I started a 5 disk (all 2TB) unRAID server build with an old Gigabyte board I had laying around. I checked the BIOS and it did not indicate there were any BIOS protection schems or HPA issue.

 

Anyway, to make a long story short...when i went to assign disks in to the array, one of the disk's capacity ended in "492" while all others were "552": an indication of HPA at work. I have since used a new board.

 

My questions is, wht do I need to do to return the affected disk back to normal so its capacity reads "552" like the others?

 

Thanks!

Details on how to do this are listed in the wiki. I used it myself to remove hpa from a drive and it worked great. Still use the gigabyte board to.

Thanks Mazma, I used HDAT2 instead as I was getting nowhere fast using "HDPARM"

 

I'm wondering why preclear just doesn't clear/erase the ENTIRE drive and storage area, including any HPA data.

 

Regards,

The_keymaker

I'm wondering why preclear just doesn't clear/erase the ENTIRE drive and storage area, including any HPA data.

 

Regards,

The_keymaker

It has no ability to touch anything beyond the HPA.  The OS won't let it.

As far as the pre-clear script knows, it IS clearing the entire drive.

The HPA has legitimate uses to reduce the overall size of a disk. For example, using an HPA allows me to downsize 3T drives to 2.2T so they can be added to the array (this won't be necessary once unRaid supports GPT partitions, but for now is the only way to use these big disks in an array

Thanks for the extra insight guys.

 

It just seems there should be a simple option/command/operation (like HDAT2) that is part of unRAID or the Linux OS that would clear the HPA area when requested.

 

The_keymaker

Thanks for the extra insight guys.

 

It just seems there should be a simple option/command/operation (like HDAT2) that is part of unRAID or the Linux OS that would clear the HPA area when requested.

 

The_keymaker

There is.

 

The

hdparm -N pCORRECT_FULL_SIZE /dev/sdX

will sometimes do it.

 

The size can only be set once per power cycle of a disk.  On some disk controllers, specifically those that cannot read the full size correctly, hdparm is unlikely to work.  In those cases, you must use HDAT2 or Seatools or something similar.

I have a disk that shows as having HPA. After reading a few threads I attempted to remove it using HDAT2, however the disk does not show up in the drive listing in HDAT. It shows fine if I boot with UnRAID. Not sure what to do now. The only disk that does show up in HDAT2 is another precleared disk, that I have not added to the array, yet. None of the array disks show either.

 

The steps I took were to unassign it from the array, power down the system, then boot form my DOS USB stick to run HDAT. Tried to rerun (D)etect form HDAT and still no disk. If I then boot UnRAID it appears again.

 

This disk was precleared last night, UnRAID is set to use 4K-aligned. It was added to the array and formatted.

 

I don't see anything amiss in the syslog relating to HPA, but in the MyMain view in UmMenu it shows as HPA. Also hdparm-N gives the following error:

/dev/sdf:
max sectors   = 625142448/4385456(625142448?), HPA setting seems invalid (buggy kernel device driver?)

 

any ideas?

 

p.s. using X8SIL MoBo, but the disk has been used in my MacPro. previously.

syslog-2011-03-26.txt.zip