Overview¶
Dashboard¶
Log in to MRS Manager. Choose Cluster, click the name of the desired cluster, and choose Services. The service management page is displayed, including the functional area and service list.
Functional Area¶
In the functional area of the service management page, you can select a view type and filter and search for services by service type. You can use the advanced search to select required services based on the running status and configuration status.
Service List¶
The service list on the service management page contains all installed services in the cluster. If the tile view is selected, the services will be displayed in pane style. If you select the list view, the services will be displayed in a table.
Note
In this section, the Tile View is used by default.
The service list displays the running status, configuration status, role type, and number of instances of each service. On this page, you can perform some service maintenance tasks, such as starting, stopping, and restarting services.
Status | Description |
|---|---|
Normal | Indicates that the service is running properly. |
Faulty | Indicates that the service cannot run properly. |
Partially Healthy | Indicates that some enhanced functions of the service are abnormal. |
Not started | Indicates that the service is stopped. |
Unknown | Indicates that the initial status of the service cannot be detected. |
Starting | Indicates that the service is being started. |
Stopping | Indicates that the service is being stopped. |
Failed to start | Indicates that the service fails to be started. |
Failed to stop | Indicates that the service fails to be stopped. |
Note
If the running status of a service is Faulty, an alarm is generated. Rectify the fault based on the alarm information.
HBase, Hive, Spark, and Loader may be in the Subhealthy state.
If Yarn is installed but is abnormal, HBase is in the Subhealthy state. If the multi-instance function is enabled, all installed HBase service instances are in the Subhealthy state.
If HBase is installed but is abnormal, Hive, Spark, and Loader are in the Subhealthy state.
If any HBase instance is installed but is abnormal after the multi-instance function is enabled, Loader is in the Subhealthy state.
If an HBase instance is installed but is abnormal after the multi-instance function is enabled, the Hive and Spark instances that map to the HBase instance are in the Subhealthy state. That is, if HBase 2 is installed but is abnormal, Hive 2 and Spark2 are in the Subhealthy state.
Status | Description |
|---|---|
Synchronized | Indicates that all service parameter settings have taken effect in the cluster. |
Expired | Indicates that the latest configuration is not synchronized and does not take effect after the service parameters are modified. You need to synchronize the configurations and restart the services. You can click |
Failed | Indicates that a communication or read/write exception occurs during the parameter configuration synchronization. Use Synchronize Configuration to rectify the fault. |
Synchronizing | Indicates that the service parameter configuration is being synchronized. |
Unknown | Indicates that the initial status of the service cannot be detected. |
next to Configuration Status to view expired configuration items.You can click a service in the service list to perform simple maintenance and management operations on the service, as described in Table 3.
Menu Item on the UI | Description |
|---|---|
Start Service | Start a specified service in the cluster. |
Stop Service | Stop a specified service in the cluster. |
Restart Service | Restart a specified service in the cluster. Note If a service is restarted, other services that depend on this service will be unavailable. Therefore, select Restart upper-layer services. Determine whether to perform this operation based on the displayed service list. Services are restarted one by one due to their dependency. Table 4 describes the restart duration of a single service. |
Service Rolling Restart | Restart a specified service in the cluster without interrupting services. For details about the parameter settings, see Table 1. |
Synchronize Configuration |
Note After some services are synchronized, restart the services for the settings to take effect. |
Service | Restart Duration | Startup Duration | Remarks |
|---|---|---|---|
ClickHouse | 4 min | ClickHouseServer: 2 min ClickHouseBalancer: 2 min |
|
HDFS | 10min+x | NameNode: 4 min + x DataNode: 2 min JournalNode: 2 min Zkfc: 2 min | x indicates the NameNode metadata loading duration. It takes about 2 minutes to load 10,000,000 files. For example, x is 10 minutes for 50 million files. The startup duration fluctuates with reporting of DataNode data blocks. |
Yarn | 5 min + x | ResourceManager: 3 min + x NodeManager: 2 min | x indicates the time required for restoring ResourceManager reserved tasks. It takes about 1 minute to restore 10,000 reserved tasks. |
MapReduce | 2 min + x | JobHistoryServer: 2 min + x | x indicates the scanning duration of historical tasks. It takes about 2.5 minutes to scan 100,000 tasks. |
ZooKeeper | 2 min + x | quorumpeer: 2 min + x | x indicates the duration for loading znodes. It takes about 1 minute to load 1 million znodes. |
Hive | 3.5 min | HiveServer: 3 min MetaStore: 1 min 30s WebHcat: 1 min Hive service: 3 min |
|
Spark2x | 5 min | JobHistory2x: 5 min SparkResource2x: 5 min JDBCServer2x: 5 min |
|
Flink | 4 min | FlinkResource: 1 min FlinkServer: 3 min |
|
Kafka | 2 min + x | Broker: 1 min + x | x indicates the data restoration duration. It takes about 2 minutes to start 20,000 partitions for a single instance. |
Storm | 6 min | Nimbus: 3 min UI: 1 min Supervisor: 1 min Logviewer: 1 min |
|
Flume | 3 min | Flume: 2 min MonitorServer: 1 min |
|