Use the script hack/build-images.sh
. It has options for building using imagebuilder
or imagestreams/buildconfigs
with the dev-builds template.
Using imagebuilder
is the default and preferred method. hack/build-images.sh
knows
how to build images based on UBI. If you use the dev-builds template
you'll need an
OpenShift cluster. Also if you want to push images into a cluster (as opposed to pushing
images into a public registry like quay.io
).
See the cluster-logging-operator hacking document for deploying OpenShift Logging using the operator make targets.
Use the script hack/update-fluentd-vendor-gems.sh
to update the vendored source
code for Fluentd. It is highly recommended to use something like rvm in order to isolate specific gem versions being used by Fluentd.
Note: Be sure to add fluentd/rh-manifest.txt
or add it to the commit if it was updated.
- Use a ruby version:
rvm use $RUBY-VERSION
- Create a gemset:
rvm gemset create $GEMSETNAME
- Use gemset:
rvm gemset use $GEMSETNAME
- Edit:
./fluentd/Gemfile
- Vendor dependencies:
./hack/update-fluentd-vendor-gems.sh
Follow the previous steps changing the task: BUNDLE_TASK=update ./hack/update-fluentd-vendor-gems.sh
Set the clobber env var: CLOBBER_VENDOR=true ./hack/update-fluentd-vendor-gems.sh
Edit the file fluentd/source.jemalloc
to also update the
vendored jemalloc source. You will have to use git add
or git rm
or otherwise
fix any conflicts, then commit and submit a PR.