keystone/tools/config/README
Morgan Fainberg 10f9c1f1c3 Setup code for auto-config sample generation
Sync in the oslo code for auto-config sample generation and
make the appropriate changes to setup to allow for automatic
option discovery.

`test-requirements.txt` now includes stevedore to
ensure the oslo libraries can all be loaded as expected.

Change-Id: Ieca0aea4d3c67b930c12467317d2f4d83763b601
2014-02-16 20:41:00 +00:00

39 lines
1.3 KiB
Plaintext

This generate_sample.sh tool is used to generate sample config files
from OpenStack project source trees.
Run it by passing the base directory and package name i.e.
$> generate_sample.sh --base-dir /opt/stack/nova --package-name nova \
--output-dir /opt/stack/nova/etc
$> generate_sample.sh -b /opt/stack/neutron -p nova -o /opt/stack/neutron/etc
Optionally, include libraries that register entry points for option
discovery, such as oslo.messaging:
$> generate_sample.sh -b /opt/stack/ceilometer -p ceilometer \
-o /opt/stack/ceilometer/etc -l oslo.messaging
Watch out for warnings about modules like libvirt, qpid and zmq not
being found - these warnings are significant because they result
in options not appearing in the generated config file.
This check_uptodate.sh tool is used to ensure that the generated sample
config file in the OpenStack project source tree is continually kept up
to date with the code itself.
This can be done by adding a hook to tox.ini. For example, if a project
already had flake8 enabled in a section like this:
[testenv.pep8]
commands =
flake8 {posargs}
This section would be changed to:
[testenv.pep8]
commands =
flake8 {posargs}
{toxinidir}/tools/config/check_uptodate.sh