I require some input from you regarding best practice to backup my TrueNas apps and some other datasets.
I understand the 3 - 2 - 1 method and I am following it to an extent.
Currently, I run x2 local backups via replication from my datasets to
A HDD on the Host system
A HDD on another local server with truenas installed.
My question is, what is the best way to get my no.3 backup to the cloud. I am already backing up some media using webdav / Cloud backup push via the GUi, however, this procedure does not seem correct when wanting to backup up apps with databases that change rather frequently.
What would your best advice be?
Ideas which I had include:
Running another truenas cloud server with Hetzner and connecting my storage box. Maybe too slow?
Connecting direclly to the storage box was unfortunately scrapped l, as replication is not supported.
using cloud backup to push the replication snapshots to the cloud?
I believe in belts and suspenders for my data. I backup to Backblaze and to Storj. I am backing up a few TB of critical data and I spend less than $15 a month.
No, and no. For Backblaze, you’d set up a “Cloud Sync Task,” and for Storj, you’d set up a “TrueCloud Backup Task.” Each would then back up the specified data on the specified schedule.
Got it, thanks for the clarification. Would you then backup the individual snapshots via Cloud sync task? Its mainly my app config / data that I want to back up, in case of disaster recovery.
I know. It just makes my life easier in case of recovery. I recently messsed up with my immich database, and realised afterwards that my backups were incomplete. I never want to be in that position again.
I don’t believe the cloud sync tasks support backing up snapshots as such–you’d back up the data as it is on your pool at that time. But not having used them before, I’m only guessing on that point.
This is not recommended with live databases. I think I have found a workaround though. Most of the apps allow database dumps. I will do local dumps to a dedicated dataset and then back up that dataset containing the dumps.
Agreed, but snapshots are no safer–they don’t quiesce the database before they’re taken. Your plan of making database dumps and backing them up is much safer.