openshift-must-gather is a tool for collecting cluster data.
It dumps clusteroperator data, and associated namespace data, into a specified --base-dir location.
The directory structure, as well as specific details behind this tool can be found in this doc.
Data collection scripts are kept in ./collection-scripts. The content of that folder is placed in /usr/bin in the image.
The data collection scripts should only include collection logic for components that are included as part of the OpenShift
CVO payload. Outside components are encouraged to produce a similar "must-gather" image, but this is not the spot to be
included.
Place in GOPATH under src/github.com/openshift/must-gather.
Build with:
$ make
Before Running: Make sure you have a KUBECONFIG environment variable set, and that it points to a valid admin kubeconfig file.
Alternatively, provide a filepath to a valid admin.kubeconfig file via the --kubeconfig flag.
Run with:
./bin/openshift-must-gather inspect clusteroperator/<name>
Note: it is possible to run this tool to collect all clusteroperator data by omitting a <name>:
./bin/openshift-must-gather inspect clusteroperators
WARNING: The following tool is provided with no guarantees and might (and will) be changed at any time. Please do not rely on anything below in your scripts or automatization.
Beside the oppenshift-must-gather binary the openshift-dev-helpers binary is also provided. This binary combines various
tools useful for the OpenShift developers teams.
To list all events recorded during a test run and stored in events.json file, you can run this command:
./bin/openshift-dev-helpers events https://storage.googleapis.com/origin-ci-test/pr-logs/.../artifacts/e2e-aws/events.json --component=openshift-apiserver-operator