Use postgres instead of postgresql for the image name to align with the
Docker official library upstream.
Signed-off-by: George T Kramer <george.t.kramer@intel.com>
The image name was hardcoded. We should receive a variable with the
image name instead, this will let the end user select which DLRS
flavour (mkl or oss) to use.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
The argument (swupd_args) defaults to the latest validated version on
the CLI, but the user still has the option to pass X version as swupd argument.
A note suggesting to use the latest validated version is added to the
README as a disclaimer.
Fixesclearlinux/dockerfiles#65
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
The stacks team validates and recommends a certain Clear Linux OS
version, stating it on the Release Notes ensures functionality
and avoids possible issues with not validated versions.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
A message stating that the TensorFlow used by the
stacks-dlrs-mkl image does not support AVX instructions
can be missleading for users.
This change adds a note clarifying how the performance
optimization is handled in this image.
Signed-off-by: Jose Lamego <jose.a.lamego@intel.com>
Removing parts of the description that do not add value until a
legally-approved description is available that differentiates each
offering. Making formatting changes to enhance the appearance of the
long description. This will ensure that when automated builds occur on
Docker Hub, the description remains correct and does not revert to an
older, non-approved version.
Signed-off-by: George T Kramer <george.t.kramer@intel.com>
This provides a Clear Linux Deep Learning Stack. To offer more flexibility,
there are two versions of the Deep Learning Stack: a "fully open source"
version that includes only open source components, and a second, "mostly open
source" version that includes the TensorFlow* framework optimized using Intel®
Math Kernel Library for Deep Neural Networks (Intel® MKL-DNN) primitives.
Previously, an already-existent 'clr-sdk' container could not be restarted
because the 'setup.py' ENTRYPOINT script would error out attempting to
create the 'clr' user again.
This patch adds new exception handling specifically for this case, ignoring
these errors. Any other error encountered creating the 'clr' user and group
is still considered a critical failure.
With this, an existent container can be restarted with 'docker start -i <id>'.
This patch also makes minor corrections and clarification to the README.
Signed-off-by: Kevin C. Wells <kevin.c.wells@intel.com>
This pull request introduces an image to be used for running
the Clear Linux OS Installer continuous integration on Travis.
Previously, a very large part of the test time in Travis was
spent building adding the required bundle to the standard
Clear Linux Docker image to properly setup the test environment.
Please ensure that auto-trigger is enabled so that this image
is automatically rebuilt with each new base clearlinux docker image.
Signed-off-by: Mark D Horn <mark.d.horn@intel.com>
This commit removes the restriction of root nopw access to only the
mixer and swupd binaries. The 'clr' user is now part of the
wheelnopw group.
Signed-off-by: Kevin C. Wells <kevin.c.wells@intel.com>
Recent changes to the Clear Linux base image cause the final line
in the setup.py script to yield an error message and exit.
There are numerous ways this could be fixed, but this fix (suggested
by @shrmrf) is the minimal change needed.
Signed-off-by: Kevin C. Wells <kevin.c.wells@intel.com>
This patch introduces an image used for running the Mixer CI on Travis.
Previously, a very large part of the time Travis spent building and
validating Mixer PR's was spent 'swupd bundle-add'ing the packages
needed. The _mixer_ build was very short once this step was done.
This image removes the need to build and configure the image with each
Travis run; only the code needs to be copied over before the mixer
build and tests can run.
Signed-off-by: Kevin C. Wells <kevin.c.wells@intel.com>
George T Kramer