debhelper has changed the way it performs path validations and
building the deb package fails when it tries to compress the files.
Signed-off-by: David Calavera <david.calavera@gmail.com>
Instead of being split between three files, let's let `hack/make/.detect-daemon-osarch` be our single source of truth for multiarch detection/vars. Not only does it make it slightly easier to make sure we change everything properly when these bits have to change, but it also makes it so that all bits of `hack/make.sh` (especially `hack/make/.ensure-frozen-images`) work properly outside the context of the `Makefile` on all platforms.
Signed-off-by: Andrew "Tianon" Page <admwiggin@gmail.com>
The Releases file(s) and other bits for EOL-ed distros such as Ubuntu
Vivid should remain untouched when we are releasing debs.
However, few files in https://apt.dockerproject.org/repo/dists/ubuntu-vivid/
were being updated for the docker 1.10 release including the Release files.
This is due to apt-ftparchive generating index files for vivid as well,
due to the stale apt-ftparchive.conf
This change always creates config using suites in contrib/reprepro/suites.sh.
Signed-off-by: Anusha Ragunathan <anusha@docker.com>
Progress toward being able to run integration-cli campaign using a
client hitting a remote host.
Most of these fixes imply flagging tests that assume they are running on
the same host than the Daemon. Also fixes the `contrib/httpserver` image
that couldn't run because of a dynamically linked Go binary inside the
busybox image.
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
This PR adds support for using the DOCKER_BUILD_PKGS env var to
better help defining what packege to build. It also adds support
for the integration-daemon so we can run it as a bundle.
Signed-off-by: Paul Liljenberg <liljenberg.paul@gmail.com>
add directory test
Adds missing directory test. This helps verifying that
DOCKER_BUILD_PKGS is a directory and exists before continueing.
Signed-off-by: Paul Liljenberg <liljenberg.paul@gmail.com>
fix indent
Signed-off-by: Paul Liljenberg <liljenberg.paul@gmail.com>
This will only trim off the last '-' separated token to be checked for
RC status. This allows including a 'cs#' token in the version string.
Signed-off-by: Mike Dougherty <mike.dougherty@docker.com>
dockerinit has been around for a very long time. It was originally used
as a way for us to do configuration for LXC containers once the
container had started. LXC is no longer supported, and /.dockerinit has
been dead code for quite a while. This removes all code and references
in code to dockerinit.
Signed-off-by: Aleksa Sarai <asarai@suse.com>
For the CS Engine we need to have an install script like OSS does, but
the locations are all different, as is the GPG key used. This is
accomplished here by slightly altering the script itself and adding a
simple 'sed' based bundle for make.sh.
This install script is used in to change the URLs instead of sed in
release.sh.
Signed-off-by: Mike Dougherty <mike.dougherty@docker.com>
This results in a significant time savings during repeated builds (since we don't have to re-download gcc for every test run).
Signed-off-by: Andrew "Tianon" Page <admwiggin@gmail.com>
.ensure-syscall-test need to build an image, we have
to add http_proxy and https_proxy to make it work.
Inseted of adding http_proxy and https_proxy env to the Dockerfile,
adding a DOCKER_BUILD_ARGS env to pass the build-arg is much more
elegant.
Signed-off-by: Lei Jitang <leijitang@huawei.com>
file to require the Unbreakable Enterprise Kernel Release 4 on both Oracle Linux 6
and Oracle Linux 7.
The UEK R4 provides the required kernel functionality for VxLAN support
required by Docker 1.9 and user namespace support required for 1.10+.
The build of Docker on Oracle Linux 6 requires some manipulation of the build
environment so that the CGO compiler uses the UEK R4 headers instead of the old
default kernel headers.
Signed-off-by: Avi Miller <avi.miller@oracle.com>
Nightly and unofficial builds of Docker bear the suffix `-dirty` in the
version string. Change this suffix to `-unsupported` to make it explicit
that no support will be provided on such versions, and that it is for
example unnecessary to file an issue for it.
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
Ubuntu 15.04 (Vivid) will be EOL'd in January,
so we should remove it from our builds in the
Docker 1.10 release.
For information about the EOL data, see:
https://wiki.ubuntu.com/Releases
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Fedora 21 is EOL'd as of December 1st, 2015.
Announcement:
https://lists.fedoraproject.org/pipermail/announce/2015-November/003296.html
Fedora 21 will reach end of life on 2015-12-01, and no further updates
will be pushed out after that time. Additionally, with the recent
release of Fedora 23, no new packages will be added to the Fedora 21
collection.
Please see https://fedoraproject.org/wiki/DNF_system_upgrade for more
information on upgrading from Fedora 21 to a newer release.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Without section/priority, reprepro will reject the packages. Add
sections following the Debian docker.io packages
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Assume that the linker can make sense of us passing in the -z,muldefs
option to tell it to ignore symbol-multiply-defined errors triggered by
https://github.com/golang/go/issues/9510. We should be able to stop
doing this once we move to Go 1.6.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com> (github: nalind)
Add a flag to allow keeping bundles around (helps with CI)
Fix several problems in repo make targets
* quote DOCKER_EXPERIMENTAL variable
* pass-through arguments for gpg provided to dpkg-sig are now quoted
properly, so passphrases with shell-interpolated symbols can be used
* when determining deb suites, don't rely on 'origin' to be
github.com/docker/docker
Fix some issues with deb repository creation from scratch
* Don't add empty components to the repository configuration as they
will cause failure when generating.
Add old docker-engine-cs name to package conflicts
Signed-off-by: Mike Dougherty <mike.dougherty@docker.com>
We're having to override it in so many places that it no longer seems worthwhile to bother. On top of that, the reason we did it in the first place was for being able to compile devicemapper statically, which still works after this change (either due to other changes in the way we build, or improvements in Go itself).
Signed-off-by: Andrew "Tianon" Page <admwiggin@gmail.com>
David Calavera