On Fri 06 Feb 2009 at 10:25AM, Jason Zhao wrote:
> It does not seem like a reasonable idea for customers/users to
> do the workaround after setting a new default IPS repo, if they
> try to create a new ipkg zone.
>
> I have filed a bug: 6428
> <http://defect.opensolaris.org/bz/show_bug.cgi?id=6428> to track it.
>
> Please tell me if I misunderstand it.
The problem stems from your changing authorities in this way. I'll
think about how best to address it.
Usually when I switch from one repo to another, I keep the authority
name-- opensolaris.org-- and just change the URL. That helps keep
everything more consistent but it is not perfect.
-dp
>
> Thanks
> Jason
> Jason Zhao wrote:
> > Hi, Experts,
> >
> > On OpenSolaris, after set the default IPS repo to my local one(localips),
> > and the previous IPS repo is "colorsword" in this case. When I
> > need to create an ipkg zone for test. It is failed because it
> > still try to get packages from previous "colorsword" IPS repo.
> >
> > As following.
> > # pkg set-authority -P -O http://localips:10000 localips
> > # zonecfg -z z1
> > z1: No such zone configured
> > Use 'create' to begin configuring a new zone.
> > zonecfg:z1> create
> > zonecfg:z1> set zonepath=/zones/z1
> > zonecfg:z1> set brand=ipkg
> > zonecfg:z1> exit
> >
> > # zoneadm -z z1 install
> >
> > The "/usr/lib/brand/pkgcreatezone" shows the following error message,
> > ......
> > + echo localips=http://localips:10000/
> > Authority: Using http://localips:10000/.
> > Image: Preparing at /zones/z1/root ...+ printf ' Authority: Using %s.'
> > http://localips:10000/
> > + printf '\n Image: Preparing at %s ...' /zones/z1/root
> > + pkg image-create -z -F -a localips=http://localips:10000/ /zones/z1/root
> > done.
> > + printf ' done.\n'
> > + nawk '{print $1}'
> > + pkg list -Hv entire
> > Cache: Using /var/pkg/download.
> > Installing: (output follows)
> > + entire_fmri=pkg://colorsword/[email protected],5.11-0.105:20090108T205827Z
> > + PKG_IMAGE=/zones/z1/root
> > + export PKG_IMAGE
> > + [ -d /var/pkg/download ]
> > + PKG_CACHEDIR=/var/pkg/download
> > + export PKG_CACHEDIR
> > + printf ' Cache: Using %s.\n' /var/pkg/download
> > + printf ' Installing: (output follows)\n'
> > + pkg install -q pkg://colorsword/[email protected],5.11-0.105:20090108T205827Z
> >
> > pkg: pkg: The following pattern(s) did not match any packages in the current
> > catalog. Try relaxing the pattern, refreshing and/or examining the catalogs
> > pkg://colorsword/[email protected],5.11-0.105:20090108T205827Z
> > + fail_incomplete 'failed to install package'
> > + print 'failed to install package'
> > failed to install package
> > + exit 254
> > ......
> >
> >
> > In the past, there was a workaround, manually set index file and make it
> > consistent with "localips" repo.
> >
> > The workaround step is as following
> > - manually update the index dir
> > "/var/pkg/pkg/entire/0.5.11%2C5.11-0.xxx%xxxxx"
> > and "installed" file inside, then make them consistent with the
> > "localips" IPS repo.
> > # cat installed
> > VERSION_1
> > _PRE_localips
> >
> > Or, update the index by "pkg install pkg://localips/entire", but it will
> > results in adding all the packages which "entire" depends.
> >
> > Then, the whole zone installing will be successful by "zoneadm -z z1
> > install".
> >
> >
> > My question is:
> > - Is it a IPS brand zone bug?
> >
> > If not
> > - How can I simply update the "entire" index entry much easier?
> >
> >
> >
> > Great Thanks
> > Jason
> > _______________________________________________
> > pkg-discuss mailing list
> > [email protected]
> > http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
> >
>
> _______________________________________________
> pkg-discuss mailing list
> [email protected]
> http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
--
Daniel Price, Solaris Kernel Engineering http://blogs.sun.com/dp
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
