Hi,
TL/DR:
I’m unsure if my off site backup strategy with spare disks is a good idea or not and wanted to check for any feedback.
Long Version:
I have the following TrueNAS Server
- HL15 HomeLab Server
- Data: 6 Disks in RAIDZ2
- Log: 2 NVMe in Mirror
Workload:
- Media Library (Plex)
- Time Machine / End Device Backup Location
- Large K8S Volumes (CSI Driver Integration – in the end NFS Shares)
- Minio
- Backup Location for Longhorn (K8S persistent volume manager)
- Backup Location for etcd (k8s control plane backup)
- Backup Location for PostgreSQL
→ so in short the storage is used for backups and for data I could loose if all goes wrong
Strategy:
I plan to have two disk which are much larger than my data disks. Have both as pools (with one disk stripe each) with a replication tasks of the main pool. The replication tasks would be scheduled differently: one for even and the other for odd months. I would keep the corresponding disk in an offsite storage and switch them out at the end/beginning of each month.
I’m aware that I would reduce fault tolerance but as this is for the most part a backup of a backup I thought it could be ok.
Questions:
- Is this strategy ok or could I improve it (without heavy investment)?
- Are there other good approaches for an offsite backup without the investment in a co-location?
- What is the best way to remove the disk? (I saw that it doesn’t seem to work to turn the last/only disk in a pool offline)