I have Scale Electric Eel 24.10.2.1 for my main NAS and hypervisor, installed on a custom server.
It runs four VMs, one of which is ultra-critical: namely OPNsense which is the main router and firewall for my house and is a complicated setup, with two internet feeds, failover, and dynamic DNS. Also, my spouse works a corporate job from home and is in meetings on Zoom throughout the day, so there is no way that I can afford any internet downtime.
In addition, another VM on Electric Eel is a Nextcloud server that my spouse and I both rely on for work.
I have seen other threads suggesting that even manual migration from an Electric Eel VM to a Fangtooth instance doesn’t work, with people saying that they had to recreate VMs from scratch after upgrading to Fangtooth, which would be disastrous in my case.
So I have two questions: (1) Has anyone successfully manually migrated from an Electric Eel VM to a Fangtooth Instance, and (2) does anyone have any information regarding if and when migration will become more automated and when I should do my own upgrade?
This really is a case where your router should be on its own hardware. Such a complicated setup with the extremely-limited virtualization capabilities TrueNAS provides is risky at best, and I’d recommend against it even with a competent hypervisor.
I have. It was a Debian Linux VM, Fangtooth imported the zvol without a problem, and otherwise matching the old VM’s configuration wasn’t an issue. However, the network interface unexpectedly (and for reasons unknown) changed names, so I had to VNC into the VM and reconfigure the networking manually. Once I did that, it’s worked fine.
There are advantages and disadvantages of running a router on a hypervisor vs. bare metal, and I’ve tried it both ways. The advantages of running it on a hypervisor are substantial: less hardware and media to maintain, less power and heat, and most importantly imo the ability to take snapshots before updating the router software so as to be able to roll back if the update causes problems.
Regarding Truenas SCALE as a hypervisor, it has fewer features than some other hypervisors, but in my experience, it’s rock solid stable. I’ve had no problems with it and have been running OPNsense on it as a VM, as well as running other VMs, for two years.
This is good information. Thanks. I’ve seen other threads in which people said the upgrade to Fangtooth rendered the VM completely unusable even with manual configuration. So this is good to know. I can deal with manual configuration of interfaces from within the VM over VNC. I’m still going to wait a while, however.
A potential problem with waiting for Goldeneye is that I don’t believe it’s possible to leapfrog major versions when upgrading, e.g. from Fangtooth to Goldeneye. I could be wrong in my understanding of this, but I’m pretty sure it’s necessary to upgrade major versions sequentially.
The other possibility is to reinstall and restore from conifg, but that presents other potential risks.