The profiles are identical (as shown by the unified diff of `emerge
--info`).  Paludis is not really a solution for me as these are
production machines.  I'm really trying to understand Portage's behavior
in the scenario I outlined.

-J

--
On Sun, Aug 12, 2007 at 12:30:44PM +0200, Beso wrote:
> do you have the same profile on both the systems? if one was installed in
> 2003 and hadn't the profile updated you may have troubles getting the same
> installeddb.
> another way of seeing the db is paludis --report.
> emerge paludis and do a paludis --report and see if that package manager can
> see all your packages. i'm currently using that for the increased speed in
> resolving the deps tree, for the repositories cache and for the better use
> of overlays, or at least, the more intuitive one. of course before working
> on the system you have to install also the paludis hooks with all the hooks
> installed so that paludis doesn't do something weird (it never happened to
> my system, but you can never say), and convert the system config with the
> portage2paludis script and then you're ok.
> if you want to give a try the gentoo wiki link for making paludis work is
> here: http://gentoo-wiki.com/Paludis
> 
> 2007/8/11, Joshua Hoblitt <[EMAIL PROTECTED]>:
> >
> > The scenario:  I have two amd64 systems one newish machine and one that
> > was installed in late 2003.  I really want to sink these systems to have
> > the exact same package set.  All packages are installed from bin pkgs
> > built on another system.  They have identical:
> >
> > /etc/make.profile (same symlink)
> > /etc/make.conf
> > /etc/portage
> > /var/lib/portage/world
> >
> > The portage tree and overlay are both mounted from NFS as is the package
> > dir.
> >
> > The only different in `emerge --info` is:
> >
> > --
> > Portage 2.1.2.9 (!../usr/portage/profiles/default-linux/amd64/2007.0,
> > gcc-4.1.2, glibc-2.5-r3, 2.6.22-gentoo x86_64)
> > =================================================================
> > -System uname: 2.6.22-gentoo x86_64 Dual Core AMD Opteron(tm) Processor
> > 285
> > +System uname: 2.6.22-gentoo x86_64 AMD Opteron(tm) Processor 248
> > Gentoo Base System release 1.12.9
> > Timestamp of tree: Mon, 09 Jul 2007 18:00:01 +0000
> > distcc 2.18.3 x86_64-pc-linux-gnu (protocols 1 and 2) (default port 3632)
> > [disabled]
> > --
> >
> > These systems have both had these commands run on them:
> >
> > # dump portage cache
> > rm -rf /var/cache/edb/
> > # prune duplicate packages, dangerous but if we really need it the world
> > # emerge should pick it up
> > emerge -P
> > # check/fix packages deps
> > emerge -KnuD world
> > # remove packages still orphaned at this point
> > emerge --depclean
> > # look for blockers
> > emerge -KDep world
> > # fix any blockers...
> > emerge -KDe world
> >
> > Yet they both emerge a different number of packages (671 vs 668) and
> > report a
> > different number of installed packages they -e emerges!
> >
> > Calculating dependencies... done!
> >
> > Packages installed:   701
> > Packages in world:    2
> > Packages in system:   55
> > Unique package names: 701
> > Required packages:    701
> > Number to remove:     0
> >
> >
> > Calculating dependencies... done!
> >
> > Packages installed:   676
> > Packages in world:    2
> > Packages in system:   56
> > Unique package names: 676
> > Required packages:    676
> > Number to remove:     0
> >
> >
> > So clearly there are install packages that aren't part of "world" and
> > aren't found by depclean.  How do I get rid of them other than by
> > comparing `equery list` with `emerge -e world` and manually removing
> > the difference?  And for some reason one system does include setarch as
> > part of "system".  Seeessh.
> >
> > -J
> >
> > --
> >
> >
> 
> 
> -- 
> dott. ing. beso

Attachment: pgpGaZrt1nifw.pgp
Description: PGP signature

Reply via email to