Ok this is part 2 of my original post, I’m ready to start this build of TrueNAS Scale, here are my components:
-For space concerns, I cannot rack this build so it has to be a tower, setteled for a Rosewill Helium NAS Black ATX Mid Tower
-Running ESXI 8.0 U3 OS
-Asrock B450M Pro4 motherboard with a 2700x CPU
-80gb of DDR4 2400MHz RAM ( 2x32GB and 2x8GB, different vendors)
-8+ 2.5 SSD’s 3.8TB each (I have a few spares to have as hot spares)
–Read Intensive
–TLC flash memory
-LSI 9305-16i HBA card
-SFF-8643 to SATA cables
-M.2 256GB SSD Drive for OS
-Dual 10G Intel x520-DA2 network adapter, but my network stack is limited to 1G at the moment
Setup:
-The motherboard has dual m.2 slots, will be using the non-nvme slot as the boot drive and then with the HBA Card pass all drives to a TrueNAS Scale VM
–Unfortunately, the motherboard states:
—M2_2 and SATA3_3 share lanes. If either one of them is in use, the other one will be disabled. (which is not a problem but I do loose one on-board sata
-Also, since the CPU does not have on-board video, this will be a headless system as the HBA card will utilize the PCI Express 3.0 x16 Slot and the network card will utilize the PCI Express 2.0 x16 Slot
Questions:
The idea is to spin up a TrueNAS Scale VM and pass all the drives to the VM, I’m contemplating what setup I should do. Been watching videos of how ZFS works and it looks like you create a pool with vdevs. What is important to me is a balance between space and resiliency, and it seems doing a Raidz1 is a good balance (but comment if otherwise)
So some questions would be:
- Should I create 1 pool in raidz1 with 8 drives and 1 vdev, with my capacity I should be around 24.313TB
- Create 1 pool in raidz1 with 8 drives and 2 vdevs, the calculator drops this to 20.9TB
- Create 2 pools in raidz1 with 4 drives, wity my capacity I should be around 10.42TB
- Not sure if raidz2 or 3 advisable with only 8 drives
- In any of this configurations, I’m wondering what the resilvering time would be like in case of a failed drive? Since these are SSD drives, could I just mimic a failed drive by disconnecting it and connecting one of my spares ones
- Realistically, I also read that if you loose one complete vdev in a pool you loose the whole pool? Not sure the best way to avoid this
If you read the whole thing thanks for taking the time.