Cluster Pi [gav-pi9 & gav-pi4 / 13 / 14 / 15]

Cluster Pi

One thing that I've always been interested in doig - for no reason other than I've never done it before, was to set up a Raspberry Pi cluster.

I know that this is no super computer - that's not the point... the point was to learn how to cluster computers together and get them to do something remotely (possibly) useful.

There are many ways to do this, ranging in cost from not much to lots and lots. I decided to play with the Cluster HAT which is "not much" range.

Cluster Pi

The Cluster HAT is basically a USB hub which allows up to 4 Raspberry Pi Zero (or Zero W) to be attached in a headless configuration, using USB Gadget mode to create a network interface between the host (Raspberry Pi 3 in my case) and the worker nodes. The HAT also has power control to allow sequenced powering up of the worker nodes, and provides easy access to the console pins.

Cluster Pi

This was a good use of my Pi Zero which was sitting doing nothing, but I "had" to buy 3 more to complete the hat. The old one is an original v1.2 board (purchased within 5 minutes of them going on sale), and the 3 new ones are v1.3 (with the camera header).

So, what does it do? ... well I had a brief experiment with Munge and Slurm before pushing the boat out and attempting kubernetes but the Pi zeros lack resources to do anything with them... the k3s use it all up before you can deploy anything. So instead, I'm running the cluster as a web server - so not really a cluster, but a load balancer. Perhaps one day I'll get a Turing Pi and construct a far more powerful Rapsberry Pi cluster!

Specifications

Case: Cluster HAT Case
Controller:
Board: Rapsberry Pi 3B (version 1.2)
SoC: Broadcom BCM2837
CPU: 64-bit Quad-core (ARM Cortex-A53, ARMv8-A arch) @ 1.2Ghz
GPU: Broadcom VideoCore IV @ 400 MHz
RAM: 1GiByte LPDDR2 SDRAM
HAT: Cluster HAT (v2.4)
Workers:
Board: Rapsberry Pi Zero (version 1.2 [x1] & 1.3 [x3])
SoC: Broadcom BCM2835
CPU: ARM1176JZF-S (ARM11, ARMv6Z arch) @ 1Ghz
GPU: Broadcom VideoCore IV @ 250 MHz
RAM: 512MiBytes SDRAM
Storage: 5x Sandisk Ultra, class 10 (16GByte micro SDHC - one per Pi)
Sandisk Ultra Fit (64GByte USB 3.1 - shared storage)

Helpful Resources




Del.ico.us Digg Facebook Google LinkedIn LiveJournal NewsVine reddit StumbleUpon Twitter
Valid XHTML 1.0 Transitional Valid CSS! [Valid Atom 1.0] [Valid RSS 2.0]
[ Page last updated Fri 15th Sep 2023 | viewed 971 times ]