At the moment there is no way how to kill ironic discovery than to wait for timeout (few hours). So now when user realizes that discovery got stuck, he very likely has to wait few hours without possibility to kill the process and start it again.
Upstream blueprint: https://blueprints.launchpad.net/ironic-discoverd/+spec/abort-introspection But we're past FF for discoverd 1.1 series. Are we ok with cherry-picking such a big change? Or will we track discoverd 1.2 instead? also note that timeout defaults to 1 hr and can be changed via 'timeout' option of discoverd.conf
Patch merged in upstream-inspector upstream: https://github.com/openstack/ironic-inspector/commit/7a3d93700e37a32257ea16ef5cca43e060d3ea9a
We now have this: https://docs.openstack.org/developer/tripleo-docs/troubleshooting/troubleshooting-nodes.html so this can be closed.
Wow, that's an old bug. Thanks, it can be closed indeed.