EVE-NG goes shutdown after starting Cisco SD-WAN viptela-edge

[Ursl235]

Hello, appreciate your help!

I'm struggling with launching Cisco SD-WAN. I run EVE-NG in a VM using KVM on my Linux Debian laptop (8-core Ryzen 4750U, 32 GB DRAM). I enabled nested virtualization and installed EVE in a VM with 16 vCPU / 24 GB DRAM. IOL labs run in this VM just fine.

I have viptela images 19.2.0 like viptela-edge-19.2.0-genericx86-64.qcow2 . vedge/vbond, vmanage, vsmart are all have the same version.

I have followed the installation guide https://www.eve-ng.net/index.php/docume ... mages-set/ exactly. The only step I skipped is Step 4 because it requires creating additional 100GB storage, and my VM has virtual storage 100GB total, so no space for that. I need to create a new EVE VM because of this requirement.

After that, I created a lab and seen that Viptela images are available. I created a router and started it. I launched a terminal session and this is what I seen for a minute:
Trying 192.168.122.77...
Connected to eve.
Escape character is '^]'.

I monitored the VM in an SSH session to the VM using top command. I seen that the VM had ~100% CPU utilization for approximately 1 minute:

top - 21:16:47 up 3 min, 1 user, load average: 0.92, 0.40, 0.16
Tasks: 280 total, 1 running, 279 sleeping, 0 stopped, 0 zombie
%Cpu(s): 3.0 us, 3.3 sy, 0.0 ni, 93.7 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 24040.5 total, 1240.6 free, 22194.8 used, 605.1 buff/cache
MiB Swap: 8192.0 total, 8192.0 free, 0.0 used. 1505.8 avail Mem

After that, empty error messages appeared in the GUI and I noticed that the VM is in shutdown state.

Appreciate your advice -- is this behavior caused by missing Step 4 with 100GB HDD creation? Or it is more likely to be caused by a faulty image of the vedge?

If creating 100GB HDD is really necessary, I'll go ahead and reinstall the VM, but it would be best to do it for sure.

Thank you!

[Uldis (UD)]

100gb hdd need for vManage only, not for vEdge

[Ursl235]

Thank you, Uldis!

So it looks like this behavior might be caused by faulty vedge image. Will try another image and let know.

[Ursl235]

Uldis, I deletd viptela images 19.2.0 and installed 18.4.4. The behavior is the same -- after I start vedge, EVE VM goes shutdown after a minute. I noticed that during this time the vedge consumes all the DRAM:
top - 22:04:52 up 12 min, 1 user, load average: 0.70, 0.28, 0.15
Tasks: 258 total, 1 running, 257 sleeping, 0 stopped, 0 zombie
%Cpu(s): 2.0 us, 3.7 sy, 0.0 ni, 94.3 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 24040.5 total, 831.7 free, 20752.6 used, 2456.2 buff/cache
MiB Swap: 8192.0 total, 8192.0 free, 0.0 used. 2920.1 avail Mem

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
8640 root 20 0 1673932 51676 35064 S 100.0 0.2 1:05.60 qemu-system-x86
...

So it failed after ~800MB DRAM left of 24GB total. It looks like these vedge images take all available RAM of my VM and fail the VM..

vedge is the only node in the lab, no any other nodes, no network connections, nothing. Could it be caused by QEMU settings? I have
QEMU version 2.12.0
QEMU Arch x86_64
QEMU custom options: -machine type=pc,accel=kvm -cpu host -vga std -usbdevice tablet -boot order=dc -cpu host

Also, I started vsmart and the result is the same -- it consumes all available DRAM and after that EVE VM goes shutdown. No output in the console window while it happens:
Trying 192.168.122.77...
Connected to eve.
Escape character is '^]'.


Could be related to nested virtualization somehow? I run EVE bare metal image in KVM with nested virtualization on my primary system -- Linux Debian 12 laptop based on Ryzen 4750 CPU. Otherwise, I don't know what the root cause might be. However, it would be unfortunate to buy a separate machine to run EVE bare metal only to find out that the issue with SD-WAN is the same

Thank you for your advice!