[ILUG] Last block on disk bad?

[Kenn Humborg]

I run a cron job weekly to scan my disks for badblocks.  On my brand-new
IBM UDMA-100 75GXP 30GB disk, the very last block is showing up as
bad.

Partition table is:

   [root at excalibur /root]# fdisk -l /dev/hda

   Disk /dev/hda: 255 heads, 63 sectors, 3737 cylinders
   Units = cylinders of 16065 * 512 bytes

      Device Boot    Start       End    Blocks   Id  System
   /dev/hda1   *         1        31    248976   83  Linux
   /dev/hda2            32        62    249007+  82  Linux swap
   /dev/hda3            63      3737  29519437+  83  Linux

Running badblocks on /dev/hda3 says that 29519436 is bad.  Running it on
/dev/hda2 shows 249004, 249005 and 249006 are bad.

I'm thinking that this is due to the partitions not ending with an
multiple-of-4 blocks (hence the + after the block count above).

Is this hypothesis correct, or do I really have a failing disk
(and it's just pure chance that the bad blocks are exactly at the
ends of partitions)?

Later,
Kenn