Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Turing Pi: Kubernetes Cluster on Your Desk

turingpi.com

Post reply on HN
Print view View original

41–50 of 65 posts

Re: Turing Pi: Kubernetes Cluster on Your Desk

#41

Post by 0x202020 »

hinkley wrote: post #19
rubyn00bie wrote: post #14

While Raspberry Pi's are awesome, and the power consumption is nothing to scoff at (when considering a cluster), you can accomplish this same thing for a lot less, and have quite a lot more compute power, by purchasing a used server or even something like an AMD 3600X... A single 3600X will grossly outperform this cluster (and cost less) with less headaches (you don't have N physical machines) by using KVM to deploy…

The other part of 'on your desk' is hearing damage. Server hardware vendors have traditionally not given two shits about their servers being north of 90 decibels, and I'm pretty sure I've witnessed a few that were pushing 100. That Raspberry Pi is probably going to absorb more noise than it makes.

I managed a few HP C7000 blade systems once in an inadequate environment. They have jet like fans and at the time a firmware bug that meant the fans had two levels: 79% and 100%. Compared to after the firmware was fixed it stayed around 30% at full load and was bearable at least

Re: Turing Pi: Kubernetes Cluster on Your Desk

#42

Post by nsky-world »

kube-system wrote: post #39
nsky-world wrote: post #25

Earlier quoted context omitted.

What's the practical point of running containers on VM?

To expand on my personal use-case, I don't want or need a whole stack of physical servers sitting around just for a test k8s environment. There are some things that you can only really test properly in a real multi-node environment rather than single-node solutions like minikube: failovers, shared storage, some networking particularities, etc. But there are reasons to run containers on VMs in production too. Hypervis…

Can you please elaborate a little bit more on the reason to run containers on VMs in production?

Re: Turing Pi: Kubernetes Cluster on Your Desk

#43

Post by kube-system »

nsky-world wrote: post #42
kube-system wrote: post #39

Earlier quoted context omitted.

To expand on my personal use-case, I don't want or need a whole stack of physical servers sitting around just for a test k8s environment. There are some things that you can only really test properly in a real multi-node environment rather than single-node solutions like minikube: failovers, shared storage, some networking particularities, etc. But there are reasons to run containers on VMs in production too. Hypervis…

Can you please elaborate a little bit more on the reason to run containers on VMs in production?

Lets say you have a complicated distributed application that needs to run on multiple cloud providers, and maybe it needs to run on-prem sometimes.

You could write your application multiple times, and publish multiple images for the same application across different infrastructures.

Or, just build your application once to run in containers, and run those containers on k8s everywhere.

Re: Turing Pi: Kubernetes Cluster on Your Desk

#44

Post by kjs3 »

0x202020 wrote: post #41
hinkley wrote: post #19

Earlier quoted context omitted.

The other part of 'on your desk' is hearing damage. Server hardware vendors have traditionally not given two shits about their servers being north of 90 decibels, and I'm pretty sure I've witnessed a few that were pushing 100. That Raspberry Pi is probably going to absorb more noise than it makes.

I managed a few HP C7000 blade systems once in an inadequate environment. They have jet like fans and at the time a firmware bug that meant the fans had two levels: 79% and 100%. Compared to after the firmware was fixed it stayed around 30% at full load and was bearable at least

Oh, man...reminds me of the IBM BladeCenters. Maybe they had 2 fan settings...90% and 100% of a 737 at takeoff. They had an extra cost "office kit" that included baffles and other mods to (hahahahaha) quiet them down enough for an office environment. Maybe knocked it down from a 737 to a Gulfstream G4. Nice kit if buried in a datacenter someplace.

Re: Turing Pi: Kubernetes Cluster on Your Desk

#47

Post by justinclift »

LargoLasskhyfv wrote: post #26

There is also the https://www.pine64.org/clusterboard/ for $99 which takes these modules https://www.pine64.org/sopine/ at $29 a piece, which are quad-core ARM Cortex A53 with 2GB LPDDR3. This is their wiki: https://wiki.pine64.org/index.php/PINE_A64-LTS/SOPine

That looks really interesting :), but it seems like they have an pretty-much-blocker bug with the individual modules not turning on again when rebooted:

https://forum.pine64.org/showthread.php?tid=5849

:(

Re: Turing Pi: Kubernetes Cluster on Your Desk

#48

Post by moondev »

detaro wrote: post #9
moondev wrote: post #2

Looking at the specs, it seems almost dishonest to promote this for kubernetes. > The nodes interconnected with the onboard 1 Gbps switch. However, each node is limited with 100 Mbps USB speed. Not only that but Compute Module 3+ are limited to 1GB RAM, is it really expected someone could run a realistic workload? How stable is the control plane node with such limited resources? It seems like picking up 3 raspi4s (4G…

If you insist on using Pis (IMHO you're quickly at the point where a (potentially used) NUC or small office PC is the better choice), why would you go for extra expense and effort for PoE? Just connect a 5V PSU to the power pins? I guess it gives you individual power switching without any DIY, but other than that?

Simply for not having to deal with all the usb-c adaptors and cables.

It's also pretty cool being able to turn the pis on or off by cycling the switch port from your network control plane (Unifi in my case)

You could even do it dynamically via the api for a "bare-metal autoscaling" type workflow.

Re: Turing Pi: Kubernetes Cluster on Your Desk

#49

Post by tuananh »

for this use case, i would rather use something like SimpleNUC[0] .

It's basically just NUC machines with rack mount. It's a lot more powerful than Rpi, quite power-efficient (when compare to actual server rack) and dead silent.

More importantly, this setup can handle some actual workload

[0]: https://simplynuc.com/server-shelf-solution/

Re: Turing Pi: Kubernetes Cluster on Your Desk

#50

Post by Proven »

hinkley wrote: post #19
rubyn00bie wrote: post #14

While Raspberry Pi's are awesome, and the power consumption is nothing to scoff at (when considering a cluster), you can accomplish this same thing for a lot less, and have quite a lot more compute power, by purchasing a used server or even something like an AMD 3600X... A single 3600X will grossly outperform this cluster (and cost less) with less headaches (you don't have N physical machines) by using KVM to deploy…

The other part of 'on your desk' is hearing damage. Server hardware vendors have traditionally not given two shits about their servers being north of 90 decibels, and I'm pretty sure I've witnessed a few that were pushing 100. That Raspberry Pi is probably going to absorb more noise than it makes.

The same can work in a VM without any extra noise (and with probably much less heat dissipation)
Post reply on HN