1TB Drives showing up as different sizes

I have 7 Western Digital WD10EACS drives.  3 show up as 976,761,496 bytes, and 4 show up as 976,762,552 bytes. 

I had assumed this was normal, until someone posted that all 1TB drives are now supposed to be standardized to the same size.  I can't find the post now.  This wasn't a problem until I decided to upgrade the array and to use a faster drive (a Seagate 1TB 7200.11 32mb) for my parity.  With my luck, the Seagate came up at the smaller size (976,761,496).

Is this normal?  Are others seeing differnt sizes on their 1TB drives?

I have 2 of the Seagates and 2 of the WD.

They are all 976,762,552 (as reported by unRAID)

I had heard some reports that the WD varied in size, but this is the first I'm hearing tha the Seagates vary in size.

As I was writing this post, I realized that all 500G+ drive sizes are ending in "552",

500G = 488,386,552

750G = 732,574,552

1T = 976,762,552

Weird.

You may be seeing drives with an HPA, Host Protected Area.  The first 2 times I saw this were here and here.  I have since seen a number of them.  Although odd, they don't appear to be harmful, and usually only involve the loss of access to about a megabyte or two at the far end of the drive.  Your deficit seems much smaller than that.

If it is an HPA, there should be a message in your syslog at the first configuration of the drive, about HPA found.

Recent versions of hdparm can remove the HPA, but I would be very careful, and take the drive completely off line first, and be sure all data on the drive is backed up.

By the way, those numbers aren't in bytes, they are in unRAID blocks, which are equal to 2 sectors, therefore 1K in size.  You actually have over a thousand times that many bytes.

The only power of two I can find ending in "552" is 2^41 or 2TB.

Interestingly, 2^40 (1TB) ends in "776", so ...776552 is double weird.

Of course, there is a simple mathematical relationship explaining this, but I don't stress my brain that much on weekends.

BTW, all of my 1TB WD's end in 552 and those aren't bytes (unless you are running 1GB drives).

Bill

Thanks for the responses especially regarding the fact that the "size" measurement is related to sectors.  The specs for the Seagate states 1,953,525,168 "guaranteed sectors," yet an hdparm -I of the drive only shows 1,953,523,055 LBA48 user addressable sectors.  According the the hdparm of the other drives, this only amounts to a 1MB difference, but in the world of unRAID, that's enough to create a stop condition.

I *really* want this faster drive to be my parity drive, so I'm going to see what Seagate says to an RMA.

(Likewise, Western Digital's specs states 1,953,525,168)

The weird thing, though, is that each of the "smaller" drives are all reporting the exact same number of sectors.  If the units were bad, I would expect different/random numbers of sectors.

EDIT:

"Bad" is not the right word because I know the drives themselves are fine.  I've been using 3 of the "smaller" WDs since December without any problems.  I should have said "out of spec" instead of "bad."

My seagate and WD drives all report - 976,762,552.

If you contact seagate, just tell them that you are mirroring the drives and the difference sizes prevent the drives from being mirrored.

Thanks for the responses especially regarding the fact that the "size" measurement is related to sectors.  The specs for the Seagate states 1,953,525,168 "guaranteed sectors," yet an hdparm -I of the drive only shows 1,953,523,055 LBA48 user addressable sectors.  According the the hdparm of the other drives, this only amounts to a 1MB difference, but in the world of unRAID, that's enough to create a stop condition.

If you check that link I gave (specifically the last paragraph of this link - http://lime-technology.com/forum/index.php?topic=1837.msg13138#msg13138), you will see the exact same numbers as the ones with an HPA, so I have to assume that you will also find a note about an HPA in your syslog.  That HPA may possibly be removed with hdparm (described in the very next post by WeeboTech) and the -N parameter.  A search turned up a few other tools, like HDat2, Atola CapacityRestore, PARemove, etc.  This subject needs a little more research...

I pulled the syslog, and it is HPA. 

