This is true. Really annoyed that arm as a hole isn’t being utilized like it could be by really anyone but apple. We could be making arm Linux powerhouses that sip power like a mid tier x86 laptop. The worry by some is that there is now way to do this without having every component solderd on, but dell has already made a new open laptop ram standard that has almost the same latency as Apple’s soldered ram.
Arm is the future, and needs to be treated as such more than it is.
Sounds like k3s would be right up your alley, it’s API compatible with k8s but has a lot less overhead than k8s, designed for use on low power devices like the Pi.
I found that for my use case (jellyfin, gitea, portainer, nextcloud, adguard, …) the pis are still nearly idle but the bottleneck for me was ram. Anyone with similar experience?
I setup a k8s rpi cluster for this reason, and now I just have 4 overloaded pis 🙃
Could you not just actually build a dedicated PC for that price? Lol
But then he won’t have a k8s rpi cluster
This is the real reason
I mean you could have a smaller cluster
and the power consumption adds up, too.
Pis are only 5W, right? 4 of them should still add up to about as much as a midweight laptop.
This is true. Really annoyed that arm as a hole isn’t being utilized like it could be by really anyone but apple. We could be making arm Linux powerhouses that sip power like a mid tier x86 laptop. The worry by some is that there is now way to do this without having every component solderd on, but dell has already made a new open laptop ram standard that has almost the same latency as Apple’s soldered ram.
Arm is the future, and needs to be treated as such more than it is.
I mean, it’s not just Apple, Google is all in on ARM and has been for like a decade and a half.
As for the laptop, look up framework
Yeah but Chromebooks suck, apple is making computers that aren’t just for web browsing
I do also have a dedicated PC as a NAS, the rpi cluster was more for learning. And k8s does provide some cool flexibility
Sounds like k3s would be right up your alley, it’s API compatible with k8s but has a lot less overhead than k8s, designed for use on low power devices like the Pi.
I found that for my use case (jellyfin, gitea, portainer, nextcloud, adguard, …) the pis are still nearly idle but the bottleneck for me was ram. Anyone with similar experience?