Hello Everybody first of all bare with me... I know that both of these questions have been answered many many times but still there are some background pieces that are surely missing to me.
It all started with me wanting to build my own Nas around a j5005 platform but then I ended up purchasing a Dell T320 on ebay. At that point I was really torned between running FreeNas (that is why I bidded for the T320) or OpenMediavault bare metal.
I mean if I had gone the j5005 route i would just go throug the bare metal installation of Openmediavault but using a full blown server just as a Nas, well it seems a pity to me....
So i started with proxmox but i've found it hard to manage storage: it was/is difficult to understand the logic how storage is expected to work through the gui so at first i gave up and gone back to OMV bare metal... after all it can do virtualization with KVM through cockpit and it seem to work well with docker.
Still i thought that the right way had to be a proper virtualization platform so i tried HYperV but i don't like to have to go through two different kind of software to manage the virtual machines.
long story short i got back to proxmox and started to understand storage a little better (it is still confusing though) and for my use case i could manage it quite well.
I am going to use OVM mainly for storage purposes as creating and managing shares is easy and comfortable. but i'm not sure wich would be the bes strategy for passing the drives.
at the beginning i have passed the drives through with the qm set VMID -scsiX /dev/disk/by-id/.... command. and while it worked well OMV can't see SMART data and i'm concerned about power control over the drives.
so the next step i took was to pass de pci card (DELL H310) to OMV as for this guide: https://pve.proxmox.com/wiki/Pci_passthrough. I also created a nfs share for proxmox to use as storage.
on the OMV side i have 2 500gb, 2 1tb and 1 3tb drives pooled with Unionfs plugin and Snapraid providing redundancy on another 3tb drive for parity.
Proxmopx has a dedicated 120gb ssd plus another 120ssd added as thin lvm for VMs.
the main reason i set up the drives this way is because a nas that died on me (hardware side) and i could'nt retrieve the data because i needed the same type of hardware.
This way i should be protected from drive failure (snapraid) and hardware or os OS drive failure by just mounting the data drives on another pc and recreating an OMV virtual Machine.
At the time seemed a good idea but now i have second thoughts... for example i would like to setup nextcloud but i don't have any drive to give to nextcloud. should I go back to single drives passhtrough? should i use a different method? what are your suggestions?
It all started with me wanting to build my own Nas around a j5005 platform but then I ended up purchasing a Dell T320 on ebay. At that point I was really torned between running FreeNas (that is why I bidded for the T320) or OpenMediavault bare metal.
I mean if I had gone the j5005 route i would just go throug the bare metal installation of Openmediavault but using a full blown server just as a Nas, well it seems a pity to me....
So i started with proxmox but i've found it hard to manage storage: it was/is difficult to understand the logic how storage is expected to work through the gui so at first i gave up and gone back to OMV bare metal... after all it can do virtualization with KVM through cockpit and it seem to work well with docker.
Still i thought that the right way had to be a proper virtualization platform so i tried HYperV but i don't like to have to go through two different kind of software to manage the virtual machines.
long story short i got back to proxmox and started to understand storage a little better (it is still confusing though) and for my use case i could manage it quite well.
I am going to use OVM mainly for storage purposes as creating and managing shares is easy and comfortable. but i'm not sure wich would be the bes strategy for passing the drives.
at the beginning i have passed the drives through with the qm set VMID -scsiX /dev/disk/by-id/.... command. and while it worked well OMV can't see SMART data and i'm concerned about power control over the drives.
so the next step i took was to pass de pci card (DELL H310) to OMV as for this guide: https://pve.proxmox.com/wiki/Pci_passthrough. I also created a nfs share for proxmox to use as storage.
on the OMV side i have 2 500gb, 2 1tb and 1 3tb drives pooled with Unionfs plugin and Snapraid providing redundancy on another 3tb drive for parity.
Proxmopx has a dedicated 120gb ssd plus another 120ssd added as thin lvm for VMs.
the main reason i set up the drives this way is because a nas that died on me (hardware side) and i could'nt retrieve the data because i needed the same type of hardware.
This way i should be protected from drive failure (snapraid) and hardware or os OS drive failure by just mounting the data drives on another pc and recreating an OMV virtual Machine.
At the time seemed a good idea but now i have second thoughts... for example i would like to setup nextcloud but i don't have any drive to give to nextcloud. should I go back to single drives passhtrough? should i use a different method? what are your suggestions?