Seagate's SeaTools (disk utilities) has a function to reset the drive to its native capacity.  The function errored out when the drive was connected directly to the motherboard (Gigabyte GA-945GZM-S2), but it worked when I connected the drive through the Promise TX4 card.  I did a soft reboot, and SeaTools continued to report the drive at the correct (larger size).  However, after reconnecting the drives the original way and booting unRAID, it reported the smaller size, and re-running SeaTools also showed the smaller size.

I currently have the drive connected through the Promise card.  unRAID is seeing the larger capacity, and I'm letting it calculate parity.  After it finishes, I'm going to reconnect the drives the old way and see if the correct size continues to be reported.

Is it possible that my motherboard could be the issue?  In searching, I see that the board can implement a recovery function that requires a HPA be created, but the area would probably be much larger and I'd think it would need to be specifically invoked.

The question I would be asking is why are these HPAs being created on your disks.  You have 3 WDs + 1 Seagate with HPAs. 

I did some Googling and it sounds like some nefarious software may create these HPAs to conceal themselves.  Musicmann, any idea what is creating these?

The question I would be asking is why are these HPAs being created on your disks.  You have 3 WDs + 1 Seagate with HPAs. 

 

I did some Googling and it sounds like some nefarious software may create these HPAs to conceal themselves.  Musicmann, any idea what is creating these?

The new Seagate was just ordered from Fry's.  It was a retail kit, and the drive was put directly in my unRAID machine straight from the box.  I'm running a stock unRAID (4.2) with absolutely no addons, so I don't see how software could be causing them.  That's why I'm wondering if the motherboard itself could possibly be doing it.

The WDs were all purchased on eBay.  They were all supposedly new, but at least one seems to have come out of a WD MyBook based on its serial number/warranty info.  Another already had a filesystem on it, but this may not be uncommon given that the smartctl showed that it had zero hours of use, and I didn't check the others.

At the time, I just thought the different sizes was just a normal variation, so I didn't pay much attention to where/when the smaller drives occured.  However, I do know that the very last WD added to the array (back in December) and the Seagate were both orginally added to the 1st SATA port on the mobo.  This was because I had added all the data drives and the last drive I was adding was parity back in December, and I ended up having to swap drives because the parity appeared as smaller.  However, I don't know where the other two WDs were originally connected.  I know it wasn't to this same port, but I don't know whether it was to the mobo or to the Promise card.

Maybe the MB has a feature for storing some configuration data there.  I remember an old laptop I had used a part of the hard disk to store a set of advanced diagnostics accessible from the BIOS menu.  But it was its own partition and I was able to see it (and could have deleted it if I had wanted to).

I just purchased 2x1Tb Seagates, and the same thing happened with me..  Most drives have spare sectors just in case any blocks fail, perhaps it has something to do with that?

I don't think that is related.  I have no HPA but one of my 2 1T Seagate drives has remapped one bad sector.  I do not think it is possible to "reclaim" space that the drive his reserved for sector remapping.

Check out THIS link on HPA.

It sounds like perhaps something that some MBs may do, or it could be that Seagate is doing this on purpose for some other reason.

After my parity finished, I switched the new drive back to the original port and I lost the sectors again! 

I really wasn't expecting that because that port has had a viable parity drive (correct size) on there before.  I'm not sure what it is.  I wouldn't see why it would be the drive because why would the same drive model produce different results.

I had already ordered a new motherboard (also Gigabyte) as part of the upgrades I'm doing now.  I'm going to wait and install that and then start toying with this again.

Thanks for all the input.

From that WIKI, here is some information which may be of use.

Using Linux, there are a couple of ways to detect the existence of an HPA. The latest Linux versions will print a message when the system is booting. For example:

dmesg | less

[...]

hdb: Host Protected Area detected.

    current capacity is 12000 sectors (6 MB)

    native  capacity is 120103200 sectors (61492 MB)

After my parity finished, I switched the new drive back to the original port and I lost the sectors again! 

