Before posting something, READ the changelog, WATCH the videos, howto and provide following:
Your install is: Bare metal, ESXi, what CPU model, RAM, HD, what EVE version you have, output of the uname -a and any other info that might help us faster.
Moderator: mike
-
yonokamayashika
- Posts: 2
- Joined: Fri May 01, 2026 11:41 am
Post
by yonokamayashika » Fri May 01, 2026 12:02 pm
I am building a new lab for my CCNP studies. I want to use Ansible to manage 10 Cisco routers. To do this, I am using a Docker node inside EVE-NG as my control station.
I have a technical question about resources. My server is getting very slow when I start the automation. I want to know:
- Image Choice: Do you prefer using a light Ubuntu Docker image for Python scripts, or is it better to use a full Linux VM node? Which one is more "stable" for long labs?
- CPU/RAM: What is the smallest amount of RAM I can give to an Ansible node without the scripts failing? I want to save space so I can add more routers later.
- KVM Settings: In 2026, are you still using "uksm" to save memory, or is there a better setting in the new EVE-NG version for handling many small Docker nodes?
I want my lab to be fast and professional. If you have a "pro-tip" for building automation labs, I would love to hear it!
-
rusty725
- Posts: 487
- Joined: Thu Mar 29, 2018 4:19 pm
Post
by rusty725 » Fri May 01, 2026 8:20 pm
What's your cpu model and how many ram do you have ?
-
yonokamayashika
- Posts: 2
- Joined: Fri May 01, 2026 11:41 am
Post
by yonokamayashika » Sat May 02, 2026 7:26 am
rusty725 wrote: ↑Fri May 01, 2026 8:20 pm
What's your cpu model and how many ram do you have ?
Thanks for your reply!
Here are my server specs:
- CPU: Intel Xeon E5-2678 v3 (12 cores / 24 threads)
Lab setup:
- 10 Cisco IOSv routers running inside EVE-NG
- 1 Docker node (Ubuntu) acting as the Ansible control node
- Ansible running with default forks (5)
Problem:
When I execute my playbooks, CPU usage spikes close to 100% and the lab becomes slow. Router consoles lag and sometimes tasks take much longer than expected.
Goal:
I would like to optimize resource usage so I can scale to ~15–20 routers while keeping the lab responsive.
Questions:
- Do you think the bottleneck here is CPU rather than RAM?
- Would reducing Ansible forks help stabilize performance?
- Are Docker nodes still the best choice for an Ansible control node or would a small Linux VM be more stable for longer sessions?
Also, if you have any tuning tips for EVE-NG (CPU pinning, UKSM alternatives, etc.), I would really appreciate it.
Thanks!
-
Uldis (UD)
- Posts: 5198
- Joined: Wed Mar 15, 2017 4:44 pm
- Location: London
-
Contact:
Post
by Uldis (UD) » Sat May 02, 2026 5:52 pm
First advice is to move away from vIOS, but use IOL roters which are far better and powerful.
Also in performance IOL wins vIOS about x200 times.