Hello
i am a bit confused about this topic. Hopefully i understood how zfs snapshots are working (quite interesting presentation : https://www.youtube.com/watch?v=NXg86uBDSqI )
Retention policy same as source → got it, easy, just a backup of those on the source
but
None or Custom is hard to understand how they might work.
The difficult task of zfs is to determine which blocks can be deleted/ set be free. When a snapshot is about to get deleted, the system looks if there are blocks that arent referenced in any following snapshots. if not, those blocks get deleted.
But how the heck does the local system knows what blocks are referenced on the snapshots on a remote system with longer retentiontimes ?
Essentially, the remote system must have retention greater than or equal the source, otherwise the source will end up complaining about a missing base snapshot.
hm … now i am more confused than before
Server A = Source
Server B = Backup
So Server Bs retentiontime needs to be as long or even longer than of Server A ?
And this is working ? i can delete snapshots from A and let the Server A clean up after something isnt referenced anymore and i still would be able to recall/restore it by using the snapshots of Server B ?
Sorry, cant believe that, but maybe i just should try XD
i am used to it from university nvm
doesnt matter for the information given by this video … but the inability to understand the remote snapshot stuff remains ^^ but most probably i just got a flaw in my logic…
i am digging out my old thread, as i am still not sure
how does Server A (Source) knows what snapshots are left on the Backup site and therefore what files or so are still needed …
but most probably, i will try it …
nvm