 

I really wasn't expecting that because that port has had a viable parity drive (correct size) on there before.  I'm not sure what it is.  I wouldn't see why it would be the drive because why would the same drive model produce different results.

 

I had already ordered a new motherboard (also Gigabyte) as part of the upgrades I'm doing now.  I'm going to wait and install that and then start toying with this again.

 

Thanks for all the input.

I doubt it's going to be the motherboard. I've got a Gigbyte board, but I've tried the hard drives on different boards to be sure. It still came up with the missing blocks.. I have noticed that the smaller drive makes occasional clicking noises, which is indicative of possible failure. I'll let you know how that goes. It would be good to test the rebuild while I still have everything backed up on other drives!

possible failure

Try definite.  I have never heard of clicking NOT portending the end of a drive.

I have seven (7) Seagate Barracuda 7200.11 ST31000340AS 1TB drives in my array.  Some were purchased from Newegg as OEM, some as retail kits from Frys, and some as retail kits from MicroCenter (I couldn't tell you which ones are which).  When I was first setting up unRaid, six of them reported the smaller size, so I used the 'larger' one as my parity drive.

I'm using a Gigabyte GA-EP35-DS3R MB...

Come to think of it, those first two drives with the 65 temps were replacement drives from an RMA from NewEgg.  The first two were DOA from shipping.  Those temps were reported when I first installed the drives fresh out of the box.  I do not trust those temps.  They must be on par with the other 40 degree drives.

When I was first setting up unRaid, six of them reported the smaller size, so I used the 'larger' one as my parity drive.

I did pretty much the same thing except with WD GP drives.  I really wouldn't mind the issue except that I specifically bought the Seagate to be used for parity since it should be faster.  Secondly, it potentially creates issues if I ever need to replace a drives.

HDAT2 seems to have had more success than SeaTools.  With it, I was able to set the drive to native size while the drive was plugged directly to the motherboard, and the size survived a reboot.  I did a parity check and upon completion, it stated 128 errors.  I guess this could be due to fiddling with the drive size, so I'm currently running another parity check.

The second parity check came back with 0 errors.  I also did a "full" power cycle including disconnecting all power for about 10 minutes and then rebooting.  It looks like I'm now locked in at the correct size.  Unfortunately, it's too late to cancel my 3 WD RMAs.  I guess I'll be out some shipping costs.

Looks like HDAT2 didn't solve the problem after all, and it does seem to be a motherboard issue.  When I was playing around with SeaTools and HDAT2, I had turned off auto detection for the new Seagate Drive, and I forgot to turn it back on.  I eventually noticed it when POST only showed 3 drives on my motherboard (although unRAID still saw the drive).  As soon as I turned detection back on and rebooted, the drive space vanished again.

My new motherboard and parts are in now, so after my current parity sync, I'm going to just rebuild, cross my fingers, and hope for some luck.

After changing from a GA-945GZM-S2 to a GA-P35C-DS3R, the new board does the same thing with drives.  I didn't extensively test, but it seems that, if the BIOS was set to use SATA drives in PATA mode, it would create an HPA.  I'm not sure if this happened on all the onboard ports as I was only focusing on the first port (Port 0) on both boards.  Also, it didn't happen with every drive as I was able to keep one of the TB drives as parity on Port 0.  I don't know if some of drives have the HPA feature disabled in a way that the motherboard could not override it, but all the model numbers for the WD drives are the same.

I had been using the PATA mode because, as far as I can tell, that's the only way to use a combo of PATA and SATA drives, and I liked having a CD-ROM attached.  Now I've disconnected the IDE CD-ROM and have set the board to use AHCI mode. 

Interestingly, the hard drives appear on POST but appear as "None" under Standard CMOS Features in the BIOS even though they are set to Auto detect (but I guess that is under "Extended IDE Detection," so maybe that's why).  The drives also appeared as "None" when I had Auto Detect off in PATA mode, and the drives would mount without an HPA being created.