Hello community,
I am pretty new and noob to Proxmox and ZFS and trying to setup a new homelab. Here is my configuration for it
1xNVME 1TB
4xSATA HDD 4TB each
CPU: 12th Gen Intel i5-12400 (12) @ 5.600GHz
RAM: 64 GB (planning to upgrade to 128, but not in the nearest future)
My motherboard does not support more SATA connectors (only 4 of them) and 1 place for NVME. So basically everything is up and used now.
My main goals for the build are:
- Host several services like gitea, nginx reverse proxy etc
- be a home kinda 'nas' for having qtorrent, photoprism, jellyfin etc
- use it for software development purposes (have 2+ vms + docker within)
I read tons of articles, but still don't know what would be the best choice to split the storage.
Option 1:
- use NVME as PVE root + remaining space for my VMs. Install ZFS on that single NVME.
- create a 'raid-5' kinda pool using ZFS on 4 remaining 4 TB HDDS and use it as sharable storage for future VMs
PROS: using ZFS allow use it's cool features like compression, backup etc. Have enough memory to deal with the pool.
CONS: only one drive (single one for root and VMs) thus now allowing to backup PVE installation.
Option 2:
Same as Option 1, but use LVM-thin for NVME
PROS: Using less RAM for ZFS, not too intencive writes onto NVME
CONS: not using all ZFS cool features!
Option 3:
Use one HDD as PVE root and use NVME purely for vms. But for this option I don't know how to deal best with remaing space.
couple of notes: I don't really care if my NVME dies in 3+ years because of writes to it, but I do care that I can reinstall PVE on a new one and restore my VMs (from HDDs?). Also want to use most of possible space for my HDDs
What do you think about it? or maybe there are other possible better setups.
I am pretty new and noob to Proxmox and ZFS and trying to setup a new homelab. Here is my configuration for it
1xNVME 1TB
4xSATA HDD 4TB each
CPU: 12th Gen Intel i5-12400 (12) @ 5.600GHz
RAM: 64 GB (planning to upgrade to 128, but not in the nearest future)
My motherboard does not support more SATA connectors (only 4 of them) and 1 place for NVME. So basically everything is up and used now.
My main goals for the build are:
- Host several services like gitea, nginx reverse proxy etc
- be a home kinda 'nas' for having qtorrent, photoprism, jellyfin etc
- use it for software development purposes (have 2+ vms + docker within)
I read tons of articles, but still don't know what would be the best choice to split the storage.
Option 1:
- use NVME as PVE root + remaining space for my VMs. Install ZFS on that single NVME.
- create a 'raid-5' kinda pool using ZFS on 4 remaining 4 TB HDDS and use it as sharable storage for future VMs
PROS: using ZFS allow use it's cool features like compression, backup etc. Have enough memory to deal with the pool.
CONS: only one drive (single one for root and VMs) thus now allowing to backup PVE installation.
Option 2:
Same as Option 1, but use LVM-thin for NVME
PROS: Using less RAM for ZFS, not too intencive writes onto NVME
CONS: not using all ZFS cool features!
Option 3:
Use one HDD as PVE root and use NVME purely for vms. But for this option I don't know how to deal best with remaing space.
couple of notes: I don't really care if my NVME dies in 3+ years because of writes to it, but I do care that I can reinstall PVE on a new one and restore my VMs (from HDDs?). Also want to use most of possible space for my HDDs
What do you think about it? or maybe there are other possible better setups.