OMV move from Physical to VM?

Another (my) approach using Proxmox & OMV is:

- Proxmox is installed on a 32GB SSD.
- All the external storage (xxxTB)is installed and given to proxmox, defined via LVM.
- OMV (VM) has 2 disks: 8 GB for System and xxTB for DATA. Storage is defined in Proxmox GUI (LVM)
- Every VM uses same approach. OMV is only used as a NAS for physical machines, Video, ....

Advantages:
All storage is handled/controlled/checked by proxmox. No need to passtrough. Very easy to maintain.
If any machine needs more storage, size (of the second disk) can be incremented (via LVM)
If all storage defined to proxmox is used, increase storage to Proxmox (under same VG) and again you can resize LVM for a given VM
For VMs access to storage is quicker trough Proxmox ( For rest of physical machines is a bit slower ).
Much more flexible for creating/destroying VMs, I.E. testing new releases of OMV (actually testing ZFS 3.03)

Regards

it would be lovely, but I suck when it comes to LVM. or ZFS.
I wold risk using LVM or ZFS for OS drive but how do I manage my data safely.
I have this setup right now.

BOX: Supermicro SC846
MB: H8DME-2 with CPU: 2x AMD Opteron Hex Core 2431 @ 2.4Ghz for total of 12 cores
RAM: 49GB DDR2 PC-5300 @ 667mHz ECC
NIC : 2x onBoard Realteck +
1 Intel Pro Dual port PCI-e card for total of 4 ports
DRIVES :
OS : 2×240 SSD
DATA: 2x3T HDD --> SnapRaid Parity
4x2T HDD
2x1T HDD

currently running OMV 3.0.40 with SnapRaid+MergerFS
all data drives are formated with BTRFS

this post really made me think in moving to Proxmx+OMV setup.
even span the ProxMox VM to set it out.

but having some reservations for now.

my concerns are, in order of importance

#1. managing and expanding data storage
what the most easiest way of doing it?
OMV provides nice an some what easy UI for that but how would I do it on Proxmox setup?

#2. data pool safety and integrity.
if drive fail how not to loose data and keep running while recovering?
how to monitor drives and catch failure before it happens? I know nothing is 100% but one can try.
again OMV has a nice plugin for running SMART and email if something comes up. Proxmox???

#3. ??? not sure yet what to ask here.

-LVM is very simple to use. It's main advantge is the abilty to expand storages. I repeat is very easy to use!

Proxmox is a Virtualization GUI that makes you create/manage virtual machines. Of course have (can have) all the linux utilities that you can need. Proxmox handles LVM from it´s GUI. (lot of tutorials)
-ZFS is not so simple, but seems to have much more power (currently testing)

Proxmox supports from the box ZFS (I've not yet had the time to test).

If you like to 'play' definitively go to Proxmox:
You can create a virtual machine, install OMV, add 4 disks to create a RAIDZ2 pool , detroy 1 or 2 disks and probe that data is still there. Add two new drives and check how they are resilvered. And finally add 4 more disks join them them to the pool to obtain doble storage.

Once decided the configuration make your OMV serve your data, and when another OMV release is out, create a new VM and test it. If you like the new release, you can 'move' your disk to your new VM, and you are running the 'new' OMV release with your 'old' data.

(Note ZFS for OMV 3.0x is currently in Beta stage).

Zitat von ryecoaaron

LVM can also be dangerous if the disks underneath it don't have redundancy and the LVs/VGs are spread across multiple disks.

Correct !!
So I will definitively go to ZFS. (Hope to have some time....)

