Linux: Boot a 2TB+ partition or Larger Array Using Grub

in Categories CentOS, data center, fedora linux, File system, Gentoo Linux, Hardware, High performance computing, Howto, kernel, Linux, Linux Scalability, Linux Virtualization, RedHat/Fedora Linux, Storage, Tips, vmware, xen last updated December 4, 2008

I’ve already written about creating a partition size larger than 2TB under Linux using GNU parted command with GPT. In this tutorial, I will provide instructions for booting to a flat 2TB or larger RAID array under Linux using the GRUB boot loader.

How to Boot a 2TB or Larger Array in Linux

Run fdisk -l to see disk drive geometry and note down heads, sectors and cylinders. If your device name is /dev/sda, enter:
# fdisk -l /dev/sda
Note down drive geometry from the output including heads, sectors, cylinders etc. Type the grub, enter:
# grub
At grub prompt type the following commands:
grub> geometry (hd0) <#cylinders><#heads><#sectors>
grub> root (hd0,0)
grub> setup (hd0)
Replace <#cylinders><#heads><#sectors> with actual values. /dev/sda is hd0 i.e. first scsi hard disk. command geometry can be used to print the information for the drive drive. In the grub shell, you can set the geometry of the drive arbitrarily. The number of cylinders, the number of heads, the number of sectors and the number of total sectors are set to CYLINDER, HEAD, SECTOR and TOTAL_SECTOR, respectively. If you omit TOTAL_SECTOR, then it will be calculated based on the C/H/S values automatically.

GRUB should update MBR and other information in order to boot your system. Before, you issue reboot command make sure grub.conf (menu.lst) has correct syntax.

Posted by: Vivek Gite

The author is the creator of nixCraft and a seasoned sysadmin, DevOps engineer, and a trainer for the Linux operating system/Unix shell scripting. Get the latest tutorials on SysAdmin, Linux/Unix and open source topics via RSS/XML feed or weekly email newsletter.

Share this on (or read 3 comments/add one below):

3 comment

  1. mate, this stuff totally does not work. the geometry command gives a negative number of total cylinders and error 22 during further installation.

  2. Does this even make sense? fdisk does not work with GPT drives, and indeed, a fair number of machine BIOS’es probably don’t work either, so booting might never be possible?

    For folks looking for an answer who happen to use 3ware RAID cards, these can be configured when you create your array to contain a small virtual drive at the start, then your massive terabytes drive afterwards – so you can use MBR on the smaller drive to boot from, and hopefully GPT on the huge one after. eg – here I use the ThreeWareCommandLineUtility (tw_cli) to create a 12gig disk (using v0=12) to put my boot/swap partitions on, and a 14TB one for the rest (/):-

    root@sysresccd /x/3w/packages/cli/linux/x86/tw_cli
//sysresccd> show

Ctl   Model        (V)Ports  Drives   Units   NotOpt  RRate   VRate  BBU
------------------------------------------------------------------------
c2    9550SXU-12ML 12        8        1       0       1       1      -
//sysresccd> /c2 show

Unit  UnitType  Status         %RCmpl  %V/I/M  Stripe  Size(GB)  Cache  AVrfy
------------------------------------------------------------------------------
u0    RAID-5    OK             -       -       256K    13038.4   ON     OFF

Port   Status           Unit   Size        Blocks        Serial
---------------------------------------------------------------
p0     OK               u0     1.82 TB     3907029168    9WM04TVH
p1     OK               u0     1.82 TB     3907029168    9WM02YH3
p2     OK               u0     1.82 TB     3907029168    9WM02WQ2
p3     OK               u0     1.82 TB     3907029168    9WM04K1M
p4     OK               u0     1.82 TB     3907029168    9WM04EKH
p5     OK               u0     1.82 TB     3907029168    9WM067LK
p6     OK               u0     1.82 TB     3907029168    9WM04GMQ
p7     OK               u0     1.82 TB     3907029168    9WM03T26
p8     NOT-PRESENT      -      -           -             -
p9     NOT-PRESENT      -      -           -             -
p10    NOT-PRESENT      -      -           -             -
p11    NOT-PRESENT      -      -           -             -

//sysresccd> /c2/u0 del
Deleting /c2/u0 will cause the data on the unit to be permanently lost.
Do you want to continue ? Y|N [N]: Y
Deleting unit c2/u0 ...Done.

