

systemctl stop nginx

Auto start rootless container

 * Steps 詳見上一段

Step1. Start CT

CT=openai-proxy

./start_docker.sh

#!/bin/bash

name=openai-proxy
image=openai-proxy:v3
base=/home/vhosts/openai-proxy
code=$base/code
port=4431

#### Code

podman run --name $name -dit \
  --privileged \
  --volume /sys/fs/cgroup:/sys/fs/cgroup:ro \
  --cap-add=AUDIT_WRITE \
  --hostname $name \
  -v $code:/home/code \
  -p $port:$port \
  $image

Step2. 建立 .service file

podman generate systemd  --new --name $CT > ${CT}.service

mkdir ~/.config/systemd/user/ -p

mv ${CT}.service ~/.config/systemd/user/

Step3.

# To relaod systemd a non-root user

systemctl --user daemon-reload

# enable a systemd service for a non-root user

systemctl --user enable $CT --now

Created symlink /home/openai-proxy/.config/systemd/user/default.target.wants/openai-proxy.service →
 /home/openai-proxy/.config/systemd/user/openai-proxy.service.

# Check status

podman ps | grep $CT             # 查看是否 start 了

systemctl --user status $CT

Step4. rootless 才須要此步驟

loginctl enable-linger openai-proxy

詳見: loginctl