Has anyone successfully deployed cnMaestro version 6 on Hyper-V. I tried extracting the OVA file today and using StarWind to convert the disks to Hyper-V format, but I was unable to boot the OS. Anyone have any details? Thanks…
My new plan is to install on VMware Workstation Pro first and then use the StarWind V2V converter to convert to Hyper-V. If anyone has anything to add, I would greatly appreciate it. Thanks…
I’m in a similar boat, except we run Nutanix AHV. I think it has an OVA converter but I have to figure that out… this would be much easier if they would just provide a raw disk image and instructions for what NIC(s) / etc are required.
I’ll definitely let you know if I make any progress.
It seems to work well in VMware Fusion and Oracle VirtualBox, but neither of those are enterprise really.
cnMaestro On-Prem definitely works with Hyper-V, but is a bit of a pain to get going as Cambium’s default mount points are not very robust, causing the drive order to slip sometimes between reboots. And on 6.0.0, the Hyper-V console will dump to a black screen with Cambium’s default grub config.
Here are some old partial notes, should at least give you a starting point.
Convert disk image format from VMware .vmdk to Hyper-V .vhdx
- Download and install QEMU for Windows ( need to use qemu-img.exe to handle the disk format conversions)
- Download and extract the two .vmdk files from the latest primary cnMaestro ova build;
- Use qemu-img.exe to convert the two disk images from .vmdk to .vhdx:
qemu-img.exe convert -f vmdk -O vhdx -o subformat=dynamic -p "C:\REPLACE_ME_SOURCE_FOLDER\cnmaestro-on-premises_6.0.0-r6_amd64-disk1.vmdk" C:\REPLACE_ME_DESTINATION_FOLDER\cnmaestro_disk1.vhdx
qemu-img.exe convert -f vmdk -O vhdx -o subformat=dynamic -p "C:\REPLACE_ME_SOURCE_FOLDER\cnmaestro-on-premises_6.0.0-r6_amd64-disk2.vmdk" C:\REPLACE_ME_DESTINATION_FOLDER\cnmaestro_disk2.vhdx
Create a legacy Gen1 Hyper-V VM, and attach the two converted drives;
- Make sure both drives are attached to the same IDE controller AND are in the correct order.
- Attach a live boot disk for some kind of flavor of Debian/Ubuntu/Fedora; make sure boot order is set to boot from DVD (going to fix the busted mount points first)
- Power on the VM, live boot from the DVD
- via Terminal, elevate as sudo;
enumerate UUIDs for the two drives; typically the second drive is the one you need the UUID for, but this order can flip
lsblk -ftemporarily mount os partition 1:
mount /dev/mapper/system-os1 /mnt/edit /etc/fstab with your favorite text editor, and fix that /mnt/data mapping to instead reference the UUID!
nano /mnt/etc/fstabfinal product should look something like
UUID=SHOULD_ALREADY_BE_SET_CORRECTLY /boot ext4 defaults 0 2 /dev/mapper/system-os2 / ext4 noatime,errors=panic 0 1 /dev/mapper/system-os1 /mnt/os1 ext4 noatime 0 2 /dev/mapper/system-tmp /tmp ext4 noatime,nodev,nosuid 0 2 /dev/mapper/system-log /var/log ext4 noatime 0 2 /dev/mapper/system-swap none swap sw 0 0 UUID=REPLACE_WITH_UUID_FOR_DRIVE2 /mnt/data ext4 noatime,errors=panic,nofail 0 2after you save your edits to fstab, we need to fix grub.
umount /dev/mapper/system-os1
mount UUID="REPLACE_WITH_FSTAB_UUID_FOR_/BOOT" /mnt/
nano /mnt/grub/grub.cfgin grub, find busted line
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"and set instead to
GRUB_CMDLINE_LINUX_DEFAULT=""
save, then gracefully shutdown the live boot environment. remove the live boot iso from the VM configuration.
Boot the VM; once the VM finishes booting and if all goes well should you should be able to login with default creds and then see the cnMaestro setup TUI
Yeah, you have to extract their crap VMWare OVA using qemu-img to get the disks.
Has to be legacy, due to requirement of BIOS and not UEFI
This worked perfectly, thank you.
Just a note if anyone is using the latest v6 file: cnMaestro On-Premises 6.0.0-r6 / 2026-07-15
When editing the grub.cfg the “quiet splash” line you need to remove will be under menuentry “cnMaestro Active Service (6.0.0-r6)”
Only remove “quiet splash”, and nothing else.
Thanks for all of the replies everyone! If I follow this process to get cnMaestro running on my Hyper-V cluster and then Cambium releases an update to remove cloud dependency, will this be easily updateable or will I have to start from scratch again? Not being familiar with cnMaestro on-prem, I’m unsure how updates work.
FYI – the backup and restore functionality of cnMaestro does NOT work.
So if you ever need to uplift and migrate… you have to snapshot / clone and move. The upload / restore doesnt bring over the entire state. I learned this in testing.
I don’t understand the question about cloud dependency.
Currently, cnMaestro depends on Cloud Anchor.
