I started piloting kopia as a replacement for our borgbackup environment.
We have a fairly large CephFS (about 1TB daily incrementals). I created a kopia repository server instance on a remote (25ms RTT) machine with 8 x 22TB drives configured in a ZFS RAIDz. The initial backup took about a week to complete and then incrementals were decently fast (under 10 hours).
Following similar maintenance issues as in #3563, I decided to run kopia snapshot fix invalid-files --verify-files-percent=100 --commit. It's now running since 6 days and backups are blocked. The current repository size on disk is 100T.
The main questions are how are issues like this normally supposed to be handled on very large repositories? If we go live with this and later run into an issue that requires running snapshot fix or other repository maintenance jobs, how can downtime to the entire backup system be avoided? Any tips on how to handle/manage very large repos?