UUIDs on drives (was Hibernate on batery low)

[Derek Broughton]

Brian Astill wrote:

> On Thursday 14 August 2008 09:47:49 Rashkae wrote:
>> I've never seen any problems
>> reported with the use of UUID that doesn't amount to old school
>> gurus not wanting to learn something different.
> 
> That's incorrect - unless you haven't been looking!

I've been looking, and completely agree with Rashkae.

> If there is a key to unravel the info implicit in /dev/hdb3
> (Primary partition on the Slave drive) from its UUID, would
> someone please publish it.

There is no "info" implicit in the UUID - it's just a nearly unique
identifier.

> 
> Also, would someone please tell me how to construct a UUID for a
> partition I have just created - short of rebooting and seeing if
> I can identify it from fstab, assuming the partition is mounted.

While it's strictly just a bit string in the superblock - and so writeable
if you know what you're doing - a UUID is generally created by mkfs
automatically.  So you _don't_ construct a UUID, you construct an fstab
entry from the known UUID.

> Why is a simple 8-digit code sufficient to uniquely identify a
> vfat partition, eg "UUID=3FF4-3BE7  /media/hda1     vfat"

It's not.  The intent is to be "nearly" unique.

> whereas a Linux partition needs something HUGE
> UUID=d757b47c-5808-4271-a247-9df9072826de /media/hdb1     ext2
> 
> I could go on, but please notice that change does not always
> involve progress.

You haven't shown us anything to disabuse us of the notion that your only
problem is not wanting to learn something different.  UUID was not
introduced to initiate progress - it was introduced to _deal_ with
progress.
-- 
derek