> ME? Moving My Hyper-V VMs From One Array To Another In Preparation To Replace Two SMART-Warned SSDs!! YUO? No Fuqing Clue What I Am Talking About.

Me? Buys reliable equipment.

Spanky wrote:

S/F/O



Yeah, I burned the SSDs out my setting up the array wrong and being lazy about fixing it.

The SSDs were supposed to be a cache tier backed by spinning rust. Instead I set it up where all four drives to be actual storage.

I will fix that since I am replacing the drives.

Expensive mistake these days.  PI harvesting requires that much I/O?

Spanky wrote:

The SSDs were supposed to be a cache tier backed by spinning rust. Instead I set it up where all four drives to be actual storage.

What kind of array is this?

Phillip_McCavity wrote:

Spanky has amassed one heck of a porno collection.

But it's all of himself.

wrote:

What kind of array is this?

This is a Windows Storage Spaces array. It's a four-drive mirror with two SSD drives, each 4TB; and two NVMe drives, each 500GB. However, in WSS you can set this up two ways. The incorrect way I set it up for this use case was just a normal array. Every drive participates in every IOP. The way I wanted to set it up is using the NVMe as a cache layer and the SSDs as the persistent layer. This pulls a lot of IOPs off the SSDs.

CIFotoDumpRandom89265701.png

Spanky wrote:

Yeah. The 4TB SSDs were $100 cheaper each three years ago. And it was only $100 cheaper because I got the SSDs at half the current price due to one of those weird things on Amazon where the price was temporarily extremely low.

https://www.amazon.com/dp/B08QBL36GF

Currently $920!!

Wow.

When I bought these the normal price was $750. But I got them for $420. Shipped / sold by Amazon.

CIFotoDumpRandom31713916.png

Fuqing financed it!
   \


Yuo=poor!
   \

Hyper-V is… okay. Switched from VMware over a year ago because of Broadcom fuqery. HA clusters are a little less robust, but since we already had windows server licenses anyway, it just made sense.

Fritz wrote:

Hyper-V is… okay. Switched from VMware over a year ago because of Broadcom fuqery. HA clusters are a little less robust, but since we already had windows server licenses anyway, it just made sense.

Simple peripheral pass through is the big gap.

I heard in 2025 that GPU partitioning is a lot better.

That is consumer-grade garbage. Notice how spanky pisses away many $1000s on cheap consumer shyt from amazon instead of building a real home-lab with enterprise-grade hardware.

wrote:

That is consumer-grade garbage. Notice how spanky pisses away many $1000s on cheap consumer shyt from amazon instead of building a real home-lab with enterprise-grade hardware.

Why would anyone want a home lab?