From: Jeff Squyres; Tuesday, April 20, 2004 11:28 AM
>
> I think the point is that you may run "install_cluster" multiple times
--
> even after the initial install.
>
> So it seems reasonable that you might not want to rebuild it every
time.
> We certainly don't want to force the user to comment out code. :-)
I don't think anyone's disagreeing here...
In all my unit tests on various systems, the first invocation of
update-rpms will add any items not already in the cache, and the
remaining runs will simply validate that the cache is correct. Since my
unit tests work, the thing I can't see is how update-rpms is actually
called in the OSCAR infrastructure.
Clearly there's a disconnect between my expected usage and the actual
usage, which is why I asked Matt to provide me something to add to the
current cvs to test what he's running.
--
David N. Lombard
My comments represent my opinions, not those of Intel Corporation.
>
> On Tue, 20 Apr 2004, Mat Garrett wrote:
>
> > Jason Brechin wrote:
> > > Matt and Dave -
> > >
> > > I'm noticing that update-rpms rebuild the cache every time
> > > install_cluster is run. There is no reason for this, esp. since
you
> > > said the cache updating code was there. IT seems like update-rpms
> isn't
> > > finding or reading the existing cache. I can tell that the tool
is
> > > being run as `update-rpms --cache=u --url /tftpboot/rpm`. I let
it
> run
> > > through once, and it got past the point where it wrote the package
> > > cache. The next time I ran install_cluster, update-rpms printed
> > > "WARNING: No cache found". The cache exists in
/var/cache/update-rpms
> > > (all 14MB of it).
> > >
> > > What's going on, and when can this be fixed?
> > >
> > > Thanks,
> > > Jason
> > >
> >
> > What's runing is the update-rpms prereq setup script. This should
only
> > be run at initial install (yes/no?), so it always assumes an
> > unconditional build of the cache. When debugging, I let it build
once
> > naturally, then I go into the setup script and comment out the three
> > lines that has to do with (re)building the cache. and uncomment them
> > again when debugging is done.
> >
> > If anyone has opinions about the propriety of relying on whatever
> > update-rpms cache might alreayd be there when a user starts their
first
> > install, I'd love to hear them.
> >
> >
>
> --
> {+} Jeff Squyres
> {+} [EMAIL PROTECTED]
> {+} http://www.lam-mpi.org/
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id70&alloc_id638&op=click
_______________________________________________
Oscar-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/oscar-devel
