layout: docs31 title: Set Up System Cube categories: tutorial permalink: /docs31/tutorial/setup_systemcube.html

Available since Apache Kylin v2.3.0

Main content of this section:

  • [What is System Cube](#What is System Cube)
  • [How to Set Up System Cube](#How to Set Up System Cube)
  • [Automatically create System Cube](#Automatically create System Cube)
  • [Details of System Cube](#Details of System Cube)

What is System Cube

For better supporting self-monitoring, a set of system Cubes are created under the system project, called “KYLIN_SYSTEM”. Currently, there are five Cubes. Three are for query metrics, “METRICS_QUERY”, “METRICS_QUERY_CUBE”, “METRICS_QUERY_RPC”. And the other two are for job metrics, “METRICS_JOB”, “METRICS_JOB_EXCEPTION”.

How to Set Up System Cube

In this section, we will introduce the method of manually enabling the system cube. If you want to automatically enable the system cube through shell scripts, please refer to [Automatically Create System Cube](#Automatically Create System Cube).

1. Prepare

Create a configuration file SCSinkTools.json in KYLIN_HOME directory.

For example:

[
    {
       "sink": "hive",
       "storage_type": 2,
       "cube_desc_override_properties": {
         "kylin.cube.algorithm": "INMEM",
         "kylin.cube.max-building-segments": "1"
       }
    }
]

2. Generate Metadata

Run the following command in KYLIN_HOME folder to generate related metadata:

./bin/kylin.sh org.apache.kylin.tool.metrics.systemcube.SCCreator \
-inputConfig SCSinkTools.json \
-output <output_folder>

By this command, the related metadata will be generated and its location is under the directory <output_folder>. The details are as follows, system_cube is our <output_folder>:

metadata

3. Set Up Datasource

Running the following command to create source hive tables:

hive -f <output_folder>/create_hive_tables_for_system_cubes.sql

By this command, the related hive table will be created.

hive_table

4. Upload Metadata for System Cubes

Then we need to upload metadata to hbase by the following command:

./bin/metastore.sh restore <output_folder>

5. Reload Metadata

Finally, we need to reload metadata in Kylin web UI.

Then, a set of system Cubes will be created under the system project, called “KYLIN_SYSTEM”.

6. System Cube build

When the system Cube is created, we need to build the Cube regularly.

Step 1. Create a shell script that builds the system Cube by calling org.apache.kylin.tool.job.CubeBuildingCLI

For example:

{% highlight Groff markup %} #!/bin/bash

dir=$(dirname ${0}) export KYLIN_HOME=${dir}/../

CUBE=$1 INTERVAL=$2 DELAY=$3 CURRENT_TIME_IN_SECOND=date +%s CURRENT_TIME=$((CURRENT_TIME_IN_SECOND * 1000)) END_TIME=$((CURRENT_TIME-DELAY)) END=$((END_TIME - END_TIME%INTERVAL))

ID=“$END” echo “building for ${CUBE}_${ID}” >> ${KYLIN_HOME}/logs/build_trace.log sh ${KYLIN_HOME}/bin/kylin.sh org.apache.kylin.tool.job.CubeBuildingCLI --cube ${CUBE} --endTime ${END} > ${KYLIN_HOME}/logs/system_cube_${CUBE}_${END}.log 2>&1 &

{% endhighlight %}

Step 2. Then run this shell script regularly. For example, add a cron job as follows:

{% highlight Groff markup %} 0 */2 * * * sh ${KYLIN_HOME}/bin/system_cube_build.sh KYLIN_HIVE_METRICS_QUERY_QA 3600000 1200000

20 */2 * * * sh ${KYLIN_HOME}/bin/system_cube_build.sh KYLIN_HIVE_METRICS_QUERY_CUBE_QA 3600000 1200000

40 */4 * * * sh ${KYLIN_HOME}/bin/system_cube_build.sh KYLIN_HIVE_METRICS_QUERY_RPC_QA 3600000 1200000

30 */4 * * * sh ${KYLIN_HOME}/bin/system_cube_build.sh KYLIN_HIVE_METRICS_JOB_QA 3600000 1200000

50 */12 * * * sh ${KYLIN_HOME}/bin/system_cube_build.sh KYLIN_HIVE_METRICS_JOB_EXCEPTION_QA 3600000 12000

{% endhighlight %}

Automatically create System Cube

Kylin provides system-cube.sh from v2.6.0, users can automatically create system cube by executing this script.

  • Create System Cube：sh bin/system-cube.sh setup

  • Build System Cube：sh bin/system-cube.sh build

  • Add crontab job for System Cube：sh bin/system-cube.sh cron

Note: System-cube.sh will call ${KYLIN_HOME}/bin/build-incremental-cube.sh to submit build job. In build-incremental-cube.sh, the user name and password of ADMIN:KYLIN are used by default as authentication to call kylin's rebuild API. If you have changed the password of the ADMIN user or want to use a user other than ADMIN to submit the build job, please find the ADMIN:KYLIN in build-incremental-cube.sh and replace it with the correct user name and password.

Details of System Cube

Common Dimension

For all of these Cube, admins can query at four time granularities. From higher level to lower, it's as follows:

METRICS_QUERY

This Cube is for collecting query metrics at the highest level. The details are as follows:

METRICS_QUERY_RPC

This Cube is for collecting query metrics at the lowest level. For a query, the related aggregation and filter can be pushed down to each rpc target server. The robustness of rpc target servers is the foundation for better serving queries. The details are as follows:

METRICS_QUERY_CUBE

This Cube is for collecting query metrics at the Cube level. The most important are cuboids related, which will serve for Cube planner. The details are as follows:

METRICS_JOB

In Kylin, there are mainly three types of job:

  • “BUILD”, for building Cube segments from HIVE.
  • “MERGE”, for merging Cube segments in HBASE.
  • “OPTIMIZE”, for dynamically adjusting the precalculated cuboid tree base on the base cuboid in HBASE.

This Cube is for collecting job metrics. The details are as follows:

METRICS_JOB_EXCEPTION

This Cube is for collecting job exception metrics. The details are as follows: