I have been postponing the migration of my TrueNAS Core server to TrueNAS SCALE primarily because of my Nextcloud instances running in iocage jails nearly a decade now.
My current setup is as follow:
HAProxy running on Pfsense (dedicated system) is handling reverse Proxy to the various Nextcloud instances (jails).
Each Nextcloud instance has its dedicated Letsencrypt service.
each jail is based on LAMP.
Each Nextcloud db and files are hosted outside the jails on their dedicated dataset.
Jails access the db and files dataset via a mount point.
I want to hear what users have to say about pros and cons and things to watch for when running Nextcloud on TrueNAS SCALE in a VM, Docker and whatnot.
I have been successful in creating on a spare system running TrueNAS SCALE a new VM running Ubuntu with LAMP and accessing the db and files dataset using VM NFS mount point, but I have read, that the NFS mount point may not always be reliable (I believe it was a warning on core to scale migration related to VM). Not sure if it is true or not.
In the TrueNAS core forum, a lot of users have experienced loss of their data when using the Apps. How does Docker handle data integrity in this regard?
Right now, I am trying to assess what would be the most reliable way to recreate my Nextcloud environment.
I am leaning toward VM (running Ubuntu or Debian, which one to choose?), with LAMP and Nextcloud installation file.
I was running a nextcloud instance on a jail too, a very simple setup.
As far i remember, i was also using mount points for db and data; instead i had caddy as reverse proxy, and the certificates were handled manually (i had a free annual wild card).
When i sidegrade to Scale-docker, i faced a lot of problem (migrating talking):
the NC version used in Core was kinda different from the docker version, on file structure
also on settings i founded a lot of discrepancy, especially on the networking parts
in core i was using a different db from the docker version
At the end, with the fear to mess up - a bit also intimorared by all the things to handle, after dig a bit into some thread in the NC forum… i decided to take the fastest way: i restarted from scratch.
In a new fresh instance in docker, after check the app was working and fix all the setup warnings provided by the app Itself
i manually recreate the users (they were just 3) as in the previous jail
i moved just the upload folders into the new file structure
i launched some command to align the files into the db (honestly i don’t remember exactly what commands, was more than a year ago)
i also have had to launch another command for each user after change settings on the phone apps
And thats all, up and running again with minimal effort.
Talking about difference from jail-docker: the docker version works reliably as the jail one, at least for me that i don’t make a massive usage of this app. I just stick to a specific version of mariadb (11.4) and not perform any automatico upgrade with watchtower to avoid break anything
