Hi John,
The files in the base release (../i386/SL/ and ../x86_64/SL) are not 
hard linked with anything.  This is because of the flow of the release 
and the way it's pushed out.

The files that *are* hard linked are all the security errrata. 
(..i386/updates/security and ../x86_64/updates/security)

Could I hard link the rpm's in the base release?
Yes, but sometimes I'm a bit paranoid, and that's one of the things I'm 
a bit paranoid about.  What if it changes something, even just a 
timestamp, and that breaks something else.  The install is this symphony 
of interrelated dependancies and sometimes very minor changes can lead 
to unforseen results.

Troy

John Rowe wrote:
> I have machines running two or three of the most recent versions of SL
> and would like to mirror the distros making efficient use of hard links
> (there seem to be very few RPMs that are in just one version).
> 
> I'm trying:
> 
> rsync -avzH --delete  \
>     --exclude=sites/Fermi --exclude=archive/debuginfo \
>     --exclude=archive/obsolete \
>     --include='/5[2-9x]/x86_64' \
>     --exclude='/*/*' \
>    rsync://rsync.scientificlinux.org/scientific/ scientific/
> 
> but I'm finding that the same file in two distributions is coming over
> as two different files, rather than two hard links to the same file. For
> example:
> 
>> ls -il 5[45]/x86_64/SL/evolution-2.12.3-19.el5.x86_64.rpm
> 71805541 -rw-r--r-- 1 root root 12388735 Sep  4  2009
> 54/x86_64/SL/evolution-2.12.3-19.el5.x86_64.rpm
> 77856909 -rw-r--r-- 1 root root 12388735 Sep  4  2009
> 55/x86_64/SL/evolution-2.12.3-19.el5.x86_64.rpm
> 
> This leads me to believe that either I'm doing something wrong or that
> they genuinely are different, but identical, files on the SL server.
> 
> Any suggestions?
> 
> Thanks
> 
> John


-- 
__________________________________________________
Troy Dawson  [log in to unmask]  (630)840-6468
Fermilab  ComputingDivision/LSCS/CSI/USS Group
__________________________________________________