I was attempting to upgrade unraid to the latest 5.0 beta from version 4.5.  First I upgraded to 4.7, but when the server booted back up it said one of my disks (a 1tb drive) was too small.  Great!  I did a quick search and saw it was an HPA issue.  I telnet into the server to use hdparm to try and fix the disk. I was able to downgrade to 4.6 and everything was working fine.  All drives showed up as green and it was all hunky dory.  I verified the HPA feature was shut off in BIOS, then used telnet to hdparm.

 

I did:  hdparm -N p1953525168 /dev/sdf

 

BUT! I foolishly forgot to check which disk I was actually running this command on.  I accidentally used this to change the sectors on one of my 2tb drives, and not the 1tb drive as intended.  It was late at night and brain was not working.  After rebooting, now the 2tb drive is blue balled because the disk space is obviously larger than the other 2tb drives in the system.

 

I attempted to use hdparm to change the setting back to the original number of sectors, and it keeps failing.  I've tried rebooting multiple times and still doesn't work.

 

Is there anything I can do to fix it?  Luckily the 2tb drive is not my parity drive.  I ordered a 2tb to replace the disk and should be arriving in a couple days.  My plan as of now is to toss the 2tb drive I borked and pop in the new disk and rebuild the array.  I would like to salvage the old 2tb drive.

 

Thanks!!

I was attempting to upgrade unraid to the latest 5.0 beta from version 4.5.  First I upgraded to 4.7, but when the server booted back up it said one of my disks (a 1tb drive) was too small.  Great!  I did a quick search and saw it was an HPA issue.  I telnet into the server to use hdparm to try and fix the disk. I was able to downgrade to 4.6 and everything was working fine.  All drives showed up as green and it was all hunky dory.  I verified the HPA feature was shut off in BIOS, then used telnet to hdparm.

 

I did:  hdparm -N p1953525168 /dev/sdf

 

BUT! I foolishly forgot to check which disk I was actually running this command on.  I accidentally used this to change the sectors on one of my 2tb drives, and not the 1tb drive as intended.  It was late at night and brain was not working.  After rebooting, now the 2tb drive is blue balled because the disk space is obviously larger than the other 2tb drives in the system.

 

I attempted to use hdparm to change the setting back to the original number of sectors, and it keeps failing.  I've tried rebooting multiple times and still doesn't work.

 

Is there anything I can do to fix it?  Luckily the 2tb drive is not my parity drive.  I ordered a 2tb to replace the disk and should be arriving in a couple days.  My plan as of now is to toss the 2tb drive I borked and pop in the new disk and rebuild the array.  I would like to salvage the old 2tb drive.

 

Thanks!!

On most disks you can ONLY change the HPA once per power cycle.  So,

power down, then upon power up you can try to reset the HPA to its correct normal value.

 

You should be able to see the correct value by typing

hdparm -N /dev/sdX

 

The actual device names can change from one boot to the next, so always verify you have the correct disk.

 

 

Thanks Joe!  I'll give it a shot when I get home.  I only rebooted and didn't try a full power down in between.  I hope it works.

OK, I tried it and this is the output:

 

/dev/sde:

setting max visible sectors to 3907029168 (permanent)

SET_MAX_ADDRESS failed: Input/output error

max sectors  = 0/14715056, HPA is enabled

OK, I ended up having to use hdat2 through USB to change the sectors back.  Now my disk appears as red and the "started" has an orange ball next to it.  The button to check parity is gone.

OK, I figured it out.  Drive is rebuilding now.  I had to stop array, un-assign, start array, stop array, re-assign.

OK, I rebuilt the array and everything is green.  I'm going to continue using the 2tb drive i thought was borked. 

 

The new 2tb drive I ordered, arrived and upon installation in place of the 1tb drive with HPA, it showed slightly less free space than the other 2tb drives.  I guess that drive spot has HPA assigned to it.  I was looking into disabling HPA for my mobo (Gigabyte 965P ds3), and it looks impossible.

 

New plan is to just live with the HPA on the drive and hope nothing else happens.  From what I've read, HPA seems to be a big problem only if it's on the parity drive.  Any thoughts? 

OK, I rebuilt the array and everything is green.  I'm going to continue using the 2tb drive i thought was borked. 

 

The new 2tb drive I ordered, arrived and upon installation in place of the 1tb drive with HPA, it showed slightly less free space than the other 2tb drives.  I guess that drive spot has HPA assigned to it.  I was looking into disabling HPA for my mobo (Gigabyte 965P ds3), and it looks impossible.

 

New plan is to just live with the HPA on the drive and hope nothing else happens.  From what I've read, HPA seems to be a big problem only if it's on the parity drive.  Any thoughts?

You are incorrect... it is a problem on any disk.  If the one disk with the HPA dies, then the BIOS will write one to another.  (and either clobber parity, or a file-system on a data disk)  Basically, it is a ticking time-bomb... waiting for one of the disks in the array to fail, and eventually one will.

 

Find another MB, or see if the BIOS on your MB can be upgraded.

Got it.  I'm going to search for a new mobo now.  It will save me headaches in the future.

 

Thanks Joe!

I both agree and disagree with Joe's note 

 

It IS a problem if your BIOS doesn't allow disabling the HPA ... Joe's absolutely correct that it CAN write an HPA to new disks, and could easily cause a "not big enough" issue.

 

While I agree it's best to eventually get a new board, you CAN work around this issue as long as you understand it.    First, as Joe suggested, be sure you're using the latest BIOS update and that it still doesn't allow disabling the creation of the HPA.

 

To bypass this problem even if you can't disable the function in the BIOS, you need to understand Gigabyte's "behavior" during POST.  It searches all of the drives attached to the onboard SATA controllers for one that has an HPA with a copy of the BIOS.  If it finds one, then done !!    If not, it will create one on the first disk it detected.

 

So ... if you simply be sure that the disk you're using that HAS an HPA is connected to a motherboard port; then you won't have the issue anymore UNLESS that drive fails.  Of course, that drive could fail ... but there are two ways to mitigate the potential problem that would create:  (a)  Be sure your parity drive is NOT connected to a motherboard port (this eliminates the issue);  or (b) install a small, spare drive to SATA Port 0 on the motherboard (got an unused little laptop drive??).  This will almost certainly always be the first drive detected;  so if your drive with the HPA fails, this is what will then get an HPA.  You can, in fact, ensure this by simply disconnected all other drives; booting the system;  then shutting it down [if you check the blocks before/after that you'll noticed that an HPA was created on the drive].  You'll now have TWO drives with HPA's on them -- the unused (for UnRAID) laptop drive; and the 2GB drive you already have.    The odds of BOTH of them failing and causing yet-another HPA are fairly small.  In fact, you could assign the spare drive as your cache unit, since whether or not it has an HPA doesn't really matter [then you'd KNOW if it failed ... further reducing the odds that both it and your 2TB data drive would both fail at once.].

 

Sounds harder than it is ... basically, just install a single drive (to be your cache);  boot;  then shut down and connect all your drives, being sure the 2TB w/HPA is on a motherboard port;  and you're ready to go.    A lot less hassle (and expense) than replacing the motherboard 

 

... but when you DO decide to upgrade, don't get a board with an HPA problem !!!

 

Thanks for the info gary!