Training and Certification

Hey guys,

I'm having issues with vJunosEvolved on EVE-NG. Once I configure a device (hostname, root authentication, new user, etc.), the VM enters some kind of reboot loop after a restart. It only stabilizes again if I wipe the configuration and start from scratch.

Has anyone faced a similar issue? If so, could you please share any suggestions or fixes?

So far, I've tested vJunosEvolved versions 23.2R1 and 23.4R2 and seen the same behavior on both. EVE-NG is running on a bare-metal server with enough compute resources.


Best,
------------------------------
Ricardo Simba
------------------------------

For future reference, I think I figured it out, and it seems the issue has nothing to do with vJunosEvolved itself, but rather with EVE-NG.

When creating a new vJunosEvolved node, EVE-NG automatically generates the QEMU settings below, which apparently emulate an older Intel CPU architecture (IvyBridge), regardless of the actual physical processor installed in the server:

' -machine type=pc,accel=kvm -serial mon:stdio -nographic -smbios type=1,product=VM-VMX,family=lab -cpu IvyBridge,ibpb=on,md-clear=on,spec-ctrl=on,ssbd=on,vmx=on '

And replaced with the below:
' -machine type=pc,accel=kvm -serial mon:stdio -nographic -smbios type=1,product=VM-VMX,family=lab -cpu host,kvm=on '

The above setting forces the vJunos node to detect the actual physical host CPU and passes the Intel VMX hardware acceleration features natively to the internal forwarding processes.

In addition, if you want to stop a vJunos node, you should avoid clicking "Stop" directly from EVE-NG, as this can corrupt the vJunos internal virtual disk and leave the VM stuck in a permanent reboot loop.

Instead, first issue the request system power-off operational command and wait until the console displays the 128.0.0.1: No route to host message. Only then should you stop the node from EVE-NG.

Best,
------------------------------
Ricardo Simba
------------------------------