Ubuntu VM issues after CORE to SCALE migration

TrueNAS General
1

I’ve just migrated from latest stable CORE to SCALE and I’m struggling to get my ubuntu 24.04 VM running proberly.

I found the hard way that I needed to manually update /etc/netplan/00-installer-config.yaml with the new network interface name (it’s using VirtIO).

And now that I have a proper network config and can ping my network, I realize that the VM is unable to ping the SCALE Host. Only the SCALE host IP can’t be pinged. The rest of the local net is accessible.

The SCALE IP is fully available from other devices, only not from its Virtualized host.

Any idea where to look for? I’ve never encountered such a network issue.

2

You need to create a bridge. And then assign the bridge to the VM.

3

I made a video explaining why and how

TrueNAS Scale: Setting up a Static IP and Network Bridge // Access NAS host from VM

1 Like
4

Thanks, I realize SCALE also cannot ping the VM.

I’ll create the bridge but how was it working before ??

5

Core auto makes the bridge

Technically, a bridge is slower than no bridge.

6

Why wouldn’t SCALE also auto makes the bridge then?

I received lot’s of issues when manipulating static IPs on the main interface:

Specifically, when trying to remove the static IP from the main interface before creating the bridge (I use .14 instead of your .33 example) :

‘192.168.0.14’ is being consumed by Applications, please use a different node IP in applications configuration. (I have no Applications set yet

And upon br0 creation I receive:
[EINVAL] interface_create.aliases.0: The network 192.168.0.0/24 is already in use by another interface.

7

Well, not taking into accounts the interface bugs (I’ll figure how to get past them) I still don’t get why I need a bridge with an additional IP address where everything was fine and transparent on CORE.
I had one IP address with no bridge and it was accessible from within Bhyve Vms.
So was it Bhyve able to do this vs QEMU/KVM needing a network bridge ?

In any case, using your video IPs example, I need SCALE to keep being accessible on 192.168.0.166 and VMs to access its storage and services on 192.168.0.166 to keep existing configuration working.

In your example you don’t reassign an PI (DHCP or Static back to the original ens3) Is it intentional @Stux ?

8

Yes. It’s in intentional. The Ip is assigned to the bridge interface.

You need to unset the apps pool, just like VMs need to be shutdown, and jails stopped.

Afterwards, you can reselect the apps pool and configure the network interface.

9

Unfortunately I can’t access my CORE configuration to understand how this was done.
But it seems to me that here should at least be a mention of this difference between CORE and SCALE in the migration documentation

1 Like