Here is the history of my experience with Proxmox.
After many years of using cnMaestro On-Premises on Proxmox, we decided to switch to the cloud. It turns out that both the On-Premises and cloud versions store the same amount of history—7 days—so we didn’t see any difference; plus, NSE3000 units cannot be managed from cnMaestro On-Premises 
Good luck with your experience, and if you manage to create a step-by-step guide, don’t hesitate to share it in the cnMaestro category.
Regards.
On cnMaestro 6.0.0-r6, I did NOT need to mess with /etc/fstab - it was already using the UUID for both /boot and /mnt/data. I’m assuming this was fixed in the OVA at some point…
Nutanix AHV is actually just fancy KVM/QEMU under the hood. They have an extra management VM you can deploy called “Prism Central,” and it adds the capability to “natively” deal with an OVA image and create a VM from it. Otherwise, you have to manually crack open the OVA and setup the VM “by hand.”
I deployed from the OVA via Prism Central (I used “legacy boot” rather than UEFI based on your advice). The VM seems to boot correctly, because I’m getting pings back from it… but I can’t see the local TUI console. I’m guessing that this is due to the grub options, like with HyperV? Do I even need the TUI if I can ping the VM (can the setup be done purely via the web)?
My grub.cfg from the /boot partition does not have that variable at all, it’s an auto-generated file from grub-mkconfig with stern warnings not to edit it. 
Also, it seems like it might support UEFI boot, based on some of the stuff I see in there.
Dumping the “quiet splash” from the end of the kernel commandlines seems to have done the trick though, and while I was in there I edited /etc/default/grub to set GRUB_CMDLINE_LINUX_DEFAULT="" so this way if it ever gets re-configured, the console won’t reset back to being blank.
It depends on whether they release an “upgrade-in-place file” for the existing OnPrem deployment, or they just spit out a completely new OVA that has the licensing restrictions removed. If it’s a new OVA, you’ll probably need to backup settings from your existing and then restore into the new one after setting it up as a new / second VM. Check out the “Server Management” section of the cnMaestro User Guide, specifically page 863-869.
Based on past experience with other wifi controllers like this, if you backup the old one, shut it down, and then start the new one with the same IP address (or a different IP address, but you used a hostname in the controller URL for all of your devices rather than an IP, and you make the hostname point at the new controller’s IP) and restore the backup to it, you probably shouldn’t need to “re-migrate” all of your devices a second time (especially not if you’re using a hostname for the controller URL rather than an IP).
[edit] based on Anthony’s findings that “backup doesn’t work” though, I’m a bit concerned about what the migration path will be if they do release a license-free OVA in a few weeks.
What specifically is missing when you backup & restore from the UI? That’s more than a little concerning since it seems like there’s a good possibility we all might need to do another OVA deployment when/if they release a “no-cloud/license-free” version on or after 10/1.
Fails to import… I exploded the tar.gz it dumps and it appears to only be the database of devices and historical.
I had created two(2)'s and appropriately snapshotted. I spun one for essentials and then licensed.
I spun up another, fresh setup of cnMaestro and attempted to import a new backup… it fails and errors.
Feel free to test yourself and report back if same issue.
Assuming the VM boots properly, all the underlying services start correctly, and the VM gets its IP addressing via DHCP then no, the TUI shouldn’t be explicitly needed.
If so, I would recommend at the very least getting into the TUI via SSH to update the default OS password; would need to temporarily enable SSH via the web UI (Administration → Server → Settings → ‘Enable SSH access to server’. Then once your changes are complete, disable SSH if no longer needed.
Hope this helps others; I already “hacked the gibson” as my generation used to say (edited my earlier post to reflect findings). 
The “cnmaestro” user has passwordless sudo access so changing it from default is definitely a good idea! System is just a tweaked Ubuntu 24.04.4 LTS, doesn’t seem to pull any updates by default which is a little concerning though not unexpected. The OS implies that it should handle UEFI correctly.
Fantastic! This was definitely not the case with older versions such as 5.x.
It does, but with potential caveats. Have not tested on 6.x, but initially tried running as a UEFI-based Hyper-V Gen2 VM back in 5.x, and from what I vaguely recall quickly ran into issues with the kernel not supporting how some of the iftables or ebtables NAT rules were being applied either for HA or RADIUS proxy. So never kept that in going in prod.
If people need Hyper-V or Azure booting vhds give me a shout - I’m not going to host them publicly but happy to share the fresh working images as we’re all in the same boat.
Synology DS916+ attempt failed for me last night.
The vm started up but couldnt get to a ubunto login prompt. the web server apparently started but i got a message that (and i didn’t write it down so sorry for the vagueness) from the cnmaestro team that it wasn’t available. It was consuming cpu and copilot had me do various tests but ultimately decided that something on the back end had failed.
Now doing my first one with vmware workstation pro which came up right away.
setting up a test environment in online X with the units i had on the shelf with some waiting to be installed at client locations on the theory that we need to get all h/w into the cloud and registered and then migrated into an on prem before they pull the plug.
question remains for me whether i will be able to move devices from one on prem to another.