#6016: Use fedpkg-minimal in Fedora buildroots
----------------------+------------------------
Reporter: pbabinca | Owner: rel-eng@…
Type: task | Status: new
Milestone: | Component: koji
Keywords: meeting | Blocked By:
Blocking: |
----------------------+------------------------
Speedup build root readiness for the builds by reducing dependencies
pulled into build root by fedpkg python package.
Use [http://koji.fedoraproject.org/koji/buildinfo?buildID=485120|fedora-
minimal] shell script which fetches sources only. Similar script is
currently used in EPEL 7 buildroot. Start with rawhide later to other
supported dists.
Original idea came in
[https://lists.fedoraproject.org/pipermail/devel/2014-July/201154.html|discu…
about slow buildSRPMFromSCM].
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/6016>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
#6008: Fedora 21 Alpha .composeinfo should have family=Fedora
----------------------+------------------------
Reporter: dcallagh | Owner: rel-eng@…
Type: defect | Status: new
Milestone: | Component: other
Keywords: | Blocked By:
Blocking: |
----------------------+------------------------
The .composeinfo file for Fedora 21 Alpha Server currently has this:
{{{
[product]
family = Fedora-Server
name = Fedora-Server-21_Alpha
version = 21_Alpha
variants = Server
}}}
For future F21 composes please update it so that the family is just
"Fedora", with "Server" as the variant like it is currently.
This is consistent with how RHEL has been doing its variant products for
many years. It also just makes more sense -- all the new Fedora variants
are in the same family, that's the whole point of that field. And it will
make our lives easier in Beaker, where we use the family to figure out
which harness builds to use and which kickstart syntax to use.
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/6008>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
#5775: switch to partitioned image and "hd00" pvgrub akis for EC2; remove editing
of menu.lst from image processing
----------------------------+------------------------
Reporter: mattdm | Owner: rel-eng@…
Type: task | Status: new
Milestone: Fedora 20 Beta | Component: koji
Keywords: | Blocked By:
Blocking: |
----------------------------+------------------------
The "hd00 aki" is like the pvgrub kernel image we are using now, but meant
for a partitioned device. If we switched to doing this, we could avoid
editing the menu.lst file, and therefore would have a more-identical
image.
.
See http://www.dowdandassociates.com/content/howto-amazon-ec2-kernel-
images-pv-grub/ (or
http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/UserProvidedkernels.html,
which seems to have vanished today) for more.
You can find the images in each region with:
euca-describe-images -o amazon --filter "manifest-location=*pv-grub-
hd00*"
and of course take the newest version.
At this point, I doubt we want to change this by alpha, but maybe we could
for beta?
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/5775>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
#4985: Extend test image creation SOP to require notification of delays
----------------------+-----------------------------------------------------
Reporter: adamwill | Owner: rel-eng(a)lists.fedoraproject.org
Type: task | Status: new
Milestone: | Component: koji
Keywords: |
----------------------+-----------------------------------------------------
Working through the F16 QA retrospective here:
https://fedoraproject.org/wiki/Fedora_16_QA_Retrospective
One of the notes is:
"jlaska - Alpha TC1 - Missed the Alpha-TC1 milestone by 1 week, little to
no communication regarding status. With only two weeks Alpha ISO test
time, losing 50% of time almost certainly results in a slip. See rel-eng
ticket#4844 . None of the Alpha rel-eng release tickets have been filed at
this time.
adamw - perhaps releng image compose SOP should require updating of ticket
with progress info, especially when image compose / delivery is delayed?"
So: this ticket proposes that the releng 'Composing test images' SOP -
https://fedoraproject.org/wiki/Composing_test_images - should be extended
to require whoever is creating the images to post a comment on the image
request ticket if the compose is delayed, explaining the reasons for the
delay (so that any other interested party can attempt to contribute to
resolving them) and giving a best estimate of an ETA if possible (so other
groups can adjust their scheduling as best they can).
Thanks!
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/4985>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
#5836: Update gitolite to gitolite3
------------------+------------------------
Reporter: till | Owner: rel-eng@…
Type: task | Status: new
Milestone: | Component: git
Keywords: | Blocked By:
Blocking: |
------------------+------------------------
The package gitolite3 in EPEL should be used to update gitolite for dist-
git.
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/5836>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
#5924: CPUID flags on buildvm-20.phx2.fedoraproject.org
-----------------------------+------------------------
Reporter: jakub | Owner: rel-eng@…
Type: task | Status: new
Milestone: Fedora 21 Alpha | Component: koji
Keywords: | Blocked By:
Blocking: |
-----------------------------+------------------------
In the gcc-4.9.0-12.fc21 build logs I saw a weird error:
/builddir/build/BUILD/gcc-4.9.0-20140619/obj-x86_64-redhat-linux/gcc/xgcc
-B/builddir/build/BUILD/gcc-4.9.0-20140619/obj-x86_64-redhat-linux/gcc/
/builddir/build/BUILD/gcc-4.9.0-20140619/gcc/testsuite/gcc.target/i386/pr57275.c
-fno-diagnostics-show-caret -fdiagnostics-color=never -O3 -march=native
-lm -o ./pr57275.exe
/builddir/build/BUILD/gcc-4.9.0-20140619/gcc/testsuite/gcc.target/i386/pr57275.c:1:0:
error: CPU you selected does not support x86-64 instruction set
and I'd like to debug that, but don't know of a way to force e.g. a
scratch rpm build to a particular build host. Would it be possible for
some release engineer to run
gcc -v -march=native -S -xc /dev/null -o /tmp/foo.s
cat /proc/cpuinfo
in the F21 mock chroot on that box? Perhaps we are missing something in
the -m{arch,tune}=native logic,
a CPU which clearly does support x86_64 64-bit ISA should not be
determined as one that does not.
Thanks.
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/5924>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
#5959: Enable keep-alive connections for koji (primary and secondaries)
-----------------------------+------------------------
Reporter: till | Owner: rel-eng@…
Type: task | Status: new
Milestone: Fedora 20 Final | Component: koji
Keywords: meeting | Blocked By:
Blocking: |
-----------------------------+------------------------
Keeping connections alive will speed up koji acces when multiCall() is not
used without a significant speed penalty. Therefore it is a good idea to
support this.
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/5959>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
We should move to something more secure than md5 for the uploaded sources.
This patch series implements the client-side part of this change.
We might want to drop the md5 fallback once we have migrated completely, that
is when:
1. all archives on the lookaside have been moved to a stronger hash
2. the "sources" file in all git repos has been updated to the same hash
https://fedorahosted.org/rel-eng/ticket/5846
#5846: move away from md5 for look-aside cache
------------------+------------------------
Reporter: till | Owner: rel-eng@…
Type: task | Status: new
Milestone: | Component: other
Keywords: | Blocked By:
Blocking: |
------------------+------------------------
The lookaside cache uses md5, but something more secure like sha-256 or
sha-512 should be used instead. Maybe it should even be made to allow easy
changes in the future.
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/5846>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project
#5992: Create new fedora-ca certificates without SHA1
-----------------------------+------------------------
Reporter: till | Owner: rel-eng@…
Type: task | Status: new
Milestone: Fedora 20 Final | Component: koji
Keywords: | Blocked By:
Blocking: |
-----------------------------+------------------------
Currently .fedora-server-ca.cert and .fedora-upload-ca.cert use SHA1 for
their self-signatures. Since SHA1 support is phased out, it might be a
good idea to re-create the signatures.
--
Ticket URL: <https://fedorahosted.org/rel-eng/ticket/5992>
Fedora Release Engineering <http://fedorahosted.org/rel-eng>
Release Engineering for the Fedora Project