Check out sdparm.
Show everything with details and explanation:


I am very, very curious if you have @WCE secion now for all of your disks.

Write cache off and save (-s=set,-S=Save)


Write cache on and save:

Hi,
Any chance someone could help with changing the mode Page 0x38, line 0x14 as mentioned above.
I'm not that experienced with sg_wr_mode and hex editing.
Also have a few disks laying around, same model, firmware and control page output;

Using this thread, I finally managed to format a Hitachi VSP 5560075-A / ST6000NM0034 drive from 520B to 512B and create a working partition.
I used an LSI 9211 controller flashed in IT mode and formatted the drive to 512B using sg3-utils — the process took about 20 hours.

Afterward, I tried to modify Mode Page 0x38, byte 0x14, as mentioned above, but the command didn’t go through.
Then I switched from the LSI 9211 to a Dell PERC H730. The controller recognized the formatted HDD but couldn’t create a partition.

I set all HDDs to JBOD mode, and only then was I able to successfully change Mode Page 0x38, byte 0x14.
I checked my Mode Page 0x38 dump using:
sudo sg_modes -p 0x38 /dev/sgX # mode page 0x38 dump
And change byte 0x14 (the 21st byte) from 00 to 01.
sudo sg_wr_mode --page=0x38 --contents="B8,46,00,02,00,03,74,08,00,00,00,02,BA,A0,F4,B0,02,00,00,12,01,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00" /dev/sda
Essentially, this changes 21st byte from 00 to 01.
After that, I was able to successfully create a partition.
Thank you to everyone in this thread for providing valuable information.

Add -s to save.
sudo sg_wr_mode --page=0x38 --contents="B8,46,00,02,00,03,74,08,00,00,00,02,BA,A0,F4,B0,02,00,00,12,01,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00" -s /dev/sdX

I wonder if you are aware that the hard disks in this topic are SAS

_________________
USB 3.5" enclosures are $19 USD postage paid from China, I own several

I am going to add that the information in this tread is very close to being complete.

What you need to do in general is set bit 1 of byte 0x14 in mode page 0x38 to make the drives usage on ordinary systems after reformatting to a normal physical 512/4096 block size as appropriate.

On a Seagate drive this would appear to be changing 0x00 to 0x01 as detailed here. However on an HGST drive this is changing 0x02 to 0x03. I have only seen Seagate and HGST drives from a VSP system. Well not strictly true I have a number of SAS, SSD but these just work when reformatted to a normal block size, no extra steps required. For spinning disks it is Seagate and HGST only.

I have verified that this works across a number of different drive models from both Seagate and HGST.

I have also deduced that the product ID returned by the drive tells you the manufacturer, the physical size rotational speed and the capacity of the drive. So take the model format of

DKX#@-YZZZSS

Then

X is R for HGST and S for Seagate drives

# is either a 5 for a 2.5" drive or 2 for a 3.5" drive

@ is an uppercase letter exact meaning unknown

Y is H for a 7200 RPM drive, J for a 10k RPM drive and K for a 15k RPM drive

ZZZ indicates the capacity of the drive. If it is three digits then the capacity is in GB. If there is an R it is a decimal point and the capacity is in TB

so some examples for the last second part after the dash

K300SS is a 15k RPM drive with a capacity of 300GB

J900SS is a 10k RPM drive with a capacity of 900GB

J1R2SS is a 10k RPM drive with a capacity of 1.2TB

H4R0SS is a 7200 RPM drive with a capacity of 4.0TB

H10RSS is a 7200 RPM drive with a capacity of 10TB

The @ uppercase letter seems to indicate a drive series and is increasing over time. That is newer drives have a higher letter in the alphabet. Presumably you need to use a lookup table to work out what they refer to.

We have done these VSP drives before.

Formatting them from 520 to 512 is only half the job. The drive can format correctly, report 512-byte sectors, and still refuse normal writes because the VSP configuration is still present in Mode Page 0x38.
The change is byte 0x14.

For the drives we have seen:
Seagate: 0x00 -> 0x01
HGST: 0x02 -> 0x03

So the actual change is simple:
byte_0x14 |= 0x01;
That is bit 0, not bit 1.

For HGST 0x02 becomes 0x03. Setting the byte directly to 0x01 would clear the existing 0x02 flag.

Read Mode Page 0x38 from the drive you are working on:
sg_modes -p 0x38 /dev/sgX

Check what is changeable:
sg_modes -p 0x38 -c 1 /dev/sgX
Change byte 0x14, write the page back, save it, power-cycle the drive, then verify it again.

The process is basically:
Format 520 -> 512.
Read Mode Page 0x38.
Set bit 0 of byte 0x14.
Write and save it.
Power-cycle.
Confirm normal read/write operation.

Without altering my initial post - One thing I should add.

sg_wr_mode supports a mask, so alter only bit 0 of byte 0x14 and leave the rest of the drive's mode page untouched.
sg_wr_mode --page=0x38 \
--contents=0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1 \
--mask=0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1 \
--save /dev/sgX
That changes only 0x14 & 0x01.