I have three Raspberry Pis in my home network One is running my Home Assistant instance, as well as Deconz and MQTT It’s basically my smart home hub The other one is running my internal reverse proxy, Unifi controller, AdGuard home and DDNS, basically stuff that concerns my home network Finally, my third Raspberry Pi is running my PiKVM instance, Which gives me a way to manage my main server remotely, change its BIOS settings, Reinstall the operating system, and all of that without having to splurge on an enterprise server motherboard with an extra iKVM chip But today, everything changes Today, I will be replacing all of it with this This is Turing Pi 2 A cluster board that can fit up to four Raspberry Pi Compute Modules 4 It’s also compatible with NVIDIA Jetson boards And you can even mix and match CM4s and Jetsons To have a cluster that can do both general purpose computing and machine learning tasks. I am really excited to try it, so let’s build up our cluster and see what it can do! But first, I want to say huge thanks to folks from Turing Pi. They sent me the board, including three Compute Modules 4 as well as an NVIDIA Jetson board – for free. As usual, this doesn’t mean that I’m only going to say nice things about this board – It’s basically a prototype unit with a fair share of quirks and flaws, all of which we’re gonna talk about in this video So let’s talk about the board itself It’s sized as a standard MiniITX board, So it will fit in any standard PC case One thing that you have to keep in mind is that you’ll need at least 51 milimeters of clearance, Or 2 inches for my American friends For exmaple, It won’t fit in 1U rack enclosures As you can see, Turing Pi 2 hat lots of I/O, and each CM4 module gets some of it The first Pi has access to the miniPCIe slot and GPIO The second one also gets a miniPCIe slot, but without a SIM card slot The third Pi has access to two SATA slots, And finally, the fourth Pi gets the USB ports, both internal and external Compute Module 4 only has one PCI Express v2 lane, And I think the Turing Pi folks found a very creative way to utilize all of those lanes. Obviously, it has its drawbacks For example, the device you connect to the miniPCIe slot of Pi 1 will not be accessible by the Pi 2 But at the same time, I imagine that it was necessary For keeping that standard miniITX form-factor And making sure that any peripherals that you would connect to the Pis would run at full speed The board also has an built-in Ethernet switch Which connects all of the Pis to the network with those two Ethernet ports The ports are bridged and connected to the same 1Gbit interface, So you can just use one of them When you connect Turing Pi to your network, Each Pi will get its own unique IP and will be identified by its own MAC address So in order to use CM4 modules with Turing Pi, You’ll need to use these adapters They kind of look like SODIMM memory modules for a laptop, except they can basically fit an entire computer The CM4 module clips onto the adapter with a satisfying click, and that’s it! The adapters also have microSD card slots for Compute Module Lite models that don’t have built-in storage. To power the cluster, I’m gonna be using this wide input PicoPSU And a 60W Lenovo charger that I frankensteined a barrel plug on You can use any ATX power supply, But since the board doesn’t need much power, PicoPSU is the best choice in my opinion Once all four Compute Modules are clipped on to the adapters, it’s time to assemble the cluster! So some of you guys who’re not familiar with the idea of cluster computing, might be wondering – why bother with four low power and low performance computers Instead of just building one high power machine and using it to run a bunch of VMs? First and one of the most important reasons for me is power efficiency Even with alll four nodes running at full blast, Turing Pi 2 only consumes around 22 watts While doing normal tasks and running some Docker containers, this number goes down to 11 watts If I were to use an x86 machine with a similar level of performance, it would consume anywhere from 15 to 65 watts Second, redundancy and availability If you host all of your stuff on one computer, You’re basically putting all of your eggs into one basket If one day you have do some maintenance on the computer, Or if a kernel upgrade breaks your OS, Or maybe if one of the components fail You can say goodbye to all of the services hosted on that machine Of course, with Turing Pi you also have a single point of failure to some extent – All four nodes use the same power supply, and the same Ethernet connection, But at the same time they have independent resources Like RAM, storage, CPU and most importantly, each of them run an indpedendent operating system That makes Turing Pi a great choice For running a high availability Kubernetes or Docker Swarm cluster And you can even hotswap nodes on the go, without having to power down the whole cluster, Although I’ve been told by Turing Pi engineers not to do that Third, flexibility and I/O With Raspberry Pi you have access to things like GPIO, SPI, serial ports and DSI Turing Pi exposes all of those things, Letting you use devices that you wouldn’t be able to easily use on a standard x86 machine Like this Zigbee adapter Spoiler alert – GPIO doesn’t really work in the current hardware/firmware revision, But it should work on actual production units One more point that I would normally include would be price And if by the time you’re watching this video Compute Modules 4 are back in stock and are sold at MSRP, I guess the point is valid But as of now, it’s actually cheaper to buy four of these thin clients, Which will also have better performance than a Raspberry Pi. Sure, the power efficiency is not really there, A thin client like that consumes around 10 to 15 watts And multiplied by 4 you get around 40 to 60 watts, But considering the price difference, even when CM4 units are in stock It will probably take a while to recoup the electricity costs So now that our cluster is built, I’ll tell you a little bit about what I’m planning to do with it. I have an actual big “NAS” that runs things like PhotoPrism, Plex, Sonarr and Radarr. It also used to run a PiHole instance, Home Assistant and reverse proxy, But I discovered pretty quickly that running all of my services on one machine is a bad idea. If I have to do some maintenance