And journal.txt attached. Note that I had to do this a number of times to reproduce the race. The last time was the time that I got "Job for cups.service canceled.". There were one or two other occurrences that I missed.
Interestingly, "cups.service: Start request repeated too quickly." appears on a successful run. Where I got the "canceled" error, nothing unusual is logged.
Using my example, I did:
systemd-analyze set-log-level debug cups/org. cups.cupsd cups/org. cups.cupsd
systemctl stop cups.service
touch /var/cache/
sudo rm /var/cache/
systemctl stop cups.service
systemd-analyze set-log-level info
journalctl -b > /tmp/journal.txt
And journal.txt attached. Note that I had to do this a number of times to reproduce the race. The last time was the time that I got "Job for cups.service canceled.". There were one or two other occurrences that I missed.
Interestingly, "cups.service: Start request repeated too quickly." appears on a successful run. Where I got the "canceled" error, nothing unusual is logged.