none
Problem Formatting Hard Drive Over 1 Tb RRS feed

  • Question

  • Running Windows Home Server on an ECS A780VM-M2 with the 780V/SB700 chipset, AMD 4850e and 2GB DDR2. I recently added two 1.5TB hard drives. With both my 1.5 Tb hard drives, I have the same problem. They will successfully add to the storage pool of WHS at full capacity. If I then run CHKDSK on the drive all capacity above 1 Tb is bad clusters. I have checked both HD with SeaTools for DOS and both pass long test, filled drives with zeros, and set capacity to max. In Disk Management a quick format works but if I do a full format everything seems to go well but the after the free space is only 1 Tb. I have updated to the latest bios and drivers with no change. I am running the SATA drives in IDE mode for easy install and reinstall of the operating system.
    I have checked the HDs on an Nvidia 8200 chipset XP pc and everything works great.
    Any ideas?

    Wednesday, August 26, 2009 9:42 PM

All replies

  • Running Windows Home Server on an ECS A780VM-M2 with the 780V/SB700 chipset, AMD 4850e and 2GB DDR2. I recently added two 1.5TB hard drives. With both my 1.5 Tb hard drives, I have the same problem. They will successfully add to the storage pool of WHS at full capacity. If I then run CHKDSK on the drive all capacity above 1 Tb is bad clusters. I have checked both HD with SeaTools for DOS and both pass long test, filled drives with zeros, and set capacity to max. In Disk Management a quick format works but if I do a full format everything seems to go well but the after the free space is only 1 Tb. I have updated to the latest bios and drivers with no change. I am running the SATA drives in IDE mode for easy install and reinstall of the operating system.
    I have checked the HDs on an Nvidia 8200 chipset XP pc and everything works great.
    Any ideas?


    My guess is there is something wrong with your SATA ports and/or the BIOS on your mobo (i.e. your SATA ports may only support up to 1TB in Legacy/IDE mode).  Although it's strange that you see the full 1.5 TB for each drive in the storage pool...

    You might try contacting ECS for support.
    Thursday, August 27, 2009 12:26 AM
    Moderator
  • I have 12 drives, a mixture of 750GB, 1TB and 1.5TB.  All work flawlessly in the pool on my Dell.  I know this isn't an answer, just wanted to assure you that larger drives work fine.  I have a total of over 13TB.
    Thursday, August 27, 2009 4:13 PM
  • Kariya21

    Yes I agree it does sound like a mobo/BIOS problem.  I did submit a case to ECS but have not gotten a response.

    Thanks

    Tom,

    My drives seem to work in the pool until I ran CHKDSK as described in the file conflicts thread.  The only indication was the number of bad clusters and less storage available in the pool.  You may want to check.

    I still think it is a mobo/BIOS problem.  But I will boot the server from BartPE and format and run CHKDSK and see what happens.  That will eliminate WHS as a suspect.  I will report back.

    Thanks

    Friday, August 28, 2009 2:21 PM
  • To my surprise, booting the WHS from BartPE built using WinXP the full format worked at full capacity!  Then I ran /r CHKDSK with no bad clusters.  Then I copied ~600 GB of data to the drive and re-ran /r CHKDSK again with no bad clusters.  I guess this clears the mobo and BIOS from suspicion.  Now WHS and drivers are suspect.  I will try to load the same drivers used in BartPE from WinXP
    Monday, August 31, 2009 2:52 AM
  • I have the same problem with a 1.5TB drive using an Intel DG965OT board. I also run the sata drives in IDE mode.
    I have successfully run a full format of the drive in Windows XP using the same model motherboard.
    I guess int's something to do with Server 2003.
    Wednesday, September 2, 2009 12:30 PM
  • Solved!

    I updated the Bios and installed the latest chipset drivers - still no luck.

    So I changed from IDE emulation to AHCI and, voila, I ran a full format with no errors - free space was 1.36TB.
    Just to verify, I ran a check disk with scan for bad sectors - again, no errors.

    A search of the net reveals a few problems occurring beyond the 1TB mark. Although the MBR disk format has a limit of 2TB based on 512 byte sectors and a maximum of 2^32 sectors, it would seem that some implementations have an effective 1TB limit, perhaps due to the use of signed rather than unsigned 32 bit integers. Solaris, for example, could not boot from a drive larger than 1TB (http://arc.opensolaris.org/caselog/PSARC/2008/336/20080520_shidokht.yadegari) - may be fixed now.

