2

I bought a Seagate Expansion Portable 2TB (current version) recently.

I've begun writing some data to it, and it's about 5% full now. But, the write speed was getting awful. At first, I thought this was due to SMR, but after learning more about SMR, I think that it shouldn't affect read speeds, and write speeds should only be affected after writing dozens of GB continuously, and should recover after idling for <10 minutes.

This drive has been idling for hours, and I currently get around 2MB/s reads and writes.

Every diagnostic, including Seatools from Seagate, and SMART, passes it with no issues.

The performance problem happens both on Windows and Linux, on different machines, and on different USB ports on each machine (both USB 3 and 2).

It certainly looks like it's a faulty drive, but I don't understand why no diagnostic, including Seagate's own, pick up any issues. Surely 2MB/s reads is an issue?

For some reasons on the below invocation of CrystalDiskMark the write speeds are zero, but usually they were around 2MB/s to 6MB/s.

------------------------------------------------------------------------------
CrystalDiskMark 7.0.0 x64 (C) 2007-2019 hiyohiyo
                                  Crystal Dew World: https://crystalmark.info/
------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

[Read]
Sequential 1MiB (Q=  8, T= 1):     2.096 MB/s [      2.0 IOPS] <2529502.09 us>
Sequential 1MiB (Q=  1, T= 1):     1.046 MB/s [      1.0 IOPS] <868586.28 us>
    Random 4KiB (Q= 32, T=16):     0.002 MB/s [      0.5 IOPS] < 19650.23 us>
    Random 4KiB (Q=  1, T= 1):     0.125 MB/s [     30.5 IOPS] < 32266.59 us>

[Write]
Sequential 1MiB (Q=  8, T= 1):     0.000 MB/s [      0.0 IOPS] <    -1.00 us>
Sequential 1MiB (Q=  1, T= 1):     0.000 MB/s [      0.0 IOPS] <    -1.00 us>
    Random 4KiB (Q= 32, T=16):     0.000 MB/s [      0.0 IOPS] <    -1.00 us>
    Random 4KiB (Q=  1, T= 1):     0.000 MB/s [      0.0 IOPS] <    -1.00 us>

Profile: Default
   Test: 128 MiB (x1) [Interval: 0 sec] <DefaultAffinity=DISABLED>
   Date: 2021/06/07 14:17:46
     OS: Windows 10 Enterprise [10.0 Build 19042] (x64)
4
  • I wonder if it is a faulty usb to sata converter in the dribe thats negotiating down to USB1 speeds? I posit tools that interrogate the drive would not see this problem as they are looking at SMART info which is internal to the drive.
    – davidgo
    Commented Jun 7, 2021 at 5:08
  • Here's some weird news. A quick format in Windows 10 has resolved the performance issue, at least for now - 105MB/s read and write, and since then I've written 80GB and counting. I have no idea why a format would affect this since much of my measuring (eg using dd or the disk manager tool in Linux) was accessing the drive directly, not its partitions. The situation could degrade again; I'll keep an eye on it before I write up my solution here or anything. Commented Jun 7, 2021 at 6:23
  • I've never seen the degraded performance again. Something weird about the factory format? Who knows Commented May 10, 2023 at 0:14
  • It could be to do with the alignment of sectors - sometimes USB enclosures map a have different geometry to the physical disk, and maybe that was causing issues?
    – davidgo
    Commented May 10, 2023 at 0:26

1 Answer 1

2

So, the problem has gone away after doing a quick format in Windows 10. Performance has jumped up to around 105MB/s reads and writes, and it's sustained even after copying hundreds of GB.

I don't have any explanation why doing this should have altered the read and write speed of the drive. It was a brand new drive. And, many of the ways I had tested read/write speed were accessing the raw drive and bypassing the formatting anyway (eg, dd).

But since it's appropriate to give the solution I found for myself, here is my answer, such as it is.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .