unfortunatly, it is a normal synch app behaviors…
A server instance, with the authority to request deletes from clients, should take measures to conclude in which manner files disappeared, e.g. requested from clients or delete on the server itself, or disappeared without inotify:unlink - the redundancy inherent is really thrown out of the window with the server requesting clients to delete their files as well without any second guessing.
Remember, nextcloud is not a backup solution unless specific cron jobs and manual scripting.
I partially relied on the server to conclude what I just mentioned, and provide some backup functionality, but I was mistaken.
Second thing, a RAIS failling and recovering only a part of the files, this is weird. I am interrested in the raid info, could you be more specific.
Two physical disks with mdadm in RAID1, one disk failed entirely, the 2nd disk having errors as well (has gone unnoticed until too late), I was also surprised on the ordered manner I lost data.