2

I've picked up an HP SimpleSave sd500a backup drive. This is a 2.5", 500GB drive. It has a mysterious CD-like partition, but otherwise seems to contain a WD Scorpio Blue disk. It seems that the CD-like partition is implemented in the enclosure's firmware, but I've no way to be certain of this.

I'm repartitioning the drive for the first time. When attempting to open the drive using cfdisk /dev/sdb, it exits with status 4 after outputting this error message:

FATAL ERROR: Bad primary partition 0: Partition ends in the final partial cylinder

sfdisk -l is able to output info on the drive without errors:

Disk /dev/sdb: 60715 cylinders, 255 heads, 63 sectors/track
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls    #blocks   Id  System
/dev/sdb1          0+  60715-  60716- 487699456    7  HPFS/NTFS
/dev/sdb2          0       -       0          0    0  Empty
/dev/sdb3          0       -       0          0    0  Empty
/dev/sdb4          0       -       0          0    0  Empty

Is the error from cfdisk any reason to question the stability of the drive or the compatibility of its firmware?

intuited
  • 3,488
  • 3
  • 25
  • 36

2 Answers2

1

cfdisk reads the partition table of the device at startup, it will exit if the geometry of a partition is wrong. You can force cfdisk to not read the existing partition table by adding -z:

cfdisk -z /dev/sdb

This is a cfdisk specific behavior, fdisk will show a similar error but won't exit. The stability of the drive is not affected, it's just a partition issue.

Alternatively use a partition tool like fdisk, parted or gparted.

I've just checked my own partition and the first one (/boot) also reported this error. I never had any problems with it.

wag
  • 35,104
  • 11
  • 66
  • 51
  • I'm more concerned with the issue being a symptom of a bigger problem than with avoiding it. Why would `cfdisk` reach the conclusion that a partition table as produced by a manufacturer was invalid? – intuited Jan 20 '11 at 19:42
  • Not every operating system, filesystem or partition tool do handle geometry information like `(c)fdisk` does. As your `sfdisk -l` implies, the partition is formated NTFS, which (could) lead to the conclusion, that this drive was mainly produced and partitioned for the Windows world (so why care about the cylinder interpretation of a Linux partition tool?). – wag Jan 20 '11 at 19:59
  • hmmrmm still kind of confused.. doesn't linux basically ignore C/H/S geometry? I would have thought that linux tools would just go with the windows flow for this sort of thing. Well, I guess I need to actually read those other paragraphs in the wikipedia entry. Mostly I'm just being ultra-anal because the drive firmware is sort of abnormal in other ways. But man was it cheap. In any case, I've been using it for a while now and haven't had any problems. – intuited Feb 01 '11 at 03:45
0

Supposing there isn't anything of value in there, remove the partitions and create a new one (either ext3 or ext4), and run e2fsck -c on it to check if it has bad blocks. If that passes, and you happen to be paranoid, run e2fsck -cc to do a more thorough (and much looonger) test.

tshepang
  • 64,472
  • 86
  • 223
  • 290