Ubuntu 12.04 won't install due to odd partitioning

[D. Hugh Redelmeier]

| From: Lennart Sorensen <lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org>

| On Wed, May 16, 2012 at 01:18:59AM -0400, D. Hugh Redelmeier wrote:

| >     Disk /dev/sda: 120.0 GB, 120034123776 bytes
| >     255 heads, 63 sectors/track, 14593 cylinders
| >     Units = cylinders of 16065 * 512 = 8225280 bytes
| >     Sector size (logical/physical): 512 bytes / 512 bytes
| >     I/O size (minimum/optimal): 512 bytes / 512 bytes
| >     Disk identifier: 0x17a417a4
| >      Device Boot      Start         End      Blocks   Id  System
| >     /dev/sda1               1        1020     8193116   27  Unknown
| >     Partition 1 does not end on cylinder boundary.
| >     /dev/sda2   *        1021        6368    42957804+   7  HPFS/NTFS
| >     /dev/sda3            6369        7388     8193148   83  Linux: Ubuntu8.04
| >     /dev/sda4            7389       14594    57882195    f  W95 Ext'd (LBA)
| >     /dev/sda5            7389        8153     6144820   82  Linux swap / Solaris
| >     /dev/sda6            8154       13573    43536116   83  Linux: /home
| >     /dev/sda7           13574       14594     8195072   83  Linux: Ubuntu10.04
| > 
| > It all looks good: all the cylinder numbers make sense.
| 
| You think ending on 14594 makes sense when you have 14593 cylinders total?

True.

| Your sda4 and hence sda7 are invalid.  Any attempt to write to the last
| few mebabytes of sda7 would have failed.  You probably just never did.

I don't think so.  The last (1024-byte) block in sda7 is 8195072:

    root at redact:~# dd if=/dev/sda7 skip=8195071 of=/dev/null bs=1024
    1+0 records in
    1+0 records out
    1024 bytes (1.0 kB) copied, 9.192e-05 s, 11.1 MB/s
    root at redact:~# 

dumpe2fs says that there are 2048768 blocks in the filesystem.  I
presume that those are 4k blocks so that is the same as 8195072 1k
blocks.  Thus (not surprisingly) the filesystem is occupying the whole
partition.

This fantasy of cylinders is past being funny.  For example, using
255*63 guarantees that the second partition will not be aligned on a
1024-byte boundary.  Very inconvenient for new disk drives that only
fake having 512-byte sectors.

The GCD of the last two partitions' sector counts 4 so they don't seem
to have been allocated on the basis of some identical notional
cylinder size.  But I don't know if that is meaningful -- perhaps some
overhead sectors are not counted.
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists