Updated message about --new flag and systemd

Eric Gustavsson 2021-05-20 22:48:49 +02:00
parent 5167668d5e
commit e0cd279105

@ -6,12 +6,12 @@ Podman is easier to run in systemd than Docker due to its daemonless architechtu
$ podman run -d --name vaultwarden -v /vw-data/:/data/:Z -e ROCKET_PORT=8080 -p 8080:8080 vaultwarden/server:latest
54502f309f3092d32b4c496ef3d099b270b2af7b5464e7cb4887bc16a4d38597
$ podman generate systemd --name vaultwarden
# container-foo.service
# container-vaultwarden.service
# autogenerated by Podman 1.6.2
# Tue Nov 19 15:49:15 CET 2019
[Unit]
Description=Podman container-foo.service
Description=Podman container-vaultwarden.service
Documentation=man:podman-generate-systemd(1)
[Service]
@ -26,14 +26,21 @@ PIDFile=/run/user/1000/overlay-containers/54502f309f3092d32b4c496ef3d099b270b2af
WantedBy=multi-user.target default.target
```
You can provide a `--files` flag to dedicate a specific file to output the systemd service file to. With this we can enable and start the container as any normal service file.
You can provide a `--files` flag to tell podman to put the systemd service into a file. With this we can enable and start the container as any normal service file.
```sh
$ systemctl --user enable /etc/systemd/system/container-vaultwarden.service
$ systemctl --user start container-vaultwarden.service
$ systemctl enable /etc/systemd/system/container-vaultwarden.service
$ systemctl start container-vaultwarden.service
```
## New container every restart
If we want to create a new container every time the service starts we can edit the service file to contain the following:
If we want to create a new container every time the service starts we can use the `podman generate systemd --new` command to generate a service file that recreates containers
```sh
$ podman generate systemd --new --name vaultwarden
```
If you're using an older Podman, you can edit the service file to contain the following instead:
```sh
[Unit]
Description=Podman container-vaultwarden.service
@ -60,6 +67,6 @@ If you want the container to have a specific name, you might need to add `ExecSt
# Troubleshooting
## Debugging systemd service file
If the host goes down or the container crashes, the systemd service file should automatically stop the existing container and spin it up again. We can find the error through `journalctl --user -u container-vaultwarden -t 100`.
If the host goes down or the container crashes, the systemd service file should automatically stop the existing container and spin it up again. We can find the error through `journalctl -u container-vaultwarden -t 100`.
Most of the time the errors we see can be fixed by simply upping the timeout in podman command in the service file.
Most of the time the errors we see can be fixed by simply upping the timeout in Podman command in the service file.