Re: Very large SCSI drives and partitioning.
Richard Kaszeta <kaszeta@me.umn.edu> writes:
> Peter S Galbraith writes ("Re: Very large SCSI drives and partitioning. "):
> >
> >Richard Kaszeta wrote:
> >
> >> I've gotten a
> >> ST173404LW drive (73.4 Gig Ultra2 160 drive),
> >> However, I can't partition the new drive. Cfdisk on debian 2.1
> >> refuses to talk to it at all. fdisk will talk to it, but thinks it
> >> only has 4.5 GB of capacity.
> >
> >Did you try passing the disk geometry to the kernel at boot time?
> >e.g. using lilo:
> >
> > LILO: somekernelname sdb=14100,24,424
> >
> >Just guessing here...
>
> Well, the kernel seems to be able to figure out most of it by itself:
>
> SCSI device sdb: hdwr sector= 512 bytes. Sectors= 143374738 [70007 MB] [70.0 GB]
>
> which is indeed the correct sector count and capacity according to Seagate's spec sheet.
>
> Here's excerpts from 'scsiinfo -a /dev/sdb' for it:
>
> Serial Number '3CE02MCM00007044LENE'
> Data from Rigid Disk Drive Geometry Page
> ----------------------------------------
> Number of cylinders 14100
> Number of heads 24
> Starting write precomp 0
> Starting reduced current 0
> Drive step rate 0
> Landing Zone Cylinder 0
> RPL 0
> Rotational Offset 0
> Rotational Rate 10016
>
> Data from Format Device Page
> ----------------------------
> Removable Medium 0
> Supports Hard Sectoring 1
> Supports Soft Sectoring 0
> Addresses assigned by surface 0
> Tracks per Zone 1810
> Alternate sectors per zone 0
> Alternate tracks per zone 6
> Alternate tracks per lun 0
> Sectors per track 424
> Bytes per sector 512
> Interleave 1
> Track skew factor 95
> Cylinder skew factor 85
>
> So apparently scsiinfo and Seagate appear to agree that
> cylinder=14100, and heads=24.
>
> Sectors appear to be 424, which means that total raw capacity is
> 24*14100*424=143481600, which is slightly larger than the 143374738
> reported by the kernel and seagate.
>
> However, fdisk doesn't allow me to partition it:
>
> zombie:~# fdisk /dev/sdb
> Device contains neither a valid DOS partition table, nor Sun or SGI disklabel
> Building a new DOS disklabel. Changes will remain in memory only,
> until you decide to write them. After that, of course, the previous
> content won't be recoverable.
>
>
> The number of cylinders for this disk is set to 4471.
> There is nothing wrong with that, but this is larger than 1024,
> and could in certain setups cause problems with:
> 1) software that runs at boot time (e.g., LILO)
> 2) booting and partitioning software from other OSs
> (e.g., DOS FDISK, OS/2 FDISK)
>
> Command (m for help):
>
> So it's defaulting to a cylinder count of 4471, not the 14100 it should be.
>
> So let's try expert mode:
>
> Command (m for help): x
>
> Disk /dev/sdb: 64 heads, 32 sectors, 4471 cylinders
>
> Nr AF Hd Sec Cyl Hd Sec Cyl Start Size ID
> 1 00 0 0 0 0 0 0 0 0 00
> 2 00 0 0 0 0 0 0 0 0 00
> 3 00 0 0 0 0 0 0 0 0 00
> 4 00 0 0 0 0 0 0 0 0 00
>
> Expert command (m for help): h
> Number of heads (1-256, default 64): 24
>
> Expert command (m for help): c
> Number of cylinders (1-65535, default 4471): 14100
>
> The number of cylinders for this disk is set to 14100.
> There is nothing wrong with that, but this is larger than 1024,
> and could in certain setups cause problems with:
> 1) software that runs at boot time (e.g., LILO)
> 2) booting and partitioning software from other OSs
> (e.g., DOS FDISK, OS/2 FDISK)
>
> So far, so good...
>
> Expert command (m for help): s
> Number of sectors (1-63, default 32): 464
> Value out of range.
> Number of sectors (1-63, default 32):
>
> So it's not letting me up the sector count. So the max capacity I've
> seen so far in fdisk is 63*14100*24=21319200, or about 1/7th of the
> real capacity.
>
> 'mke2fs /dev/sdb' appears to work fine, however. I still don't know
> if this won't hurt anything... Just skiddish before trusing 60 GB of
> data to it (although it is backed up via networker)
There shouldn't be anything inherently wrong with a 70G
partition. Linux limit for a partition size is somewhere in the TB
range. I bet it'll make for some LONG fsck times though!
You might read the Large Disk HOWTO to see if it gives you any
ideas. It's mostly for IDE, but there is at least a section
SCSI. Perhaps you can feed the kernel a geometry of your own choosing
via the lilo "append" option with something like:
append="sda=8924,255,63"
Don't rely on my math above either....
The 1MB ?= 1000 kB ?= 1024000 bytes thing always confuses me and I
believe you may still be able to damage a disk if you tell the
software there are more cylinders than there are.
Read the HOWTO at
http://www.linuxdoc.org/HOWTO/Large-Disk-HOWTO.html
and see if it helps.
Gary
Reply to: