[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Do "slave" machines in a Condor pool need 80MB of sw?

> >Hmm, thinking about this a bit further.  We could split it 
> into multiple
> >packages like "condor-core" "condor-execute" 
> "condor-execute_standard"
> >"condor-execute_vanilla" "condor-submit" "condor-submit_standard"
> >"condor-submit_vanilla" "condor-central_manager", 
> "condor-compile"...  As you
> >can see, this could rapidly become quite a maze of packages.

This is one of the quandaries about software release - how to package.

I suspect that few would argue that there is scope for some repackaging
of the Condor release, but how? Who would benefit? How can it be done 
without causing confusion - remember that Condor already has the following
download options:
* Stable vs Development (and when latest Development becomes Stable, lots more
  "goodies" will become available by default and be included in the .rpm or equiv)
* OS - there are many OSes and OS flavours and versions supported
* Static vs Dynamic
* Some also have .rpm as well as .tar.gz

This is already a bit of a maze - adding other options would make it harder.

Of course there are a few who would suggest that releasing the sources would help!

If any subdivision was considered, it would have to be aimed at those who
would most benefit and that would mean (to me) a simple(r) client side download
for use in established systems, probably a submit+execute download which could
be simply tuned to be one or the other.

All-in-all, I don't think there would be much to be gained, but I suspect it is
worth (re)considering for the next development release (soon I hope chaps - hint ;-) )