New to the forum but I feel I’ve played around with Duplicati on Linux enough to give some input and suggestions for future releases and/or installation methods.
I am running duplicati on linux manjaro installing the aur package - all good. The package updater two or three days ago alerted me of an update so I proceeded to do that. I had the duplicati instance open in the browser… it lost connection and then came back, as expected when the server restarts. However, it restarted with NO JOBS present. Needless to say it made my heart sink - I instantly went to work and made sure that backup files were still present on the destination. They were, so I proceeded to do the work to figure it out.
I restarted the duplicati service as root etc and to no avail - I was looking at the forums postings about Windows and service level and all that stuff and files - no dice. I finally got the bright idea to check the /usr/lib/systemd/system/duplicati.service and sure enough the username and group lines were re-entered. I REMOVED the entries again and BAM the backup tasks that were once in Duplicati before the upgrade were present again in Duplicati.
ExecStart=/usr/bin/mono /opt/duplicati-latest/Duplicati.Server.exe --webservice$
I am wondering if we can change the way that this package either installs or the defaults to allow for Duplicati to run as root, if started in root. IF I run this as root with those fields in there it takes over and runs it as duplicati and it doesn’t have access to the entire drive. ALSO, it was running as root and all and then all of a sudden after the update it runs as Duplicati user and all the jobs are lost. Obviously the jobs are per user (or user defined) which is fine but for the average user to rely on Duplicati to backup after an update (at least on Linux) it could be frightening experience.
On that note, if I couldn’t get the jobs back could I recreate the jobs and point to the backup location that was on the NAS or Google Drive? I didn’t go down that route yet… AND yes I understand that there is a backup config option but who does that that.
Just my thoughts on the subject -hope we can have some conversation and/or back and forth to the methodology and perhaps why it is the way it is.