Merge pull request #303037 from eak13/main

Adding in MHC remediation information

Author: JillGrant615

articles/operator-nexus/concepts-rack-resiliency.md

@@ -1,8 +1,8 @@
 ---
 title: Operator Nexus rack resiliency
-description: Document how rack resiliency works in Operator Nexus Near Edge
+description: Document how rack resiliency works in Operator Nexus
 ms.topic: article
-ms.date: 06/03/2025
+ms.date: 07/21/2025
 author: eak13
 ms.author: ekarandjeff
 ms.service: azure-operator-nexus
@@ -19,9 +19,9 @@ Operator Nexus ensures the availability of three active Kubernetes control plane
 > [!TIP]
 > The Kubernetes control plane is a set of components that manage the state of a Kubernetes cluster, schedule workloads, and respond to cluster events. It includes the API server, etcd storage, scheduler, and controller managers.
 >
-> The remaining management nodes contain various operators which run the platform software and other components performing support capabilities for monitoring, storage, and networking.
+> The remaining management nodes contain various operators that run the platform software and other components performing support capabilities for monitoring, storage, and networking.
 
-During runtime upgrades, Operator Nexus implements a sequential upgrade of the control plane nodes which preserves resiliency throughout the upgrade process.
+During runtime upgrades, Operator Nexus implements a sequential upgrade of the control plane nodes. The sequential node approach preserves resiliency throughout the upgrade.
 
 Three compute racks:
 
@@ -65,7 +65,31 @@ Operator Nexus supports control plane resiliency in single rack configurations b
 
 In disaster situations when the control plane loses quorum, there are impacts to the Kubernetes API across the instance. This scenario can affect a workload's ability to read and write Custom Resources (CRs) and talk across racks.
 
-## Related Links
+## Automated remediation
+
+To maintain Kubernetes control plane (KCP) quorum, Operator Nexus provides automated remediation when specific server issues are detected. Additionally, this automated remediation extends to Management Plane & Compute nodes.
+
+Here are the triggers for automated remediation:
+
+*   For all servers (Compute, Management and KCP): if a server fails to provision successfully after six hours, automated remediation occurs. This check includes provisioning a new Bare Metal Machine (BMM) at initial deployment time or provisioning during a Replace action.
+*   For all servers (Compute, Management and KCP): if a running node is stuck in a read only root file system mode for 10 minutes, automated remediation occurs.
+*   For KCP and Management Plane servers only, if a Kubernetes node is in an Unknown state for 30 minutes, automated remediation occurs.
+
+### Remediation process
+
+*   Remediation of a Compute node is now one reprovisioning attempt. If the reprovisioning fails, the node is marked Unhealthy. Reprovisioning no longer continues to retry infinitely, and the Bare Metal Machine is powered off. 
+*   Remediation of a Management Plane node is to attempt one reboot and then one reprovisioning attempt. If those steps fail, the node is marked Unhealthy. 
+*   Remediation of a KCP node is to attempt one reboot. If the reboot fails, the node is marked Unhealthy and Nexus triggers the immediate provisioning of the spare KCP node. This process is outlined in the [KCP remediation details](#kcp-remediation-details) section.
+*   In all instances, when the Bare Metal Machine is marked unhealthy, the BMM's `detailedStatusMessage` is updated to read `Warning: BMM Node is unhealthy and may require hardware replacement.` The Bare Metal Machine's node is removed from the Kubernetes Cluster, which triggers a node drain. Users need to run a BMM Replace action to return the BMM into service and have it rejoin the Kubernetes Cluster.
+
+### KCP remediation details
+
+Ongoing control plane resiliency requires a spare KCP node. When KCP node fails remediation and is marked Unhealthy, a deprovisioning of the node occurs. The unhealthy KCP node is exchanged with a suitable healthy Management Plane server. This Management Plane server becomes the new spare KCP node. The failed KCP node is updated and labeled as a Management Plane node. Once the label changes, an attempt to provision the newly labeled management plane node occurs. If it fails to provision, the management plane remediation process takes over. If it fails provisioning or doesn't run successfully, the machine's status remains unhealthy, and the user must fix. The unhealthy condition surfaces to the Bare Metal Machine's (BMM) `detailedStatus` and `detailedStatusMessage` fields in Azure and clears through a BMM Replace action.
+
+> [!NOTE] 
+>The provisioning retry process doesn't execute on compute and management node pool nodes for systems running the 4.1 NetworkCloud runtime. This capability is available when the Nexus Cluster is updated to the 4.4 runtime.
+
+## Related links
 
 [Determining Control Plane Role](./reference-near-edge-baremetal-machine-roles.md)