Migrating TrueNAS 12 install from Intel Xeon to AMD EPYC

[gsrcrxsi]

So I currently have a system that's running TrueNAS 12.0-U1 on Intel Xeon hardware

Specs:
2x Intel Xeon E5-2680v2
128GB DDR3-1600 ECC RDIMM
dual SATA SSD boot drive mirror
LSI HBA running 8x 12TB SATA drives (mirrored vdevs, "RAID10")
a couple PCIe/Nvme drives for jails and VMs
runnig things like Plex/*arrs/a few Ubuntu VM installs, pihole, ubiquity controller, etc. basic stuff

I would like to upgrade the hardware on this system, and move from the dual Xeon setup to a single socket AMD EPYC setup. maybe like one of the 16 or 24 core processors. mainly want to upgrade for lower power consumption, better compute performance, and overall modernization of the server.

Can I expect problems doing this? or can I basically just upgrade the hardware and then boot it back up without changing anything? has anyone done this before? What about the VMs and jails, are they OK with this kind of change?

or will I have to re-install everything? I REALLY don't want to do that.

 

[JohnDigital]

In my experience Linux and Unix systems care not. You can essentially just shutdown, do the upgrades, and boot back up. The only time there might be a hiccup is when the ethernet cards or chipsets change. So on first boot you might want to look at the terminal output and reconfigure network from the terminal (if needed). There might also be hiccups because of this in jails too if the cards change. Some manual config edits can get you going again pretty easily.

 

[winnielinnie]

I second what @John Digital wrote.

I had upgraded practically every component (except my drives) with no issues, and I didn't even do a re-install! I just booted up from the existing boot drives. I'm using a few plugins/jails, one of which is Plex, and I haven't had any obvious problems streaming or managing my library. I don't use any VMs, so I cannot speak on how well such a migration will go in that regard.

One practical (and only theoretical) hiccup might be if you're switching video card vendors (say from Nvidia to Intel), and you had previously configured one of your VMs to use specific hardware encoding/decoding (NvEnc, VA-API, Vdpau, CUDA, etc). Which I'm not even sure is possible with bhyve.

 

[Evertb1]

Just upload your config file to a save location before you do anything. That way if you are somehow forced to a fresh install you can dowload your config file and you are good to go in no time.