[RFC] The Future of the ALFS project

Kendrick alfs at linux2themax.com
Tue Feb 28 17:02:34 PST 2006


Bruce Dubbs wrote:

>Jeremy Huntwork wrote:
>  
>
>>If so, in a situation like this, wouldn't it be more efficient to have
>>one machine build the master system, so to speak and then copy it over
>>to each of the others? Perhaps even create your own RPM repository
>>(don't cringe!) that you build, according to LFS instructions and then
>>deploy to the remaining clients.
>>    
>>
>
>Yup.  I did that quite successfully on several identical systems.  Just
>tar it up and untar into a fresh partition.  A couple of config files
>need to be updated for IP address, host name, etc, but overall it is a
>very efficient process.
>  
>
Jeremy  for just a couple of the same systems yes that works as Bruce
has statted.

point 1)  goal for alfs is to allow a corperate it admin for example. 
to tell machine a to be a dns krb server  machine b to be a 2ndary dns
and mail server machine c to be a web/proxy server....  machine m to be
a base system for accounting  machine n to be a base system for the
phone support department  machine o to be setup for managers... etc.  
when it comes time for the bord to decree hay we need this piece of
software for our call center which has additional parts that help the
managers to do ...  etc   the admin can rebuilt a test lab for this
quickly and then once all kinks are worked out a small production snipit
can be built so that if they have 100 offices around the nation, its
emailed to each buldings head of it and then run for their local master
image and packages up for each subsiquent departments machine.

point 2) blfs will be able to properly blume once alfs is brought up. 
dependancy tracking was suposed to be a BIG part of this reworking.  if
it wasn't immeadiatly apart of the profile/program  it would have
support for being added shortly afterwords once the proper method was
decided on/built. 





More information about the alfs-discuss mailing list