Very interesting article. I have to wonder how much benefit a user would get from letting parts of the backup run concurrently with other operations. Aren’t backups usually just run relatively infrequently on a fixed, automated schedule?
not familiar with this system, but for large setups, having the maintenance operations block usage can reduce the backup window enough that it impacts your desired scheduling pattern
> I have to wonder how much benefit a user would get from letting parts of the backup run concurrently with other operations. Aren’t backups usually just run relatively infrequently on a fixed, automated schedule?
This is good point, i think it depends a lot on the repository size and storage latency. Some other similar backup tools can take many hours to do a garbage collection on a huge repository, so I'm trying to avoid that if I can. In the end it hasn't been a huge complexity burden and I hope I can keep it that way.