[Doc] Brokerless doc update (backport #52346) (#52349)

Co-authored-by: Dan Roscigno <[email protected]>
StarRocks · Oct 26, 2024 · c4acd52 · c4acd52
1 parent 879e9cf
commit c4acd52
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/docs/en/sql-reference/sql-statements/loading_unloading/BROKER_LOAD.md b/docs/en/sql-reference/sql-statements/loading_unloading/BROKER_LOAD.md
@@ -256,7 +256,7 @@ Open-source HDFS supports two authentication methods: simple authentication and
 
     - To perform broker-based loading, make sure that at least one independent broker group is deployed, and place the `hdfs-site.xml` file to the `{deploy}/conf` path on the broker node that serves the HDFS cluster. StarRocks will add the `{deploy}/conf` path to the environment variable `CLASSPATH` upon broker startup, allowing the brokers to read information about the HDFS cluster nodes.
 
-    - To perform broker-free loading, place the `hdfs-site.xml` file to the `{deploy}/conf` paths of each FE node and each BE or CN node.
+    - To perform broker-free loading, you only need to set `hadoop.security.authentication = kerberos` in `conf/core-site.xml` under the deployment directories of all FE, BE, and CN nodes in your cluster, and use the `kinit` command to configure the Kerberos account.
 
   - If you load data from a single HDFS cluster that has multiple Kerberos users configured, only broker-based loading is supported. Make sure that at least one independent broker group is deployed, and place the `hdfs-site.xml` file to the `{deploy}/conf` path on the broker node that serves the HDFS cluster. StarRocks will add the `{deploy}/conf` path to the environment variable `CLASSPATH` upon broker startup, allowing the brokers to read information about the HDFS cluster nodes.
 

diff --git a/docs/zh/sql-reference/sql-statements/loading_unloading/BROKER_LOAD.md b/docs/zh/sql-reference/sql-statements/loading_unloading/BROKER_LOAD.md
@@ -253,7 +253,7 @@ StarRocks 访问存储系统的认证配置。
 
     - 如果采用有 Broker 的导入，您需要确保至少部署了一组独立的 Broker，并将 `hdfs-site.xml` 文件放在 HDFS 集群对应的 Broker 节点的 `{deploy}/conf` 目录下。Broker 进程重启时，会将 `{deploy}/conf` 目录添加到 `CLASSPATH` 环境变量，使 Broker 能够读取 HDFS 集群中各节点的信息。
 
-    - 如果采用无 Broker 的导入，您需要将 `hdfs-site.xml` 文件放在每个 FE 节点和每个 BE（或 CN）节点的 `{deploy}/conf` 目录下。
+    - 如果采用无 Broker 的导入，您需要在各个 FE、BE、CN 节点的部署路径下的 `conf/core-site.xml` 文件中设置 `hadoop.security.authentication = kerberos`，并通过 `kinit` 命令配置 Kerberos 账号。
 
   - 在单 HDFS 集群、并且配置了多 Kerberos 用户的场景下，只支持有 Broker 的导入。您需要确保至少部署了一组独立的 Broker，并将 `hdfs-site.xml` 文件放在 HDFS 集群对应的 Broker 节点的 `{deploy}/conf` 目录下。Broker 进程重启时，会将 `{deploy}/conf` 目录添加到 `CLASSPATH` 环境变量，使 Broker 能够读取 HDFS 集群中各节点的信息。