Awesome Tips I Built a PC that CAN’T Fail… and You Can Too!
To learn more about the power and future of Intel’s Xeon processors, check them out at:
To learn more about the power and future of Intel’s Xeon processors, check them out at:
Our servers need to stay up and running throughout the day for us to keep doing what we do. So we built additional servers to get rid of any down time if something were to crash. And we’re here to show you that you can do it too.
Discuss on the forum:
Proxmox Mesh Networking Tutorial:
Proxmox Clustering Tutorial:
Linstor/DRBD Proxmox Tutorial:
Proxmox Ceph Tutorial:
Buy a Gigabyte R163-SG2-AAC1 Rack Mount Servers:
Buy Patriot Burst Elite 480GB SATA SSDs:
Buy an Intel Xeon Platinum 8562Y+ CPU:
Buy Micron 5600MT/s CL46 96GB Registered ECC DDR5 RAM:
Buy a KIOXIA CD6-R Series SSD:
Buy a NVIDIA ConnectX-6 Dual Port 100GbE Network Card:
Buy a Dell PowerSwitch S5232F-ON Network Switch:
Purchases made through some store links may provide some compensation to Linus Media Group.
► GET MERCH:
► GET EXCLUSIVE CONTENT ON FLOATPLANE:
► GET A VPN:
► SPONSORS, AFFILIATES, AND PARTNERS:
► EQUIPMENT WE USE TO FILM LTT:
► OUR WAN PODCAST GEAR:
FOLLOW US
—————————————————
Twitter:
Facebook:
Instagram:
TikTok:
Twitch:
MUSIC CREDIT
—————————————————
Intro: Laszlo – Supernova
Video Link:
iTunes Download Link:
Artist Link:
Outro: Approaching Nirvana – Sugar High
Video Link:
Listen on Spotify:
Artist Link:
Intro animation by MBarek Abdelwassaa
Monitor And Keyboard by vadimmihalkevich / CC BY 4.0
Mechanical RGB Keyboard by BigBrotherECE / CC BY 4.0
Mouse Gamer free Model By Oscar Creativo / CC BY 4.0
CHAPTERS
—————————————————
0:00 Intro
1:50 How did the cat teleport?
2:25 The servers
3:27 Installing the CPU
5:30 RAM and networking
7:26 The rack
8:47 Clustering
12:28 Does it work?
14:37 A proper demonstration
16:05 Virtualization magic
17:50 You can do it too
18:40 Outro
Deals for Days. Big home savings are happening now.
4:30 Linus says "very very long time"
shows "up to 1000 hrs" < 42 days!
"process was repeated EVERY 500 hrs" … so it was repeated ONCE?
1000hrs/500hrs=2, so it was done twice, which means repeated one time.
What a complete nonsense , AD runs distributed anyway. just add enough domain controllers.. Linus knows bog shit
kominfo need to see this
Really cool that this has become open source now.
I remember setting up lab with a VMware HA cluster with at home with two ESXi nodes and an Openfiler NAS with iSCSI for shared storage in 2009. Not the most reliable of solutions but it was free (software) and it worked. 😅
PC that can't fail. Engineers will tell you that is complete BS.
That's amazing. Also Linus seems so much happier since he stepped down. You can see how much less stress he is under.
"Sponsored by Intel" lmao
The title of the video is completely misleading.
Intel need to wake up and start making better cpu that can last….. paying people not fixing it…. but everyone know linus is intel employee
Love this kind of videos as a network/infrastructure technician ❤
This is so cool, holy hell.
This video is just Basic N+1 Clustering
It’ll still crash if the software crashes…
does anyone know anything about that blue dell case at 1:23
You should use 4 14900k systems to maximally test the reliability of this configuration 🙂
I love it when Jake is showing us more and more network/server functionality. I would love to see Jake host videos like these!
Welcome to the mid-2010s?
A "technology" company that's about 10 years behind the industry is. . .disconcerting. Makes me wonder how they do backups and other types of disaster recovery.
So they build something new what is already used by many people for years with either Proxmox/Sphex, VMware/VSAN, or Hyper-V converged cluster.
Running this setup with 3 nodes at home.
But the best solution for a NO BSOD screen is a Ryzen 5950X system (running since day one without BSOD).
Linus so rich he hired his boss…
Wait Linus isn't the boss of Linus tech tips??!
What was old will become new again. Every 4 to 5 years a new flush of grads comes out of Uni and "disrupts" the status quo. Except this happens every 4 – 5 years, so in fact any given flush of grads is just reverting what the last cadre did to what the cadre before that did. It's tedious.
But then once you get beyond the conceptual choice of VDI vs running local – there really is nothing else but hype lipstick on an age old pig. Pick your generation of pig because nothing has fundamentally changed in this space since the 80s. It's always been – "this time with more hype and product placement".
That holding the cable up to his ear bit was hilarious 😂
For some reason when my pc restarts it hangs on dram but if I shut it down it’s fine. Ryzen 7950x 64gb which I know could take ages to memory train but more than 45 min?
go buy yourself an 8th gen core cpu, theyre cheap…….
that hit me right in the gut [still using an 8700k]
Is this the first time Linus has done anything with virtualization? It's like 20 years old now..
i am not in IT but for a medium sized business, can this act work for DR for applications and databases? to resume everything like nothing happened in case the main servers went down?
FYI you should be using 2 different makes/models of SSD for mirroring, that way they shouldn't risk failing around the same time.
the pc im watching this on IS my dads old dell, its mine now
I work for a hosting company. Thanks to virtualization and HA clusters, our services will often be back up following a hardware failure before the operations technicians even engage the admins. If it were bare metal, we would have to dispatch someone to the data center to identify and replace the defect, which would substantially increase the downtime associated with hardware failure.
So what exactly is going on? It was pitched as if the systems were talking at all times and hosting like a networked raid 1 that would jump to a different machine when one shut down but what you talk about in the beginning and what came up just did not make any sense at the end. Maybe I’m way out of touch but I really don’t understand what is supposed to happen.
Gee, I run Linux Mint 22 Cinnamon, doesn't ever crash, games much better than Windows, is much safer, I don't have to have an anti-virus sucking away system resources, and it doesn't have the mystery navigation, oh and it's prettier lol.
I had some problems with the Dell switches as they run on 8900 MTU and not 9000 and the SAN refuzed to mount the partitions in ESXI until i realized and change it to 8900 🙂
Yall remember the last time linus showed off a $10k Xeon CPU? 😅
I don't want to dog on this too much, but this has been doable with vMotion/HA on vSphere forever. I was actually expecting some magic that would buffer ram/storage state, and get close to realtime HA. Have you guys not been running with any HA for routers/dns/etc? Good lord.
I'm sorry, but this video is totally disingenuous. Yes, that technology is great, but it isn't new. It has been used in datacanters for years. The purpose is load-balancing: If you run a monolithic app (like a game server), which requires very few VCPUs 90% of the time, but has huge spikes 10% of the time, where you need 10 times as much, you can use this tech to avoid having to pay for 90% of CPUs idling, but instead get your system to upscale within seconds and a downtime of milliseconds. Does it help with failing hardware? No! Not at all. Because your failing hardware doesn't tell you seconds in advance, that it's going to fail. So no matter what you do, your service will be interrupted and the current state will be lost (i.e. the people running the game on your server will be dropped out of the game and all of their advances, since the last real save, will be lost). Additionally the vast majority of crashes happen due to software failures, rather than hardware failures, which will not make a difference, on which CPUs you crash on. Sure it has it's use cases, but preventing a crash/failure is surely not one of them.