This is a companion discussion topic for the original entry at https://restic.net/blog/2020-09-13/rest-server-0.10.0-released/
So does using separate repos for each user generally work in 0.10.0 or not? If I understand correctly, only using these encodings breaks, right?
I’m using several sub-repos per user since I suffered a repo corruption (probably because of a defective RAM module when rewriting some packs) which made me loose a good amount of data. Well, in fact the loss was so bad that it affected almost every snapshots present in the repo (including the older ones), so I deleted that repo and started from zero.
From then on, I decided to fragment each original repo into multiple sub-repos hoping that future data corruption would only damage that sub-repo, but not all of them for that user.
Sadly enough, I’ve suffered two more data corruptions, but as these were actually sub-repos I’ve only lost a small fraction of the data compared to the first one.
I’m saving some money to buy a new computer to properly handle the rest server and avoid these nasty data corruption (three in one and a half year, aprox.), but nevertheless I would like to keep using sub-repos to control damages in case of data corruption.
I would really appreciate if you could please keep the sub-repo feature on the rest server, because I find them really useful to limit data corruption if it ever happens, as previously described.
Yes, in general that works. So if you have a user
--private-repos will only grant access to the repo
Thanks for the report! We’ll probably add a sub-repo function later on. For now, you can just create a bunch of different user accounts so you get several repos.
Cool thanks! I just switched a couple of installations to rest server for the first time and so I can safely upgrade them.