    Windows Server 2003 is implicated in a few cases, notably http://forums.seagate.com/stx/board/message?board.id=ata_drives&message.id=13190. Coincidentally, for me, the motherboard in this case used an Intel G965 chip.

    Thursday, September 3, 2009 12:09 PM
  • My bad.  I should have asked about ACHI.  Mine are all configured as native SATA, not compatibility mode
    Friday, September 4, 2009 12:20 AM
  • JohnSoftek,

    Thanks for the workaround solution.  I was hoping for a solution where I could leave my mobo in IDE mode for easy system restore if the system drive fails.  I guess I will need to a slipstream of WHS with my AHCI drivers.

    I just wonder how many people have added a drive that is greater than 1 TB in IDE mode and think everything is ok because they see the full format capacity.  But in fact only have 1 TB of capacity available.  Or even worse, if data can be written into the above 1 TB space and not retrieved correctly.  I have gotten file conflicts warnings after running chkdsk /x /r that were not fixed with a restart.  I had to remove the disk from the pool to get them to clear.  Luckily, I have full file duplication on.

    Does the WHS team have anything to say on the subject?
    Friday, September 4, 2009 4:01 PM
  • ...
    Does the WHS team have anything to say on the subject?
    Probably not, because the problem is in your computer's firmware, which doesn't properly support large disks in legacy IDE mode.
    I'm not on the WHS team, I just post a lot. :)
    Friday, September 4, 2009 5:24 PM
    Moderator
  • Ken,

    Please read my second post!  On my WHS I booted from BartPE built from WinXP and everything worked fine.  I think that clears my mobo bios from suspicion?
    Friday, September 4, 2009 5:35 PM
  • What I can tell you for sure is that Windows Home Server relies on the drivers that the manufacturer supplies to talk to the manufacturer's hardware. And those drivers rely on the firmware working correctly, both BIOS and firmware onboard on the drives. Windows Home Server itself doesn't know, or care, about the hardware at that level.

    What you describe still makes me think it's a firmware issue, but it could also be drivers. If it were Windows Home Server, I would probably have issues with my servers, which both have 1.5 TB drives installed. So would the thousands of other users with 1.5 TB or larger drives...

    I'm not on the WHS team, I just post a lot. :)
    Friday, September 4, 2009 8:43 PM
    Moderator
  • The evidence seems to clear the firmware of suspicion. Windows XP can successfully format a 1.5 TB drive using the same model motherboard in IDE emulation mode - and Emdrei had success with a BartPE boot disk on his system where WHS failed.

    Server 2003 has had other problems with disks larger than 1TB (http://support.microsoft.com/default.aspx/kb/906668). All it takes is for a piece of code to use a signed rather than unsigned 32 bit integer to refer to the sector number.
    • Edited by JohnSoftek Saturday, September 5, 2009 3:48 AM fix typo
    Saturday, September 5, 2009 12:33 AM
  • I suspect a few people may have experienced the problem and maybe even "fixed" it, but at the loss of some storage.

    WHS started reporting file conflicts after I added a 1.5TB drive to my system which already had a 500MB and 2x1TB drives. I didn't at first connect the file conflicts with the new drive. The event viewer showed error '55' on the 'DATA' disk. Since I had four disks labelled 'DATA', I ran chkdsk /x/r across all drives. This would have fixed the problem (no more file conflicts) since chkdsk had marked 1/3 of my 1.5TB drive as bad clusters! Luckily I noticed the drop in free space on the Server Storage tab - some users may not notice the loss and conclude that chkdsk has fixed the problem. Since I had installed the Disk Management add-in, I could see the sudden jump in usage (bad clusters) on the 1.5TB drive.

    So when users say "chkdsk fixed my file conflict problem", they may have hit the 1TB disk limit.
    Saturday, September 5, 2009 1:01 AM
  • How do I even run Chkdsk on an individual drive in my pool?

    I have 13 drives, including some greater than 1TB.  10 are in eSata enclosures (5 in each), including all the drives over 1TB.

    The internal drives are all 750GB running in IDE mode.

