I have start to write an history of all nextcloud crash in my cluster infra with fully infos (in french but google trad) : https://www.0rion.netlib.re/forum4/viewtopic.php?f=107&t=761&p=1773#p1773
Actually i try to repare but i don’t know if i can continue with nextcloud, if it’s for it to crash again and again every month. The lonely solution i have drop for hack this regular crash is to make a LDAP server and disable server-side encryption. But this solution introduce another point of failure (ldap server) and need learn about another software (LDAP)-
I think most people do not host their cloud, they give the hosting job to OVH or other big brother corporation ^^
They just not need protection again data lost, it’s the job of the hoster.
In own-hosting world, the admin need make all the job and learn everything (for non natural english speaker, it’s not easy to drop fresh informations) and solve more instability than datacenter.
===> storage cluster without point of failure (glusterfs-server replica 3 arbiter 1 (bugged) or Ceph (2 datanodes + 1 metadata node))