Managing Kudu

This topic describes the tasks you can perform to manage the Kudu service using Cloudera Manager. You can use the Kudu service to upgrade the Kudu service, start and stop the Kudu service, monitor operations, and configure the Kudu master and tablet servers, among other tasks. Depending on your deployment, there are several different configuration settings you may need to modify.

Continue reading:

For detailed information about Apache Kudu, view the Kudu documentation.

Installing and Upgrading the Kudu Service

You can install Kudu through the Cloudera Manager installation wizard, using either parcels or packages. For instructions, see Installing Kudu.

For instructions on upgrading Kudu using parcels or packages, see Upgrading Kudu Using Cloudera Manager.

Enabling Core Dump for the Kudu Service

If Kudu crashes, you can use Cloudera Manager to generate a core dump to get more information about the crash.

  1. Go to the Kudu service.
  2. Click the Configuration tab.
  3. Search for core dump.
  4. Check the checkbox for the Enable Core Dump property.
  5. (Optional) Unless otherwise configured, the dump file is generated in the default core dump directory, /var/log/kudu, for both the Kudu master and the tablet servers.
    • To configure a different dump directory for the Kudu master, modify the value of the Kudu Master Core Dump Directory property.
    • To configure a different dump directory for the Kudu tablet servers, modify the value of the Kudu Tablet Server Core Dump Directory property.
  6. Click Save Changes.

Verifying the Impala Dependency on Kudu

In a Cloudera Manager deployment, once Kudu is installed, Impala will automatically identify the Kudu master. However, in case your queries don't work as expected, make sure you check that the Impala service is set to be dependent on Kudu.

  1. Go to the Impala service.
  2. Click the Configuration tab.
  3. Search for kudu.
  4. Make sure the Kudu Service property is set to the right Kudu service.
  5. Click Save Changes.

Using the Charts Library with the Kudu Service

By default, the Status tab for the Kudu service displays a dashboard containing a limited set of charts. For details on the terminology used in these charts, and instructions on how to query for time-series data, display chart details, and edit charts, see Charting Time-Series Data.

The Kudu service's Charts Library tab also displays a dashboard containing a much larger set of charts, organized by categories such as process charts, host charts, CPU charts, and so on, depending on the entity (service, role, or host) that you are viewing. You can use these charts to keep track of disk space usage, the rate at which data is being inserted/modified in Kudu across all tables, or any critical cluster events. You can also use them to keep track of individual tables. For example, to find out how much space a Kudu table is using on disk:

  1. Go to the Kudu service and navigate to the Charts Library tab.
  2. On the left-hand side menu, click Tables to display the list of tables currently stored in Kudu.
  3. Click on a table name to view the default dashboard for that table. The Total Tablet Size On Disk Across Kudu Replicas chart displays the total size of the table on disk using a time-series chart.

    Hovering with your mouse over the line on the chart opens a small pop-up window that displays information about that data point. Click the data stream within the chart to display a larger pop-up window that includes additional information for the table at the point in time where the mouse was clicked.