Good use for an x299 box?

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,400
Yet another lockdown inspired build... did the gaming system and the workstation in may and June, did the HTPC in August... VR system is wrapping up shortly... which leaves me bored again - just as we go into lockdown two.
So, with an X570, TRX40, x399, and Z490 already done... all that’s left is to build an X299 system! Because why the hell not. But I can’t think of what to use it for... local micro center has 10980s in stock and on sale.

Arcade cabinet setup?
hardline watercooling learner box?
Extra ESXi host?
Portable (LOL) lan box?

Trying to think of something creative to do that would give me a reason to build one last machine.
 

somebrains

[H]ard|Gawd
Joined
Nov 10, 2013
Messages
1,401
Mine was a Kubernetes lab.
128gb ram, 3x nvme drives, 4 ssds.

I ran discreet redis 5.0 install as a dummy cache on a dedicated samsung 960 Pro.

Ironic but I had to run MQ on the other 960 Pro bc I was testing DRs with an eye towards preserving data in flight.

N-tier web application, key value stores, mongo ran on ssds.
 

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,400
Oh now that’s a fascinating possibility I hadn’t considered, and actually need... damned brilliant idea!
 

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,400
Dammit. And there are even things I need to do on the CSI driver side. Why didn’t I think of that?!?
 

somebrains

[H]ard|Gawd
Joined
Nov 10, 2013
Messages
1,401
I mean, what would you like to learn that you can take advantage of the pcie bandwidth, ram density, and core ct?

I haven't spoken to any of the guys I used to work with at a gaming pub lately, but cracking the Intel multithread Dev guide was a big topic earlier this year.

A significant % of the full stack/BI/DBA set all dove at machine learning.

lots of VMware guys have to start planning on hybrid or at least drafting public cloud migration POCs.

Kubernetes is everywhere like Linux was everywhere 20 years ago. It's just an OS for all intents and purposes. The actual workflow of Docker containers from initial commit thru Prod is an arc we all benefit from understanding.

CICD is something you can work on. People still cling to 1 recipe and don't account for issues in build time, Sec checks, quality checks, etc thru multiple environments. That workflow in itself is the hot ticket right now.

We can all learn more networking.

We can all learn more storage tier architecting for a given DB and datastore use case.

Even the desktop support guys get to learn how AD/Ldap/Id federation remote at scale for a given quality level works.
 

lopoetve

Fully [H]
Joined
Oct 11, 2001
Messages
30,400
So I’m a datacenter architect. Up till last week, I ran a lab with 1800c/40TB of ram (and about a PB of all flash storage), all for my use and a couple of others. Virtualized networking, full BGP and spine leaf stack, you name it. I’ve designed and architected container solutions... just never spent much actual hands on time with Kubernetes for some reason- thats why this excites me.
Cloud and VDI expert, just... never got the hands on time with the container side once kube came out. Lots and lots of docker time, but... yeah. So there’s a gap I can fill!!!
 
Top