Description of problem: After dual-stack is enabled in cluster, etcd operator will go degraded as it checks the validity of the current certs against node ip's. Since the node receive new ipv6 IP, the check will fail, because the certificate is not valid for that IP... There are two options, we either regenerate the certificates and include additional IPv6 address in the cert OR we just skip ipv6 address, because we don't really need it for internal communication (ipv4 is still working). xref: https://bugzilla.redhat.com/show_bug.cgi?id=2039235 Version-Release number of selected component (if applicable): 4.10 (but will require backport down to 4.8) Steps to Reproduce: 1. enable dual-stack 2. watch etcd go degraded on certificate check 3. Actual results: Expected results: Additional info:
We should be setting the Version to the version we believe first introduced or have confirmed the problem exists in and intend to fix. Target version gets set to the relevant branch where it's being fixed. This ensures that even looking just at the bug metadata it's clear which version range is affected.