Proxmox Storage Disks how to.

[Nightraider]

Hi there,

Just registered here, started this week with ProxMox, coming from an ReadyNAS it was time for something bigger

I do find it a pretty steep learning curve, but ill get there.

One thing i can't get my head around is the Storage.

I have 1 256gb NVME SSD, (Proxmox Boot as partition one, partition two for VM, Container and iso's)
and 2x 5Tb 2.5"HDD.
The NVME works as intended, mainly as bootpartition for the host and guest os'es
However i dont know how to integrate my 2x 5tb disks properly.

My ReadyNAS had x-raid , so one big disk and all was on that in Directorys (Movies, TV Shows, Backup, Download)

Best thing would be to have one big 10TB Disk, but as far as i can see that only works with ZFS or Hardware Raid. i have no ecc ram, so i dont want ZFS. dont have an Raid card either.

Second best is just 2 5TB disks, one partition each, and all my vms could just store on those. i have no clue how to set this up (if its possible in the first place)
Only thing i can come up so far is every vm gets a virtual partition of the disk, but thats not very disk economic as only the first downside.

Can anyone point me to a dummy guide to set this up, or something like that? the guides i can find on the internet make my head spin with all the different options Proxmox offers.

Thank you very much,
Bob

 

[Alwin]

My ReadyNAS had x-raid , so one big disk and all was on that in Directorys (Movies, TV Shows, Backup, Download)

That means on data redundancy. If one disk goes 'by by' then you have lost the data.

Best thing would be to have one big 10TB Disk, but as far as i can see that only works with ZFS or Hardware Raid. i have no ecc ram, so i dont want ZFS. dont have an Raid card either.

Use ZFS, ECC is not a must to run ZFS. But it's recommended. Then you can create a mirrored setup for data redundancy.

Can anyone point me to a dummy guide to set this up, or something like that? the guides i can find on the internet make my head spin with all the different options Proxmox offers.

Well the steep curve is a lot of linux to learn.

 

[Nightraider]

That means on data redundancy. If one disk goes 'by by' then you have lost the data.

My X-Raid setup was 4 Disks, one had parity, so it was (sort of) a Mirroring setup.

it not a must to have data-mirroring, i had only one drive with parity but if i do that with raid1 i already need 2, and since 95% is only Media server storage, it sucks if one of them fell over, but nothing to serious. Backups of my personal photo's and stuff are also on there but also on an 2nd off site external hdd.

Use ZFS, ECC is not a must to run ZFS. But it's recommended. Then you can create a mirrored setup for data redundancy.

I read somewhere that ZFS without ECC had the risk of more corruption of the data. isnt that the case? or no more then other file systems?
ZFS also doable as JBOD? so no mirroring but also not ALL data gone if one of them fails.

Well the steep curve is a lot of linux to learn.

Yeah, i get that, learning everyday

But do i need to make partitions and such in terminal? i tought all could be done from the WebUI.

 

[Alwin]

My X-Raid setup was 4 Disks, one had parity, so it was (sort of) a Mirroring setup.

In ZFS terms, a raidz1.

I read somewhere that ZFS without ECC had the risk of more corruption of the data. isnt that the case? or no more then other file systems?

ECC does error correction on bits in memory, ZFS most likely will not be able to catch those and write corrupt data from memory to disk. But that is in principle with any other filesystem. The good part is ZFS will complain about it, since it has data integrity features.

In general, I definitely recommend the use of ECC memory on long running computers.

ZFS also doable as JBOD? so no mirroring but also not ALL data gone if one of them fails.

Yes. ZFS groups disks into pools and vdev.

But do i need to make partitions and such in terminal? i tought all could be done from the WebUI.

No partitions needed. Though you will need to get familiar with the ZFS tools and terminology. Since the second disk needs to be added to the pool by hand.

 

[Nightraider]

In ZFS terms, a raidz1.


ECC does error correction on bits in memory, ZFS most likely will not be able to catch those and write corrupt data from memory to disk. But that is in principle with any other filesystem. The good part is ZFS will complain about it, since it has data integrity features.

In general, I definitely recommend the use of ECC memory on long running computers.


Yes. ZFS groups disks into pools and vdev.


No partitions needed. Though you will need to get familiar with the ZFS tools and terminology. Since the second disk needs to be added to the pool by hand.

Thanks for your help so far, i will dive in to ZFS terminology later today.!

 

[Nightraider]

Hi There,

I Added a 3rd 5TB drive, to be able to make a zfs raidz-1 pool. worked good so far.

After some searching the only option would be to share my zfs pool with an NFS Share to the VM'S.

Thats where i am stuck now, i have root write acces to the nfs share between host and one vm. not User write acces(yet)

But is this the only way to make this happen? it looks t me it is very very complicated to make it work this way, and probably there is an more user friendly way that i have not found..

I set this all up in Terminal. have found no other way.

Isn't there something in the WebGUI of Proxmox of some sort? if i wanna do this for multiple VM's thats a long road to go..

 

[Alwin]

After some searching the only option would be to share my zfs pool with an NFS Share to the VM'S.

The zfs pool is meant for VM/CT disk storage. If you need a shared storage inside the VM/CT, then best use a NFS/SMB server.