//sysresccd> /c2 add type=raid5 disk=0-7 stripe=256K name=cnd14tb ignoreECC autoverify v0=12 storsave=perform
Creating new unit on controller /c2 ... Done. The new unit is /c2/u0.
Naming unit /c2/u0 to [cnd14tb] ... Done.
Setting AutoVerify=ON for the new unit ... Done.
Setting IgnoreECC=ON for the new unit ... Done.
Setting Storsave policy to [perform] for the new unit ... Done.
Warning: New unit /c2/u0 storsave policy is set to Performance which may
cause data loss in the event of power failure.
Setting default Command Queuing policy for unit /c2/u0 to [on] ... Done.
Setting write cache = ON for the new unit ... Done.
Warning: You do not have a battery backup unit for /c2/u0 and the enabled
write cache (default) may cause data loss in the event of power failure.

//sysresccd> /c2/u0 show all
/c2/u0 status = INIT-PAUSED
/c2/u0 is not rebuilding, its current state is INIT-PAUSED
/c2/u0 is not verifying, its current state is INIT-PAUSED
/c2/u0 is initializing with Percent Completion = 0%
/c2/u0 Write Cache = on
/c2/u0 volume(s) = 2
/c2/u0 name = cnd14tb
/c2/u0 serial number = 9WM04TVH000479002578
/c2/u0 Ignore ECC policy = on
/c2/u0 Auto Verify Policy = on
/c2/u0 Storsave Policy = performance
/c2/u0 Command Queuing Policy = on

Unit     UnitType  Status         %RCmpl  %V/I/M  Port  Stripe  Size(GB)
------------------------------------------------------------------------
u0       RAID-5    INIT-PAUSED    -       0%      -     256K    13038.4
u0-0     DISK      OK             -       -       p0    -       1862.63
u0-1     DISK      OK             -       -       p1    -       1862.63
u0-2     DISK      OK             -       -       p2    -       1862.63
u0-3     DISK      OK             -       -       p3    -       1862.63
u0-4     DISK      OK             -       -       p4    -       1862.63
u0-5     DISK      OK             -       -       p5    -       1862.63
u0-6     DISK      OK             -       -       p6    -       1862.63
u0-7     DISK      OK             -       -       p7    -       1862.63
u0/v0    Volume    -              -       -       -     -       12
u0/v1    Volume    -              -       -       -     -       13026.4

//sysresccd>
//sysresccd> quit
root@sysresccd /x/3w/packages/cli/linux/x86 % fdisk -l

Disk /dev/sda: 12.9 GB, 12884901888 bytes
255 heads, 63 sectors/track, 1566 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00000000

Disk /dev/sda doesn't contain a valid partition table

Disk /dev/sdb: 13987.0 GB, 13987037052928 bytes
255 heads, 63 sectors/track, 1700493 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00000000

Disk /dev/sdb doesn't contain a valid partition table
root@sysresccd /x/3w/packages/cli/linux/x86 %

  3. Unfortunately I don’t see how this helps, I’ve tried it. What I ended up doing is making my own grub boot disc. I used the latest Debian unstable grub .97 diff/patch on top of the original .97 GNU Grub. You really need a patched grub to write the MBR on a 2TB+ hard drive. Remember that .97 GRUB does NOT work with GPT EFI partitions either (and I personally don’t like them anyway). I believe Grub 2 uses it, but most distros are still using their own form of patched .97.

    Personally I blame GRUB, they should be making new releases so the distros can all use a standardized GRUB, instead of patching it for peculiar things with EXT3 etc…

    Have a question? Post it on our forum!

Tagged as: Tags , , , , , , , ,