    Everything seems fime.
    Saturday, September 5, 2009 2:47 AM
  • Tom, you are using AHCI for your large drives so you don't have a problem :=)

    WHS mounts DATA drives under C:\fs, so on my system:

    C:\>dir c:\fs
     Volume in drive C is SYS
     Volume Serial Number is 98F6-61D0

     Directory of c:\fs

    03/09/2009  09:13 PM    <DIR>          .
    03/09/2009  09:13 PM    <DIR>          ..
    16/01/2009  11:21 PM    <JUNCTION>     N
    26/01/2009  11:13 AM    <JUNCTION>     Q
    03/09/2009  09:13 PM    <JUNCTION>     W
                   0 File(s)              0 bytes
                   5 Dir(s)  12,740,870,144 bytes free

    I can run chkdsk on drive N using: chkdsk /x/r c:\fs\N
    • Edited by JohnSoftek Saturday, September 5, 2009 3:57 AM clarification
    Saturday, September 5, 2009 3:47 AM
  • Yeah, I know that I am, I was just curious.
    Saturday, September 5, 2009 9:43 AM
  • The evidence seems to clear the firmware of suspicion. Windows XP can successfully format a 1.5 TB drive using the same model motherboard in IDE emulation mode - and Emdrei had success with a BartPE boot disk on his system where WHS failed.

    Server 2003 has had other problems with disks larger than 1TB (http://support.microsoft.com/default.aspx/kb/906668). All it takes is for a piece of code to use a signed rather than unsigned 32 bit integer to refer to the sector number.

    I agree with Ken's assessment of your situation.  Like Ken, I also have a 1.5 TB drive in my server and WHS has already stored more than 1 TB of data on it (1.12 TB to be exact).  I never did anything to "fix" it (I just added it to the pool and it worked).

    Saturday, September 5, 2009 1:33 PM
    Moderator
  • I think we can agree that the problem does not occur if you are using AHCI for your TB+ drives. The bug involves IDE and Server 2003.
    Sunday, September 6, 2009 12:18 AM
  • I had a Gigabyte and Asus MB, both with the ICH10R chipset.  Neither liked the 1.5TB drives ... the chkdsk found 390million bad sectors.

    After a frustrating afternoon, I was able to slipstream AHCI drivers into the WHS installer based on a wiki on WeGotServed.  The installer did it's thing, WHS got installed, all the drives check out with chkdsk.  Why couldn't someone tell me this 4 months ago!!!!???  Sigh.
    Tuesday, October 20, 2009 3:36 AM
  • Do you have a link to that wiki guide?
    Tuesday, October 20, 2009 3:55 PM
  • I realize this post is about a year old but the problem clearly persists.  I've posted a couple of times and continue to get the "it's your mobo/drivers/PSU" runaround.  With a 1.5TB drive as system drive it would format and work just fine on a Win7 machine but chkdsk would, in about a week, show 390GB of bad sectors in WHS.  Swapped out for a 2TB drive figuring the 1.5TB was truly going bad (it still goes strong back in my Win7 machine - reformat = no bad sectors) and was shocked to see nearly 900GB of bad sectors!  Same thing:  Moved off to Win7, reformat, worked fine.  Swapped out motherboard, same results.  Changed SATA cables, same results.  Got a beefier PSU, same results.

    The one commonality between the two machines is both ran SATA in IDE mode.  It is CLEAR from all these posts and others in the Windows forums about mysterious bad sectors on Server 2003 that it's Server 2003, even with SP2, that's the issue.  One hotfix from 4 yrs ago acknowledges the issue with drives >1TB but that it was included in SP2. 

    From all these posts it appears switching to AHCI fixes the problem.  Unfortunately the new mobo I put in is Intel ICH7 which has 4 SATA ports but no RAID/AHCI support, just IDE.  Argh!   

    My current setup:  I put a 500GB HDD in as the system drive.  No errors reported for about 2 wks now, even though I have 1 1.5TB and 2 2TB data drives in the server.  So it APPEARS to only be an issue for the system volume.  But I am nervous about all my data potentially disappearing one day on those DATA volumes, and will likely get another motherboard that offers AHCI support.  It appears from other posts like you can simply update the SATA drivers in WHS via RDP, reboot, change the bus mode in BIOS, and you should be up and running.  Any opinions on this?  Otherwise I'm going the Slipstream route indicated above - but afraid that it won't allow me to do a "Server Reinstall" based upon the change from IDE to AHCI.

    So why, with all these posts out there regarding this issue with "ghost" bad sectors, doesn't Windows acknowledge, and put a full official post on, the fact that Windows Server 2003 SP2 does not support HDD over 1 TB???

    Fearful that this message will get lost in an old post, I may post this again as a new post.  WHS builders beware:  This product, and WS2003, was built before the age of huge-capacity hard drives!!

     

    Tuesday, November 9, 2010 5:01 PM