Garbage IronWolf NAS PRO write performance

Hardware
1

So I was trying to wipe new IronWolf 8TB NAS PRO from Seagate with dd if=/dev/zero of=/dev/sda status=progress conv=fdatasync and I’m getting really trash results.
Drive is connected through SATA. I was expecting better results than 40MB/s top write speed even with all NAS dedicated fetures the drive has.
Here are some tests with 2,5 GB file.

75s 35.8 MB/s	bs=8K count=327680
77s 34.6 MB/s	bs=16K count=163840
74s 36.2 MB/s	bs=32K count=81920
78s 34.4 MB/s	bs=64K count=40960
73s 36.9 MB/s	bs=128K count=20480
66s 40.5 MB/s	bs=256K count=10240
71s 37.9 MB/s	bs=512K count=5120
73s 36.8 MB/s	bs=1M count=2560
71s 37.5 MB/s	bs=2M count=1280
72s 37.3 MB/s	bs=4M count=640
65s 41.3 MB/s	bs=8M count=320
71s 37.7 MB/s	bs=16M count=160
72s 37.0 MB/s	bs=32M count=80
71s 37.4 MB/s	bs=64M count=40
73s 36.4 MB/s	bs=128M count=20

Is it normal for these drives to be so slow during wipe?
According to specsheet sustained data transfer rate OD is 200 MiB/s max.

2

Did you try gnome disk’s built in benchmarks too?

Throughput does look low, for sure.

I would check cables, make sure connectors are fine, power is oaky, another SATA socket / drive bay etc.

Also, does it complete smaller/shorter tests okay? like, the drives often have a few hundred meg ram cache, so can you test that reads nice and fast?

3

Not sure what kind of system you plugged it into. Some operating systems have a lot of automated SMART stuff. Make sure it isn’t running something like a SMART Extended Offline test. And check that something didn’t give it a RAID Hotspare label and is already zeroing it.

Just a couple of weird things that might happen.

1 Like
4

@Trooper_ish
With USB 3.1 docking station the drive is ~25-30 slower.
SATA ports and cables makes no difference - checked with second machine even.
Reads when I checked with real movie file (~1GB) looked OK - copying to drive and playback.

As for the benchmarking. First thing I ran diagnostics with GSmartControl:

  • Overall Health Self-Assessment Test - PASSED
  • Sector Size - 512 bytes logical, 4096 bytes physical
  • Write Cache - Enabled
  • Smartctl Version 6.6 2017-11-05 r4594

Last one looks old?

In Attributes drive reports 3 warnings:
ID# ATTRIBUTE_NAME FLAGS VALUE WORST THRESH FAIL RAW_VALUE

  • 5 Reallocated_Sector_Ct PO–CK 053 053 010 - 48400 (0 1)
  • 197 Current_Pending_Sector -O–C- 095 091 000 - 2664
  • 198 Offline_Uncorrectable ----C- 095 091 000 - 2664

In Advanced -> Physical I found:

  • Complete phy log:
  • SATA Phy Event Counters (GP Log 0x11)
  • ID Size Value Description
  • 0x000a 2 2 Device-to-host register FISes sent due to a COMRESET
  • 0x0001 2 0 Command failed due to ICRC error
  • 0x0003 2 0 R_ERR response for device-to-host data FIS
  • 0x0004 2 0 R_ERR response for host-to-device data FIS
  • 0x0006 2 0 R_ERR response for device-to-host non-data FIS
  • 0x0007 2 0 R_ERR response for host-to-device non-data FIS

@zlynx
Sorry for not mentioning. I’m on Debian, Threadripper 3xxx platform. For the services that might be relevant there are gvfsd and udisks. The first time the drive was connected to the machine was through docking station.

For RAID I have 6 other drives in RAID 10 - made with mdadm:
sudo mdadm --create /dev/md0 --level=10 --raid-devices=6 /dev/sd[c-h]1
sudo dd if=/dev/urandom of=/dev/md0 bs=64M status=progress conv=fdatasync

5

Am I reading the Reallocated Sector Count correctly and it has a raw value of 48,400? And it is a new drive?

I think it may have been damaged in shipping.

6

Yes its 48400, and its a new drive.

7

I’ve got 4 WD Red 6 TB drives which are mostly three years old now and their counts are zero.

2 Likes
8

Those speeds are definitely not what they should be, however:

A lot of things going on here. Just want to check that you dd'd from /dev/zero to the bare drive while it was connected via SATA. If that performance is sub 100MB/s, then the drive is bad (or less likely, the controller or obscure OS issue).

Have you run badblocks on it?

1 Like
9

Oh good double check. Those speeds are what we would expect to see from USB-2.

1 Like
10

@oO.o
Yes the performance is sub 100MB/s via SATA - the tests from first post were done that way. I have 8 sata ports, 6 are populated with RAID 10, and 2 are free. The IronWolf is connected to one of the free ones.
I have 6 older HGST drives with 0 issues regarding speed and performance.

Running it now.

@zlynx
First Seagate HDD after 10 years and its already making problems …

1 Like
11

Yeah that looks like a bad drive to me. Do you have another system you can test it on (preferably not linux)? That would be the nail in the coffin.

How was the packaging? I’ve definitely gotten drives that someone had swapped out and returned before from Amazon. Looking closer at SMART values that @zlynx pointed out, all of these should be zero on a new drive:

What is the power cycle count and/or powered on hours (forgot how they phrase it in SMART)?

I have a lot of Ironwolfs running and they’ve been fine overall. I think you just got a lemon.

1 Like
12

Power Cycle is 14 times, and that is roughly the times I powered it on.
The package was bad thou. Cardboard box and one layer of bubble wrap.

1 Like
13

Yeah, just return it and say it has a arrived with a lot of bad sectors. If you run into any issues there, I have successfully RMA’d drives simply for having bad sectors before, you’d just have to wait a while for the replacement.

1 Like
14

Couldnt finish badblocks the speed was around 4MB/s and it could take an eternity.
CrystalDiskInfo in Windows™ shows 53 reallocated sectors.

I guess thanks for troubleshooting, I’m filing RMA ticket.

1 Like
15

This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.