Repository is already locked, forget cmd fails but systemd service is marked successful

[varac]

I often get this issue that the repository is locked, and the forget cmd fails:

backup-all.sh[577930]: + [[ 0 -eq 0 ]]
backup-all.sh[578466]: + systemd-cat -t all
backup-all.sh[578467]: ++ date -u '+%Y-%m-%d %H:%M:%S'
backup-all.sh[578465]: + echo '2022-10-19 08:19:03 OK'
all[578466]: 2022-10-19 08:19:03 OK
backup-all.sh[578468]: + /usr/local/bin/restic forget --path / --keep-last 10 --prune
backup-all.sh[578469]: + systemd-cat -t all
all[578469]: unable to create lock in backend: repository is already locked by PID 294851 on sancho.varac.net by root (UID 0, GID 0)
all[578469]: lock was created at 2022-10-18 00:00:05 (34h19m6.088127533s ago)
all[578469]: storage ID 361f61fc
all[578469]: the `unlock` command can be used to remove stale locks
backup-all.sh[577930]: + [[ 1 -eq 0 ]]
backup-all.sh[578486]: + systemd-cat -t all
backup-all.sh[578487]: ++ date -u '+%Y-%m-%d %H:%M:%S'
backup-all.sh[578485]: + echo '2022-10-19 08:19:11 ERROR'
all[578486]: 2022-10-19 08:19:11 ERROR
backup-all.sh[577930]: + rm -f /var/run/restic_backup_all.pid
backup-all.sh[577930]: + exit
systemd[1]: restic-all.service: Deactivated successfully.
systemd[1]: Finished Backup all using restic.

However, in this case the script ends with an exit code of 0 and therefore the systemd service is marked as successful, which makes it very hard to notice when the forget fails:

❯ systemctl status restic-all
○ restic-all.service - Backup all using restic
     Loaded: loaded (/lib/systemd/system/restic-all.service; static)
    Drop-In: /etc/systemd/system/restic-all.service.d
             └─env.conf, untethered.conf
     Active: inactive (dead) since Wed 2022-10-19 10:19:11 CEST; 36min ago
TriggeredBy: ● restic-all.timer
   Main PID: 577930 (code=exited, status=0/SUCCESS)
        CPU: 14min 4.131s

It can be reproduced manually as well:

root@sancho:~# /usr/local/bin/backup-all.sh 
+ [[ -z '' ]]
+ MODE_TAG='--tag manual'
+ /usr/local/bin/restic backup / --tag manual --exclude-caches --exclude-file /etc/backup.exclude --exclude-if-present .exclude_from_backup
+ systemd-cat -t all
+ [[ 0 -eq 0 ]]
+ systemd-cat -t all
++ date -u '+%Y-%m-%d %H:%M:%S'
+ echo '2022-10-19 09:04:12 OK'
+ /usr/local/bin/restic forget --path / --keep-last 10 --prune
+ systemd-cat -t all
+ [[ 1 -eq 0 ]]
+ systemd-cat -t all
++ date -u '+%Y-%m-%d %H:%M:%S'
+ echo '2022-10-19 09:04:14 ERROR'
+ rm -f /var/run/restic_backup_all.pid
+ exit

root@sancho:~# echo $?
0

[DO1JLR]

Yes, I am currently not entirely happy with the behaviour of the script and have been planning to revise it for some time. Unfortunately, I haven't got around to it yet.

If you feel like taking a closer look, I would of course appreciate it.