Ssd/hdd for server, which one is better?

Question

Solid State Drives

Good day to all!
On what (on what media) would you advise organizing data storage in the virtualization server?
Given :
Mother Supermicro XDRH-C/CT/i/iT
Stone Xeon E5-2630 x 2 pcs.
RAM 128 GB DDR4 ECC
Supermicro CSE-825 chassis (8 x 3.5" hot-swap SAS/SATA drive bay w/ full SES-II support on SAS motherboards, 2 x 3.5" fixed drive bay)
Server is planned to use hypervisor under Windows AD + MS SQL + 1C + a couple of Linux services. There was a question of data storage:
A colleague "in the shop" advises to take cheap (compared to server solutions) Samsung 850 Pro x6 pieces, assemble 50 raids from them and put a couple of ordinary SATA screws in the remaining 2 backing baskets.
I am inclined to believe that SSDs, no matter how super-reliable they are, they will die soon enough, and they will die almost simultaneously due to the raid. And I think that you should take more expensive enterprise-class HDDs.
Actually the question is: what to put in the baskets? What will be ideal in terms of price / quality / performance ratio?

Answer 1

On what (on what media) would you advise organizing data storage in the virtualization server?

Depends on the type of load.

The server is planned to use a hypervisor under Windows AD + MS SQL + 1C + a couple of Linux services.

SQL version 1c does not heavily load the disk, memory is effectively used there for caching, so a regular disk, of course, a dedicated one, will probably suffice. Though it is necessary to look that for basis and how many users.

A colleague "in the shop" advises to take cheap (compared to server solutions) Samsung 850 Pro x6 pieces, assemble 50 raids from them

There are doubts about the adequacy of this colleague.

I'm leaning towards the fact that SSDs, no matter how super reliable they are, they will die soon enough

Of course, any technique breaks down, so even super reliable HDDs, SSDs and other hardware will fail over time.

and will die almost simultaneously because of the raid.

Sounds pretty crazy... Why would it be almost at the same time? Simultaneous failure of two disks is unlikely, although it is possible, for example, if the power supply fails, or due to mechanical impact.
But certainly not because of the raid.

what to put in the baskets? What will be ideal in terms of price / quality / performance ratio?

Initially - one HDD under the OS, the other under the base 1s. Other services look at the load whether they need a disk or not. Then look at the load.
Reid is installed if there is a clear need for it - for example, to organize uninterrupted work.

Answer 2

With your initial data, 5 bases of 30-40 gigs. 70+ users
The hypervisor will slow down, I just recently went through this.
The record seems to be small, but if logging is output to a separate disk (just an SSD), the load on the raid will be small.
I recommend abandoning the hypervisor when it comes to 1C.