Rate This Document
Findability
Accuracy
Completeness
Readability

Installing OmniData

Distribute the OmniData package to all nodes, decompress the package, and install and initialize OmniData.

  1. Copy the OmniData binary package prepared in Obtaining Software to all nodes using the file transfer function.
    1. Select offload nodes (ceph1, ceph2, and ceph3) and click Configure Service.

    2. Select Custom and click Next.

    3. Select File Transfer and click Add a Task.

    4. Set task 1 parameters and click Save.
      Table 1 Task 1 parameters

      Parameter

      Description

      Task Name

      Transferring the OmniData installation package

      Task Description

      Copying the OmniData installation package to offload nodes

      Transfer Direction

      From Local to Remote

      Local Path

      Directory where the local installation package is stored

      Remote Path

      /home/omm/omnidata-software

  2. Copy the OmniData third-party dependencies prepared in Obtaining Software to offload nodes using the file transfer function.
    1. Select File Transfer and click Add a Task.

    2. Set parameters for tasks 2 to 5 and click Save.
      Table 2 Task 2 parameters

      Parameter

      Description

      Task Name

      Copying the OmniData lib package

      Task Description

      Copying the OmniData lib package to offload nodes

      Transfer Direction

      From Local to Remote

      Local Path

      Directory where the local installation package is stored

      Remote Path

      /home/omm/omnidata-software

      Table 3 Task 3 parameters

      Parameter

      Description

      Task Name

      Copying Ceph dependencies

      Task Description

      Copying Ceph dependencies to all offload nodes (Skip this step for an HDFS cluster.)

      Transfer Direction

      From Local to Remote

      Local Path

      Directory where the local installation package is stored

      Remote Path

      /home/omm/omnidata-software

      Table 4 Task 4 parameters

      Parameter

      Description

      Task Name

      Copying the librgw file

      Task Description

      Copying the librgw dependency to all offload nodes (Skip this step for an HDFS cluster.)

      Transfer Direction

      From Local to Remote

      Local Path

      Directory where the local installation package is stored

      Remote Path

      /home/omm/omnidata-software

  3. Decompress the OmniData package boostkit-omnidata-server-1.4.0-aarch64.tar.gz and copy the third-party dependencies to the corresponding directories.
    1. Select Command Execution and click Add a Task.

    2. Set parameters for tasks 6 to 8 and click Save.
      Table 5 Task 5 parameters

      Parameter

      Description

      Task Name

      Decompressing OmniData packages

      Task Description

      Decompressing the OmniData installation package and third-party dependency packages

      Run Directory

      /home/omm/omnidata-software

      Run Command

      unzip BoostKit-omnidata_1.4.0.zip; tar -zxvf BoostKit-omnidata_1.4.0.tar.gz;tar -zxvf boostkit-omnidata-server-1.4.0-aarch64.tar.gz -C /home/omm/omnidata-install; unzip boostkit-omnidata-server-1.4.0-aarch64-lib.zip;

      Whether to Check Return Code

      Yes

      Expected Return Code

      0

      Table 6 Task 6 parameters

      Parameter

      Description

      Task Name

      Copying OmniData dependencies

      Task Description

      Copying third-party dependencies to the corresponding OmniData directories

      Run Directory

      /home/omm/omnidata-software

      Run Command

      mv /home/omm/omnidata-software/boostkit-omnidata-server-1.4.0-aarch64-lib/* /home/omm/omnidata-install/omnidata/lib;rmdir /home/omm/omnidata-software/boostkit-omnidata-server-1.4.0-aarch64-lib;cd /home/omm/omnidata-install;cp -f ./haf-offload/lib/jar/haf-1.3.0.jar ./omnidata/lib/;

      Whether to Check Return Code

      Yes

      Expected Return Code

      0

      Table 7 Task 7 parameters

      Parameter

      Description

      Task Name

      Copying Ceph dependencies

      Task Description

      Copying the Ceph dependencies to the corresponding OmniData directories (Skip this step for an HDFS cluster.)

      Run Directory

      /home/omm/omnidata-software

      Run Command

      cd /home/omm/omnidata-install;chmod 700 ./haf-offload/lib/; cp /home/omm/omnidata-software/librgw_jni.so ./haf-offload/lib/;cp /usr/lib64/ceph/libceph-common.so.0 ./haf-offload/lib/;chmod -R 500 ./haf-offload/lib/;cp /home/omm/omnidata-software/hdfs-ceph-3.2.0.jar ./omnidata/lib/;

      Whether to Check Return Code

      Yes

      Expected Return Code

      0

  4. Install the OmniData service on offload nodes (ceph1, ceph2, and ceph3).
    1. Select Command Execution and click Add a Task.

    2. Set task 8 parameters and click Save.
      Table 8 Task 8 parameters

      Parameter

      Description

      Task Name

      Installing the OmniData service

      Task Description

      Installing the OmniData service

      Run Directory

      /home/omm/omnidata-install/omnidata-setup

      Run Command

      sh omnidata_install.sh

      Whether to Check Return Code

      Yes

      Expected Return Code

      0

  5. Create configuration files for the OmniData service on the offload nodes (ceph1, ceph2, and ceph3).
    1. Select Command Execution and click Add a Task.

    2. Set task 8 parameters and click Save.
      Table 9 Task 9 parameters

      Parameter

      Description

      Task Name

      Creating configuration files

      Task Description

      Creating a configuration file for the OmniData service

      Run Directory

      /home/omm/omnidata-install/omnidata/deploy

      Run Command

      sh generate_config.sh -o Register_HOSTNAME:Register_Port -s Register_HOSTNAME:Register_Port

      Whether to Check Return Code

      Yes

      Expected Return Code

      0

    • If the big data engine is Spark, run sh generate_config.sh -s Register_IP:Register_Port.
    • The -s parameter in the script is used to specify the IP address/domain and port of the third-party registration service used by OmniData.
      • -s indicates Spark, which is used to specify the registration address of the Spark engine. The default IP address and port of the ZooKeeper node are 2181.
  6. Export the service flow and save it to your local PC. You can import service flows to quickly install other components.
    1. In the upper right corner of the displayed page, click Export.

      The window for selecting a path for saving the template is displayed.

    2. Name the template OmniData Installation and Configuration Workflow and click Save.

  7. Execute the service flow.

    Click Finish in the lower right corner and select Run Service. The procedure is complete when the execution is successful.

    • You can develop your own OmniData plugins and ensure the security of the plugins.
    • OmniData can have the Hive UDF plugin. You can add your own UDF plugin. To add a plugin, perform the following operations:
      1. Add a task similar to task 2 of step 2 in Installing OmniData, for example, copying the Hive UDF plugin.
      2. In task 5 of step 3 in Installing OmniData, add the unzip boostkit-omnidata-hive-udf-loader-1.6.1-1.4.0-aarch64.zip command to the beginning of the commands.
      3. In task 6 of step 3 in Installing OmniData, add the mv boostkit-omnidata-hive-udf-loader-1.6.1-1.4.0-aarch64 omnidata/plugin command to the beginning of the commands.