Any data on one of the 8TB disks that you want to back up to the other 8TB disk (like personal photos) can be done using replication.
There is no such thing in ZFS as RAID1. There is RAIDZ1 which is kinda equivalent to RAID5 and mirroring which is kinda equivalent to RAID1 - but neither are identical to the RAID equivalents.
Edit: Fix typo RAID0 should have been RAID1.
Raid1 is mirroring, raid0 is striping
Raid10 is a stripe of mirrors.
Which is conceptually similar to a zfs pool with multiple mirror vdevs.
Ok, everything has been reworked and the H/W is up. Plex is the only app I expect to run on this.
Pool1 is at 77%, Pool2 at 14% and Pool3 (Nvme) at 11% (music only).
I copied most of the data into another 8TB drive which will be my backup as I expect minimal updates (addition) to the existing data, maybe a few more digitized home videos, around 250GB max.
I have saved the export of the existing server config. If TrueNAS goes down and has to be stood up on NEW H/W, is there anything else I should backup so I can plug-in these drives and import these 3 Pools.
Thx
Implement @joeschmuck 's Multi-Report script which will email you your system config on a weekly basis so that you have a recent backup.
What do you mean by “non-raid”? You had a stripe of two vdevs (not one): This is the ZFS equivalent of RAID 0, and if you lose a disk in RAID 0 you also loose everything.
RAID != parity array
Actually you had the worst possible configuration.
2x discs striped PLUS 1x NVMe for metadata.
Lose any one of those THREE disks and you lose everything - and the risk of losing your data is therefore 3x the risk of losing one of the disks.
Whilst I am sure your friend thought he was doing good, in reality he was simply proving that too little knowledge is a dangerous thing.
The cache vdev is not critical. But unless there’s at least 96-128 GB RAM in there (which is unlikely), it could actually end up hindering performance.
Sorry - I was wrong. Somehow I read the cache vDev as a metadata vDev.
As @etorix said, a cache vDev is not critical. But a cache vDev uses a lot of ARC to store the cache metadata, so unless you have a lot of ARC it will cache less and end up being slower. BUT… if you have a lot of memory but not a massive amount of storage (many tens of disks, hundreds of TB of data), most of what you access from disk will be held in ARC anyway and L2ARC will not be read. So L2ARC is really only for very large scale NAS implementations.
