Scheduled reboot fails due to missing shutdown binary

[markozajc]

System information

Host distribution: Debian Trixie
Seesaw version: 0.10.3
Python version: 3.9.18 (main, Nov 29 2023, 05:59:34) [GCC 12.2.0]
Installation method: Docker, image b4becc328563 from atdr.meo.ws/archiveteam/warrior-dockerfile

Expected behaviour

After a week of uptime AT warrior restarts itself.

Actual behaviour

After a week of uptime AT warrior shuts itself down, but does not come back up. Logs show a failed attempt to run shutdown, which might be what causes this.

Logs

-- snip --
2024-02-29 01:26:49,213 - seesaw.warrior - INFO - Running for more than 7 days. Time to schedule a reboot.
2024-02-29 01:26:49,263 - seesaw.warrior - DEBUG - Update warrior hq.
2024-02-29 01:26:49,264 - seesaw.warrior - DEBUG - Warrior ID ''.
2024-02-29 01:26:49,267 - seesaw.warrior - DEBUG - Update project.
2024-02-29 01:26:49,267 - seesaw.warrior - DEBUG - Check project has update telegram
2024-02-29 01:26:49,272 - seesaw.warrior - DEBUG - git fetch
2024-02-29 01:26:49,410 - seesaw.warrior - DEBUG - Select project telegram
2024-02-29 01:26:49,726 - seesaw.warrior - DEBUG - False
2024-02-29 01:34:30,152 - seesaw.warrior - INFO - Runner has finished.
sudo: shutdown: command not found

[tech234a]

Did you include --restart=on-failure in the creation command?

See https://wiki.archiveteam.org/index.php/ArchiveTeam_Warrior#Installing_and_running_with_Docker

[markozajc]

I did, here is my full docker run command:

docker run --detach --name archiveteam-warrior --label=com.centurylinklabs.watchtower.enable=true --restart=on-failure -v /var/cache/warrior-projects:/home/warrior/data/projects -e CONCURRENT_ITEMS=1 -e DOWNLOADER=mzajc -e SELECTED_PROJECT=auto --publish 10084:8001 atdr.meo.ws/archiveteam/warrior-dockerfile

[stepney141]

I'm experiencing the same issue. As a workaround, I'm running docker start archiveteam-warrior via crontab every 7 days, but obviously this isn't a proper fix. Is there any way I can help contribute to a permanent solution?

[TheTechRobo]

Maybe try replacing --restart=on-failure with --restart=always? The container might not be exiting with a failure status code.

[stepney141]

Maybe try replacing --restart=on-failure with --restart=always? The container might not be exiting with a failure status code.

It seems to successfully work! Here are results of docker logs archiveteam-warrior in my environment:
https://pastebin.com/pi0EQD8z

[gladykov]

Not to create more issues, dumping info here:

https://github.com/ArchiveTeam/grab-base-df lacks shutdown binary which was removed from Debian (part of upstart package)

grab-base-df does not use systemd so we cannot use systemctl to poweroff https://www.debian.org/releases/stable/amd64/ch08s01.en.html

If image with systemd will be used it should be possible by altering this https://github.com/ArchiveTeam/seesaw-kit/blob/development/seesaw/warrior.py#L910

Any other solutions seems hacky