Database corrupted because of disk full - Pre-check free disk space before upgrading db schema

Hi,

I’ve updated my Duplicati installation from 2.0.5.1-beta to 2.0.6.1-beta.I’ve got two jobs configured, the database sizes are (1,9 GByte) and (34 MByte). After the upgrade, I ran “check files”. The small job ran fine while running the large job with the big database failed, resulting in a corrupted database.

Looking closer the cause was “disk full” at the partition where /root/.config/Duplicati/*.sqlite was stored.


My suggestion/Feature Request: Please calculate twice the sqlite DB size plus - let’s say - 5 % of the file size and check if this is free on the partition where the file is stored (“df -h”) check. If there’s not enough free disk space, warn the user and do NOT touch the sqlite database.

Good idea. I’d also like to see Duplicati clean up some of the old sqlite “backups” after a while.

1 Like

Hi!

Is there an option to put the database on an alternative location? I’ve had an issue running Duplicati on a Synology where it “ate” the /dev/md0 partition because /root sits there. On Synology this is the partition where DSM is installed (Synology’s OS). I now “solved” the issue by running in a Docker container and externalize /data to somewhere else.

@ErikZ How to change the local database file location? - #2 by kees-z

1 Like