Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
We should have this fixed today so that future emails about broken deps will be about actual broken deps, not broken configurations. Sorry for the mailbombing.
Once upon a time, Jesse Keating jkeating@redhat.com said:
Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
As a mirror admin: what does this mean for .../development/ppc{,64}? Will they go away at some point? If so, when?
On Mon, 2009-11-16 at 14:12 -0600, Chris Adams wrote:
Once upon a time, Jesse Keating jkeating@redhat.com said:
Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
As a mirror admin: what does this mean for .../development/ppc{,64}? Will they go away at some point? If so, when?
They will go away as of tonight I do believe.
On Mon, Nov 16, 2009 at 7:22 PM, Jesse Keating jkeating@redhat.com wrote:
Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
We should have this fixed today so that future emails about broken deps will be about actual broken deps, not broken configurations. Sorry for the mailbombing.
Seems its underway again today for the ppc/ppc64.
Cheers, Peter
On 11/16/2009 08:22 PM, Jesse Keating wrote:
Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
We should have this fixed today so that future emails about broken deps will be about actual broken deps, not broken configurations. Sorry for the mailbombing.
Seems as if you once more failed to fix this. The 4th flood of mail (ca. 1200 each) seems to be underway.
Ralf
On Mon, 2009-11-16 at 11:22 -0800, Jesse Keating wrote:
Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
We should have this fixed today so that future emails about broken deps will be about actual broken deps, not broken configurations. Sorry for the mailbombing.
*sigh*
While we were successful in building a new mash package that would avoid making ppc repos, we forgot to update one of the rawhide creation configs so that it used dist-f13 content as opposed to dist-f12. So the rawhide creation process has been using dist-f12 content all this time to build up the chroot, which would then compose dist-f13 content. This means that the dist-f12 version of mash was used, not the dist-f13 version we built to disable ppc.
I've corrected that. Third try to kill the ppc deps should be the charm.
On Tuesday 17 November 2009 06:50:58 am Jesse Keating wrote:
On Mon, 2009-11-16 at 11:22 -0800, Jesse Keating wrote:
Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
We should have this fixed today so that future emails about broken deps will be about actual broken deps, not broken configurations. Sorry for the mailbombing.
*sigh*
While we were successful in building a new mash package that would avoid making ppc repos, we forgot to update one of the rawhide creation configs so that it used dist-f13 content as opposed to dist-f12. So the rawhide creation process has been using dist-f12 content all this time to build up the chroot, which would then compose dist-f13 content. This means that the dist-f12 version of mash was used, not the dist-f13 version we built to disable ppc.
I've corrected that. Third try to kill the ppc deps should be the charm.
Could we just not send emails tomorrow, double check that it produces the correct result, and re-enable them for the next day? In case there's something else we-shouldn't-have-missed-but-we-did?
On Nov 17, 2009, at 14:47, Conrad Meyer cemeyer@u.washington.edu wrote:
On Tuesday 17 November 2009 06:50:58 am Jesse Keating wrote:
On Mon, 2009-11-16 at 11:22 -0800, Jesse Keating wrote:
Many of you received emails over the weekend and this morning regarding broken deps in rawhide. If these emails mentioned that the deps were broken on ppc or ppc64 they can be ignored. We are no longer producing ppc/ppc64 as a primary arch, however we forgot to tag the config change that enacted this on our compose tools. We were attempting to compose ppc(64) trees with only noarch packages, and well things didn't work so hot.
We should have this fixed today so that future emails about broken deps will be about actual broken deps, not broken configurations. Sorry for the mailbombing.
*sigh*
While we were successful in building a new mash package that would avoid making ppc repos, we forgot to update one of the rawhide creation configs so that it used dist-f13 content as opposed to dist-f12. So the rawhide creation process has been using dist-f12 content all this time to build up the chroot, which would then compose dist-f13 content. This means that the dist-f12 version of mash was used, not the dist-f13 version we built to disable ppc.
I've corrected that. Third try to kill the ppc deps should be the charm.
Could we just not send emails tomorrow, double check that it produces the correct result, and re-enable them for the next day? In case there's something else we-shouldn't-have-missed-but-we-did?
That's one of the changes I made today.
-- Jes
-
Chris Adams -
Conrad Meyer -
Jesse Keating -
Jesse Keating -
Peter Robinson -
Ralf Corsepius