Zitat von ryecoaaron

LVM can also be dangerous if the disks underneath it don't have redundancy and the LVs/VGs are spread across multiple disks.

and herein lies the problem.
my setup does not have hardware raid, and I don't really want one.
but LVM was designed to make the raided setup be more flexible and extendable.

further more if I want to use ZFS it would be on the HOST (Proxmox)
not the VM, unless I am passing the raw disks into VM. all management therefore would be on the HOST.
if I am passing disks into VM than I can do the same thing I do now BTRFS snapraid+mergerfs

of the issues I have on my hands is that I have a mixture of disks of different sizes.
ZFS and LVM will not let me use them in a single pool. while having several pools do make some sense in specific situations, like I can use my 1TB drives to store the VM data and config as well as ISOs
the rest of the drives I would like to pool into one big storage pool.

First of all, I must say that I no expecience with Snapraid + MergeFS. So I can miss something important.

Agree with you:
For data disks, use ZFS under proxmox (and also raidXX what ever you choose)
Pass a ZFS volume to the VM (no need to raid inside the VM)

You can also use your 2 x SSDs to hold (besides Proxmox system) system disks for some of the VMs. (I have proxmox running on 32 GB SSD). So your 'special' VMS can beneficiate from the speed of them.

I do not think I will move to ZFS right now. too many mismatched disks for my liking.
I guess I will rebuild my setup with Proxmox on ZFS/raid1 on my SSDs
and pass the data disks sunce 2x1TB to OMV VM
will use the 1TB ZFS raid1 for VM data and system backup on host
the rest is OMV...

Is a BTRFS plug-in only available in OMV version 3.x? Not in version 2.x?

There is no btrfs plug-in.
It just since v3.0 is on newer kernel, it supports btrfs . To have btrfs support on v2 you need to run back port kernel.

Sent from my SGH-T889 using Tapatalk

OK. I've installed Backports kernel. But I still don't see an option for BTRFS. What am I missing here?

If you are running v2 there is no option. The system just supports btrfs, as in you can add drives preformated with btrfs, and use command line to install btrfs tools and manage the drives. As is, without backport kernel, omv does not recognize drives with btrfs and think they are empty. So, ssh to the server, run apt-get instal btrfs-tools
And you can create fs on the new drive. Once you do so it will show up in file system tab.
Or if you already have drives with btrfs, they should be available for mounting as soon as you add them.
On my server, all my drives were mounted right after reboot.

On version 3 you can format drive with btrfs from the file system tab. No cli needed.

Sent from my SGH-T889 using Tapatalk

Hello!

I use this thread because I think my problem is quite similar. My HP N54L has unfortunately died (RIP), leaving my data inaccessible. As you can see on my description, I have a Raid 5 4*3To Software, managed my OMV 2.
I already bought a new server (HP Microserver G8- Celeron G1610t & 4GB), and I'm planning to set up a small home-lab with Esxi, without forgetting its primary function: the NAS part.

So I wonder, How can I remount my raid 5 in an OMV VM using ESXI?

I don't mind if I need to buy some hardware parts.

Thanks in advance

I think the only way is to do a pass-through of all the drives into VM and reassemble the raid there.
several issues jump out here.
#1. does the hardware supports drive passthrough in ESXi?
#2. do you know how to do that in ESXi?
#3. do you have enough drive spaces for all the data drives and system drives on the new server?

if it was me, I would load the drives onto spare PC , backup the data onto external space. and try all of it on empty drives. would not play with the live data if I could help it.

also is there any reason you go with ESXi? what advantages it gives you over other solutions.
just want to know so no obligation to answer that one

Hello vl1969

#1. No, the Celeron has no VT-D. But I can change it. Otherwise, the hardware is fully supported by ESXi. HP has even made a special ESXi image for this server.
#2. I have no clue, but I'm not afraid to learn and play around with it, until I have a setup which fulfills my needs.
#3. I will directly use the 4*3To directly for my data, and recycle my old Kingston 32GB SSD for the datastore for ESXi. ESXi will be installed on a microSD.

I already have a backup of my important data. The rest can be found again.

I go with ESXi, because, compared to Proxmox, ESXi has a better compatibility with my hardware. I can also run OMV as ost, and then use virtualbox plugin, or docker. But I alread did it I want to learn hypervisors. That's why ESXi