I have a few questions about repository server (referred to as RS from now on). It looks to me that its docs are a bit lacking (or I did a poor job of googling it =) ).
My usecase is hosting backup server that a handful of hosts will backup to. Those will be both mine and semi-trusted devices (some relative’s PCs for example).
Is it safe to make RS open to the internet through nginx reverse proxy? Assuming good passwords, of course.
How is maintenance run when RS is active? It looks like any of the clients can’t do that, because it’s supported on direct repos only. Does it mean that I should periodically run it on my own with cron or similar tools? Or is it run by the RS itself?
Lets assume that I have RS, and two clients (A and B). Clients only see RS and can’t see bare repository. RS and A have good enough clock sync, but B doesn’t - it’s wildly inaccurate (either on accident or on purpose) to the point that clocks are out of sync for hours or even days.
Docs say that data corruption is possible, but what kind? Obviously I wouldn’t rely on data that B created, but is data that A sends safe from corruption, or is there risk to a repo as a whole?
How much trust is given to clients of RS? In other words - can clients misbehave to a point of corrupting data (snapshots or blobs related to snapshots) of other clients?
I’m migrating from borg. It had this option:
–exclude-if-present NAME exclude directories that are tagged by containing a filesystem object with the given NAME
so, basically, I could place an empty ‘marker’ file with name of choosing (.nobackup for example) and borg would skip entire directory that contains it.
Is there any alternatives to it in kopia? I see that there is gitignore like system, but that requires files to have rules, empty ones won’t do as they are essentially no-op for gitignore.