on my server, I end up with no DNS, no light or heating automation, and no access to any of my services. So I started running my mission critical services on a separate Raspberry Pi 4. At the same time, I needed another Raspberry Pi for PiKVM, Since the developers don’t have any plans of releasing it as a docker container of some sort, And only offer PiKVM as a standalone OS image. And then I also added another Raspberry Pi that only runs smart home stuff. At the end, like I said in the beginning, I ended up with three Raspberry Pis. So instead, I’m planning to run all my stuff on Turing Pi 2. I also want to use it to learn clustering software, like Docker Swarm and Kubernetes, To make some of my services, like DNS, highly available. Since this board also supports NVIDIA Jetson modules, I might eventually use it to run Plex. The SoC installed on those boards doesn’t support NVENC, But there are ffmpeg builds on Github that add a hardware decoding support for those boards. One more thing I wanted to do is put the board into a rackmount case. Yes, I actually have a server rack now, Vut I won’t show it to you just yet because it’s still a work in progress. Make sure to subscribe if you don’t want to miss that video. I found this noname 2U case on eBay for 54 euro shipped. It came in a pretty rough shape – full of dirt and dust, but nothing some compressed air and isopropyl alcohol couldn’t fix It also came with a 400 watt Seasonic power supply. However, with only an 80 plus efficiency rating and a jet engine fan… …It’s not really a good fit for this machine. I’ll just use my PicoPSU instead. Please don’t shout at me, I know that it’s a very good industrial grade power supply, But this machine will not be running in a factory, it’ll be in my office. And I prefer silence. As I mentioned before, Turing Pi 2 won’t fit in a 1U enclosure because of the height limitations, But you could definitely go with a shorter case. I got this one because it’s cheap, and the rack I have is deep enough to fit it. After mounting the board in the case, I also installed this external fan hub. In its current revision, Turing Pi v2 doesn’t have a PWM fan controller, So instead, I’m going to use SATA to power the built in 80mm case fan, And control its voltage with this dial. The compute modules don’t get that hot, so I can get away with just running the fam at… Fam… What’s up, fam The compute modules don’t get that hot, So I can get away with just running the fan at its lowest RPM. Upon further inspection I found that the 12V rail on the fan connector works, but it’s reversed for some reason. So after bending the pins and forcing my fan plug in the connector, the fan works, So no need for an external SATA adapter, I guess. One more thing that doesn’t work in the current revision of Turing Pi are the front panel connectors. I tried pretty much all combinations to connect the power button, but none worked. At the end, it turned out that I need to update the board’s firmware, which requires a JTAG programmer And that’s not something I’m going to do now. The board comes on automatically when you connect it to power anyway, so that will do for now. As a work around, I connected power LED to the Vbus of the front panel USB3 connector, To at least get some indication that the board is on. So there you have it! It’s kind of a shame to hide all of those other blinkenlights, like Ethernet or storage activity, But hopefully I can figure out how to output them to the front panel soon enough. So what do I think about the board? Well, it’s definitely a work in progress. A lot of functionality that is promised in the official blog posts and press releases is missing Such as the Ethernet switch management, proper GPIO layout, fan control, and even some pretty essential things like front panel pins. However, the biggest reason to be skeptical about Turing Pi’s release Has nothing to do with how good or bad the board itself is But the component shortage. The Kickstarter for the board has been delayed multiple times, because the creators of Turing Pi are simply not able to produce enough units. One of the engineers mentioned that they eventually gave up on including the fan speed controller, Because the part required for it is impossible to find for a regular price. The availability and pricing of Raspberry Pi Compute Modules hasn’t been great in the “normal” times But nowadays you’d be lucky to even find regular Raspberry Pis in stock. And even when they are in stock, they usually cost double the MSRP. And now that the component shortage is further exacerbated by the Russian invasion of Ukraine, The fate of Turing Pi 2 remains unclear. I gotta be honest with you, I’m kinda getting tired living through these major historical events Anyway, I really hope that Turing Pi 2 takes off, And I’m definitely looking forward to the Kickstarter campaign. Video Information
This video, titled ‘Turing Pi 2 – The Ultimate Home Server?’, was uploaded by Wolfgang’s Channel on 2022-05-04 08:01:36. It has garnered 125076 views and 3611 likes. The duration of the video is 00:10:35 or 635 seconds.
Turing Pi 2 https://turingpi.com/
Follow me: Mastodon https://mastodon.social/web/@notthebee GitHub https://github.com/notthebee Twitch https://twitch.com/notthebeee
Support the channel: Patreon https://patreon.com/WolfgangsChannel PayPal (one time donation) https://www.paypal.com/donate/?hosted_button_id=S46R487SBMJZ8
Music: Aves – Day Trip Meod – Crispy Cone Less Gravity – Still Awake Kitrano – Empty City Aves – Fried Eggs
Videos are edited with Davinci Resolve Studio. I use Affinity Photo for thumbnails and Ableton Live for audio editing.
Video gear: Camera https://amzn.to/3gfcQYo Main lens https://amzn.to/3cvCozn Microphone https://amzn.to/3pAARgI Key light https://amzn.to/2TK374E Softbox https://amzn.to/3v6mBxt Secondary light https://amzn.to/3w7KLJg
Other stuff that I use: Monitor https://amzn.to/36zNekk Monitor arm https://amzn.to/3ASFKXD Laptop stand https://amzn.to/3eahdDJ Keyboard https://www.amazon.de/HHKB-PD-KB401W-Tastenkappen-Professionelle-Mechanische/dp/B082TXQD1S Mouse https://amzn.to/2Vp6NcU Audio interface https://amzn.to/3AS6Jmm
As an Amazon Associate, I earn from qualifying purchases
Timestamps: 00:00 Introduction 01:15 The Turing Pi 2 board 03:10 Why build a cluster? 05:29 My use case 06:38 Rackmounting the board 09:08 Final thoughts, chip shortage, etc.