Juniper vMX: startup config breaks

[StephenKing]

Hi,

I'm running eve-ng PRO 3.0.1-16 on Ubuntu 18.04

I've set up vMX based on the documentation and can successfully use it. I use vmx-bundle-20.2R1-S3.2

When I try to use the startup config feature, as soon as I activate it, wipe the node and start it, then it fails after only a few seconds.
The node then shows an exclamation mark, but without a tool tip or any further info, what's wrong.
I tried this a couple of times, it is reproducible that it works without startup config and it fails as soon as I enable it.

Can anyone please help me to troubleshoot this? I feel pretty lost due to the lack of error messages. According to the docs, it should it be possible to use this feature with vMX.

Thanks for your help
Steffen

[StephenKing]

Hi,

anyone at least with a suggestion, where I could look at? There's nothing helpful in /opt/unetlab/data/Logs/.

Thanks
Steffen

[dvsuray1]

I have the same thing with 18.4.

I guess, the reason could be in one of python scripts in /opt/unetlab/scripts : config_vmx.py or config_vmxvcp.py - i dont know which one controls the config upload process...

Dmitriy

[Uldis (UD)]

script works only for VCP and it is correct
VFP no need it at all

[Uldis (UD)]

it works mate, and 19 and 20 versions

[StephenKing]

it works mate, and 19 and 20 versions

Thanks for the feedback. That's good to read that it works at least for some of us. Can you help me to understand, what might be different between your setup and ours?

I've attached to screen shots of my setup.

[Uldis (UD)]

You must export config, and then you will see wat format it has been exported..
Simply adding word TEXT is not ok

[StephenKing]

Thanks Uldis, this question was too obvious.. Of course, I also tried it with a complete configuration, see attached screen shot.

It does not matter, what the configuration is. A freshly wiped VCP aborts after ~10sec, later tries after 2-3sec. So during that time, the vMX never boots.
It's also impossible for me to get a console during that time (Connection Error - Remote Desktop Server unavailable).

I haven't had the chance to check the mentioned python scripts, will do that after the holidays.

Thanks for additional ideas
Steffen

[StephenKing]

I found the command that unl_wrapper uses to invoke qemu in wrapper.txt

Booting with activated startup config fails because the CD-ROM device containing the config.iso is added as index=0. Manually invoking the qemu-system-x86_64 command gives:

> qemu-system-x86_64: -drive file=config.iso,if=virtio,media=cdrom,index=2: drive with bus=0, unit=2 (index=2) exists

Changing it to index=3 at least boots Junos in a VM.

I'm just lacking ideas, how I could fix this. I checked unl_wrapper, which I assume as the source of this command, but this is obfuscated PHP code. So I will try the Live Helpdesk the next days as I paid for the PRO version and cannot help myself

Or has anyone further ideas?

Thanks
Steffen