3

I have a pair of identical SATA harddrives connected to the same controller, but for some reason, fdisk under Debian 6 reports different CHS:

user@host:~$ sudo fdisk -l /dev/sdc

Disk /dev/sdc: 1000.2 GB, 1000204886016 bytes
81 heads, 63 sectors/track, 382818 cylinders
Units = cylinders of 5103 * 512 = 2612736 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0xef1cc199

   Device Boot      Start         End      Blocks   Id  System
/dev/sdc1               1      382819   976761560   83  Linux

user@host:~$ sudo fdisk -l /dev/sdd

Disk /dev/sdd: 1000.2 GB, 1000204886016 bytes
78 heads, 63 sectors/track, 397542 cylinders
Units = cylinders of 4914 * 512 = 2515968 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0xa3d9cbe5

   Device Boot      Start         End      Blocks   Id  System
/dev/sdd1               1      397543   976761560   83  Linux

Why is that ? Shouldn't the BIOS (or whatever does that nowadays) see harddrives of the same model with the same CHS configuration ? I even bought them together, so there's a good chance they come from the same production batch. They are connected to the SATA ports 1 & 2 on the motherboard.

Is that a problem ? Should that worry me ? I tried changing cylinders and head using the fdisk expert menu, but the changes don't seem to persist...

I've done some superficial research into the subject, but mostly, I'm just happy it just works (TM)...

Has anyone ever seen that before ?

ssc
  • 1,129
  • 3
  • 16
  • 30

2 Answers2

2

Don't worry about this. The disk layout is different from what the drive reports to the computer anyway. CHS is just stated there for historic reasons.

Take the sectors/track value as an example. In the old days, this was fixed over the whole disk, but of course for many years now a disk has more sectors in the outer tracks, because they are longer. If you measure transfer speed over the whole disk, you can clearly see the points where the sector/track count increases because the transfer speed increases.

Sven
  • 97,248
  • 13
  • 177
  • 225
  • Agreed. Drive geometry as reported is largely imaginary these days, as I understand it. – Matt Simmons Aug 02 '11 at 13:30
  • thanks for the insight! I see I can ignore the issure. Still feels weird that there's different geometries for seemingly identical harddrives, especially as yet another pair of identical drives is reported with the same geometry - and moreover, the CHS values seem to change after I moved from GPT back to a DOS partition... – ssc Aug 03 '11 at 10:15
1

There can be situations where different physical drives are sourced in manufacturing. If you can (physically) spot the serial numbers, it can provide some clues as to whether the drives are actually from the same manufacturing batch.

Loosely related, Dell and their manufacturing sourcing practices, where it is very common for the same model Dells have significantly different components. Their goal is to have collectively similar functionality/capabilities for a model.

user48838
  • 7,393
  • 2
  • 17
  • 14
  • serial numbers as reported by GSmartControl: 6VPAR7SL and 6VPCY3TX. Don't think that gives any clues without some Seagate hard drive production facility database, does it ? – ssc Aug 03 '11 at 10:25
  • The advantage of a physical sighting (when possible/practical) is where manufacturing series or dates might also be stamped as well. – user48838 Aug 03 '11 at 12:31