System requirements

#1

Hi everyone!
It’s my first time on this forum, but I’ve been looking around lately.
I’m considering to install GenieACS on a machine for deployement. However I can’t seem to find any guide/mailing list/forum that specifies the actual system requirements (CPU, RAM, Disk Space, etc.). Has anyone got an advice?

Thanks for your attention, I’m looking forward to an answer.

(I’m sorry if I made any grammar error - I’m not a native speaker).

#2

I use VM for acs with
3GB RAM and 2 core CPU
I have 512 devices for every 30 min refresh

1 Like
#3

I’ll setup a VM too and I’m probably gonna have more clients (maybe I’ll lower the refresh rate).

Thanks a lot.

#4

I think my system will withstand more than 1k device.
Cpu never loaded more than 3%. Mem 60%, where 30% - mongodb

#5

Yeah I’ve just tested the new setup with the simulator at 200CPEs - 1min inform rate and Cpu never loaded above 5%.

Thanks for the feedback again!

#6

This page may help if you want to run benchmarks to determine your hardware requirements: https://github.com/genieacs/genieacs/wiki/Measuring-GenieACS-Performance

#7

Yes, I’ve used it. I was just wondering if there were some minimal system requirements. Anyways, the current configuration is 4GB RAM, 2 CORES and an LVM (currently 20 GB) and everything works perfectly fine.

Thanks for your beautiful work!

#8

samsam, i am working with device management since early adoption. Currently i am working on a system that holds 8-10 millions devices.
To specifically troubleshoot appplication performance, looking at PERIODIC inform events is not reliable.

In general if your provisioning logic is triggered by PERIODIC event inform, the general action for the application to perform on a database is just to update the last contact time of the device.

You should see how your application performs with event informs BOOTSTRAP, REBOOT, TRANSFER COMPLETE (after firmware upgrades).

Also setting a periodic interval on low value will impact your system the more devices you had to it pretty quickly. Unless you have a way to throttle your incoming data when the event informs are flooding into your system. It is always a best practice to start a periodic interval at a value and increase that value when the number of devices in your system under management increases. This is my own experience with TR069 after 10+ years working and similar systems.

1 Like
#9

Thanks a lot for your advice!