After a few months of playing with Kubernetes on the OrangePi CM4s, I realized that running anything requiring decent CPU power wasn’t going to work very well. For example, hosting WordPress wasn’t as viable as I had hoped. Granted, WordPress can be pretty resource-intensive, but still—it was a bottleneck. That’s why I eventually settled on Ghost, which is much lighter on resources. More on that later!
At the time, there weren’t any CM5s out yet , and while the Turing Pi supports Nvidia Jetson Orin and Nano, both seemed more focused on AI/ML, which wasn’t what I needed. I was simply looking for more processing power.
Luckily, as part of the Kickstarter program I mentioned in another post, Turing Pi v2 kickstarter did so well they decided to create their own SoM—the RK1. It boasted a much faster processor than both the OrangePi CM4 and Raspberry Pi CM4, with the added bonus of some ML/AI capabilities (which I’ll touch on in a future post). The RK1 has great specs, so I decided to grab a full set for my cluster.
After the tedious experience of manually flashing each OrangePi CM4, it was a game-changer to be able to flash the RK1s directly from the BMC! You can find more info on the RK1 here.