Merge remote-tracking branch 'origin/release-5.8' into release-5.9

Author: zmstone

en_US/deploy/rolling-upgrades.md

@@ -34,6 +34,16 @@ To upgrade each node in the cluster without downtime, follow these steps:
 4. Install a new version of EMQX.
 5. Start the new version node.
 
+**Upgrading Core and Replicant Nodes**
+
+In a core/replicant cluster, upgrade nodes in alternating groups to maintain cluster stability:
+
+- Begin with one core node.
+- Then upgrade a proportionate subset of replicant nodes (for example, roughly one-third if there are three core nodes).
+- Continue alternating between core and replicant nodes until all nodes are upgraded.
+
+This approach ensures replicants always have compatible cores to connect to during the rollout.
+
 :::tip Note
 Do not perform cluster-wide config changes during a rolling upgrade. Configuration changes made from the Dashboard, HTTP API, or CLI are applied to all nodes in the cluster. Making configuration changes during a rolling upgrade may cause nodes to become out of sync.
 :::

zh_CN/deploy/rolling-upgrades.md

@@ -25,14 +25,23 @@
 
 ## 如何执行滚动升级
 
-集群节点的滚动升级过程如下：
+**集群节点升级过程**
 
 1. 使用集群重平衡将客户端从节点中清除。(可选)
 2. 停止旧版本节点。
 3. [备份](../operations/backup-restore.md)节点的配置文件和数据目录。
 4. 安装新版本的 EMQX。
 5. 启动新版本节点。
 
+**Core/replicant 集群升级过程***
+
+在 Core/Replicant 集群中，应分组交替升级节点，以保持集群的稳定性：
+- 首先升级一个 核心节点（core node）。
+- 然后升级一部分 从属节点（replicant nodes）（例如，当有三个核心节点时，可先升级大约三分之一的从属节点）。
+- 继续在核心节点和从属节点之间交替升级，直到所有节点都完成升级。
+
+这种方式可确保在整个升级过程中，从属节点始终能够连接到与其版本兼容的核心节点。
+
 :::tip 注意
 在滚动升级期间，请勿进行集群范围的配置更改。从 Dashboard、HTTP API 或 CLI 进行的配置更改将应用于集群中的所有节点。在滚动升级期间进行配置更改可能会导致节点不同步。
 :::