Due to a glibc version mismatch in the debian repos and the
unstable-slim and testing-slim docker image 3 CI jobs have been failing
for some time. The issue was fixed by switching to testing-slim base image
and started again when the bug was introduced to the image.
Using testing-slim image as a base should fix the issue.
Signed-off-by: Ed Bartosh <eduard.bartosh@intel.com>
Due to a glibc version mismatch in the debian repos and the
unstable-slim docker image 3 CI jobs have been failing last several days.
Using testing-slim image as a base should fix the issue.
As a build optimization, allow usage of pre-created licenses.
Together with usage of go mod vendor, the container build
times are significantly shorter.
Dockerfiles are generated from include-snippets on the fly.
Signed-off-by: Ukri Niemimuukko <ukri.niemimuukko@intel.com>
An earlier commit updated the license install paths in our
distroless based images. This updates the remaining ones
to stay consistent.
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
Added LABELs (name, vendor, version, release, summary, description) to all plugins and updated
scripts/set-version.sh script to include the LABEL version='<image-tag>' by passing TAG parameter
to make set-version TAG=<image-tag>
Signed-off-by: chaitanya1731 <chaitanya.kulkarni@intel.com>
In Debian, uuid and uuid-dev are totally different things.
accel-config build depends on uuid-dev headers but the runtime
library is libuuid1. This is part of debian:sid-slim so we can
just drop installing uuid. Moreover, it's sufficient that we
only ship accel-config source.
Also, correctly ship OPAE SDK COPYING in the final image.
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
accel-config v3.4.3 dropped libkmod, libudev dependencies, update
initcontainer to v3.4.3 and drop local customizations / workarounds.
Signed-off-by: Oleg Zhurakivskyy <oleg.zhurakivskyy@intel.com>
'make pre-pull' pulls the latest (base) images so that the image
build can share the cached pulls.
Update the list of pre-pull images to what we actually use in our
Dockerfiles.
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
go mod tidy adds more requirements because of the changes in 1.17:
"Go versions 1.17 and higher retain more requirements in order to
support lazy module loading."
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
To align with other images that use Debian, move intel-vpu-plugin
container image to use unstable too.
Both the build image and the final base image use libusb-1.0 which
is ABI compatible.
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
Add checksum checks for toybox and qat driver
Use toybox version 0.8.5 instead of 0.8.4
Update toybox-config
Signed-off-by: Hyeongju Johannes Lee <hyeongju.lee@intel.com>
This is the part 2 of the work to move our published container images
to new base images. The initcontainers will be based on
distroless/static and the VPU plugin is based on Debian.
Signed-off-by: Ed Bartosh <eduard.bartosh@intel.com>
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
All but one (VPU) of the published container images can be built with
static binaries which allows us to use distroless/static as the
base image. Moreover, when combined with stripping the plugin binaries,
we can get both build time and image size savings.
This is the part 1 (out of 2) of the rework. Part 2 will finish the
change by making some adjustments to VPU plugin image and moving the
FPGA/SGX/GPU initcontainers to distroless/static too.
Partial: #516
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
Signed-off-by: Ed Bartosh <eduard.bartosh@intel.com>
This reverts commit 6c02ff51fc.
crun fails to build the images [1] but when setting BUILDAH_RUNTIME
to runc the workaround is no longer needed.
[1] https://github.com/containers/podman/issues/9378
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
buildah 1.19.4 and its dependencies when installed from the
libcontainers OBS repo, defaulted to crun and that made the
image builds to fail.
By forcing buildah to use runc instead, everything works OK.
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
This adds an nfd-hook for the gpu-plugin, which will create labels
for the GPUs that can then be used for POD deployment purposes or
creation of GPU extended resources which allow then finer grained
GPU resource management.
The nfd-hook will install to the host system when the
intel-gpu-initcontainer is run. It is added into the plugin deployment
yaml.
Signed-off-by: Ukri Niemimuukko <ukri.niemimuukko@intel.com>
