Shocking speed of WD Black preclear

[ldrax]

I have a 3-cycle 4TB WD Black preclear in progress, just completed the first one a minute ago, the speed was totally not what I expected of this drive:

 

Last Cycle`s Pre-Read Time: 9:37:17 @ 115 MB/s.
Last Cycle`s Zeroing Time: 17:02:33 @ 65 MB/s.
Last Cycle`s Post-Read Time: 8:47:04 @ 126 MB/s.
Last Cycle`s Elapsed TIme: 35:27:01

 

The last ‘normal 3TB Seagate drive’ I precleared two months ago averaging 180-200 MB/s in pre, zero, and postread!

 

This WD drive still has 1 month warranty, should I stop the preclear cycle now and request for RMA?

 

[JonathanM]

Were you using the exact same version of preclear? There have been umpteen million different iterations of preclear, some markedly faster than others.

 

Those numbers don't look too out of line, attach diagnostics taken before you reboot if you want a more thorough evaluation.

[ldrax]

I happened to use the “auto update” on the GUI to update the preclear plugin, prior to running this last session.

[Fireball3]

Imho it's better to start with an extended smart test

smartctl -t -long /dev/sd...

If this test is not positive, you can stop all other actions.

 

If you consider RMAing the drive you need proof that something is wrong.

WD insists on a test with their Data Lifeguard Diagnostics Tool.

 

IIRC the improvement of the preclear speeds affected the post-read phase but I might have

lost track at some point.

You can give the comandline-script a chance and see if it makes a difference.

 

Edit:

You can check the progress

smartctl -a /dev/sdx

 

or look for the "remaining"

smartctl -a /dev/sdx | grep remaining

 

Edited July 20, 2018 by Fireball3

[pwm]

Strange that the zero speed is half the read speed.

 

I currently do not have any 3.5" WD Black online I can test on, but the following is read tests with a 2.5" 1TB WD Black:

pwm@europium:~$ sudo hdparm -t /dev/sda

/dev/sda:
 Timing buffered disk reads: 376 MB in  3.00 seconds = 125.17 MB/sec
pwm@europium:~$ sudo hdparm -T /dev/sda

/dev/sda:
 Timing cached reads:   1986 MB in  2.00 seconds = 993.08 MB/sec

pwm@europium:~$ sudo dd if=/dev/zero of=/mnt/data1/test bs=1M count=10K
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 131.017 s, 82.0 MB/s

I can't see a reason why a 4TB 3.5" WD Black should be slower than a 1TB 2,5" drive unless the drive possibly have issues with vibrations, cables, ...

 

Note that vibrations can make the drive regularly decide to rotate an additional revolution because it doesn't like the track alignment before starting to write.

[ldrax]

To give it benefit of a doubt, I let cycle 2 carry on last night.

The 25% report was 14hours ago, and now the preclear seems to be stalling. I check the system log and dmesg, there is no error or message logged. The GUI status is still stuck showing the progress of 33% (which is the same state as when I checked not long after receiving the 25% progress email).
I'm puzzled, to say the least.


Subject: Pre-Read in Progress on sdb.
Description: Pre-Read in Progress on sdb: 25% @ 147 MB/s. Temp: 51 C. Cycle 2 of 3.
Importance: normal

Pre-Read in progress on sdb: 25% complete.
Read 1,000,752,545,792 of 4,000,787,030,016 @ 147 MB/s
Disk temperature: 51 C
Cycle`s Elapsed Time: 1:53:14.
Total Elapsed time: 37:20:20.