A curse on LABEL=

[Richard Lynch]

Timothy Murphy wrote:

>I honestly expected to be shown simple, likely to occur, scenarios
>where LABELs are helpful,
>but all the examples that have been suggested
>are ones I cannot imagine a non-guru like me ever encountering.
>  
>
Lets say I have a single drive dual boot system.  The 1st two primary 
partitions are Windows.  The rest of the partitions on the drive are for 
Linux.   The 1st primary (C:) is full.  The second (D:) is barely used.  
I buy a new drive and, using a partitioning tool (e.g. PartitionMagic),  
I create a new primary on the second drive and copy all my stuff from 
the 2nd partition on the 1st drive to it.  I delete the 2nd partition on 
the 1st drive and expand the 1st partition into the new free space.  
Now, for Windows, my C: and D: drives are intact and C: is no longer 
full.  D: is on my new drive.  However, for Linux, /dev/hda3 has become 
/dev/hda2.  With the LABEL= approach the system will still boot.  With 
hardcoded devices it wont. 

It doesn't take much of an imagination to come up with other scenarios 
that would be even more disruptive to the partition numbers.  For 
example, re-arranging primary and logical partitions because one needs 
to adjust the space allocations or the system layout.  Basically, the 
LABEL= paradigm gives you a level of indirection for devices -- a 
symbolic name instead of a hardcoded device address.  This is generally 
considered a good thing and a more flexible approach to system design 
and administration.

-- 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: rich.vcf
Type: text/x-vcard
Size: 259 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/fedora-list/attachments/20040411/6458fdc9/attachment-0001.vcf>