I have an array of 4 Seagate IronWolf Pro drives in this Unraid 7.2.2 system.

When copying large files from another system to the Array, I'm seeing about 50 - 55 MB/s write speed.
The CPU is running 6-10%, and the network interface is running about 500 Mbps.

The "theoretical" write rate of IronWolf Pro is about 200 MB/s. Is there something wrong or misconfigured, or is this all these drives can actually deliver in a real-world Unraid array?

Here's the results of a copy process writing a group of 100 2 GB files to the array from another computer over the network. Everything else is idle on the Unraid system. The sending computer is not CPU limited or network limited.

I'm curious why the read rate is about the same as the write rate? Is that expected?

Diagnostics attached: t440-diagnostics-20260118-1201.zip

Yes, this is the typical write speed to a parity enabled Array. Each write requires a read of the data drive being written and the parity drive(s), a parity calculation, then a write to the data drive and parity drives(s).

If you have a cache set up for the Share being written, it does "speed up" the write access time - But just to make your "While you wait" at the keyboard time go more quickly. The write speed will be the same when the Mover transfers the data from Cache -> Array.

Edited January 18Jan 18 by ConnerVT
speeling

Another option is to enable Turbo Write. This spins up all array disks, but it also significantly speeds up write operations in a parity-protected array. There is also an Auto Turbo Write plugin which manages disk spin up to facilitate Turbo Write and does not require that all array disks always be spun up.

Details on Turbo Write:

Edited January 18Jan 18 by Hoopster

Getting back to the slow write issue on my array
Last large copy to the array: Logs analysis by AI: "Every one of those disks is topping out around 21.6 MB/s despite being large 28TB/24TB Exos-class drives rated for 250+ MB/s sequential. That gap is the real story — the physical disks aren't spinning at their limit, something is throttling how much work reaches them."

I set md_sync_limit to 10.

reconstruct_write was already set.

During copying activity, array activity goes up to 75%, and CPU rises from its nominal of 10% to about 40% (The copy is from a Windows VM to a share on the array, so anything on the Windows VM uses lots of CPI

Is there anything changed in Unraid 7.2.8 that would help in this area? Have any other techniques for improving array performance been identified?

In your original post the screenshot showed your array writes to be just over 50MB/s, which I replied is pretty typical (and aligned with what I see on my server).

Ignoring what AI told you (as its answer doesn't directly answer what you are seeing), 20MB/s is not typical. Things that can influence this are other activity taking place on the array (reads or writes from another source/application) or something related to your source (you mention coming from a VM).

I would try copying a fairly large file internally on the server, such as from a pool drive (SSD/NVMe) to the array, using the native Unraid file utilities. This would help narrow down if the issue is you array or the mechanism trying to copy data.

The first screen shows normal speed for non turbo write, post a new screen with turbo write enabled (reconstruct write) while writing a large file.