From c28205e9341924a8197e485608cb0ff6a27bd3a6 Mon Sep 17 00:00:00 2001 From: Ted Avery Date: Fri, 17 Oct 2025 07:40:38 -0400 Subject: [PATCH] OSDOCS-15028:tavery-etcd-1:RN Bug Batch --- release_notes/ocp-4-20-release-notes.adoc | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/release_notes/ocp-4-20-release-notes.adoc b/release_notes/ocp-4-20-release-notes.adoc index 74a70169d3..d0546cbbe7 100644 --- a/release_notes/ocp-4-20-release-notes.adoc +++ b/release_notes/ocp-4-20-release-notes.adoc @@ -96,6 +96,11 @@ The `siteconfig-converter` tool does not convert SiteConfig CRs that use the dep For more information, see xref:../edge_computing/ztp-migrate-clusterinstance.adoc#ztp-migrate-clusterinstance[Migrating from SiteConfig CRs to ClusterInstance CRs]. +[id="ocp-release-notes-etcd_{context}"] +=== etcd + +With this update, the Cluster etcd Operator introduces alert levels for the `etcdDatabaseQuotaLowSpace` alert, offering administrators timely notifications about low etcd quota usage. This proactive alert system aims to prevent API server instability and allows for effective resource management in managed OpenShift clusters. The alert levels are `info`, `warning`, and `critical`, providing a more granular approach to monitoring etcd quota usage, which results in dynamic etcd quota management and improved overall cluster performance. + [id="ocp-release-notes-extensions_{context}"] === Extensions ({olmv1}) @@ -1550,6 +1555,15 @@ As a result, the controller handles errors during migration better. * Before this update, importing multiple files in the YAML editor copied the existing content and appended the new file, which created duplicates. With this release, the import behavior is fixed. As a result, the YAML editor displays only the new file content without duplication. (link:https://issues.redhat.com/browse/OCPBUGS-45297[OCPBUGS-45297]) +[id="ocp-release-note-etcd-bug-fixes_{context}"] +=== etcd + +* Before this update, the timeout on one etcd member caused context deadlines to exceed. As a consequence, all members were declared unhealthy, even though some were reachable. With this release, if one member times out, other members are no longer incorrectly marked as unhealthy. (link:https://issues.redhat.com/browse/OCPBUGS-60941[OCPBUGS-60941]) + +* Before this update, when you deployed {sno} with many IPs on the primary interface, the IP in the etcd certificate mismatched with the IP in the config map that the API server used to connect to etcd. As a consequence, the API server pod failed during {sno} deployment, which caused cluster initialization issues. With this release, the single IP in the etcd config map matches the IP in the certificate for {sno} deployments. As a result, the API server connects to etcd by using the correct IP included in the etcd certificate, which prevents pod failure during cluster initialization. (link:https://issues.redhat.com/browse/OCPBUGS-55404[OCPBUGS-55404]) + +* Before this update, during temporary downtime of the API server, the Cluster etcd Operator reported incorrect information, such as messages that the `openshift-etcd` namespace was non-existent. With this update, the Cluster etcd Operator status message correctly indicates API server unavailability instead of suggesting the absence of the `openshift-etcd` namespace. As a result, the Cluster etcd Operator status accurately reflects the presence of the `openshift-etcd` namespace, enhancing system reliability. (link:https://issues.redhat.com/browse/OCPBUGS-44570[OCPBUGS-44570]) + [id="ocp-release-note-extensions-olmv1-bug-fixes_{context}"] === Extensions ({olmv1})