Installing and upgrading Netezza Performance Server Analytics on Netezza Performance Server for Cloud Pak for Data System

Learn how to install and upgrade Netezza Performance Server Analytics.

Before you begin

Ensure that you configured the nzpassword utility. For more information, see The nzpassword command.

Procedure

  1. Verify which INZA version you are on by running the following commands:
    $  source ~/.bashrc
    $  nzcm --inza
  2. Find the package in IBM Fix Central:
    1. In the Product selector field, type IBM Cloud Pak for Data System.
    2. From the Installed Version dropdown list, select IPSA_11.0 or INZA_11.2 based on the version that you want to install and click Continue.
    3. Select Browse for fixes to get a list of available INZA versions.
  3. Log in to the Netezza host, and create the directory to install the package in:
    [root@e1n1 ~]# docker exec -it ipshost1 bash
    [root@e1n1 /]# su - nz
    [nz@e1n1 ~]$ mkdir /nz/var/inza
  4. Download the package to this directory and unzip it:
    [nz@e1n1 inza]$ unzip ips-analytics-v11.0.1.0.zip
  5. Run the following command to install:
    [nz@e1n1 inza]$ ./inzaPackageInstaller.sh
    INZA package installation is executed by an interactive script that guides you through the process. When the installation starts, you are prompted to select the required options. For example:
    Do you want to install the Netezza Analytics packages? (y/n): y
    
    Confirm that the following packages should be installed.
     Netezza Analytics package: YES
    Do you want to install the above selections?
       Enter "y" to continue, "x" to exit
       or any other key to be prompted to modify your selection: y 
    
    Installing Netezza Analytics packages...
    Available compressed installation file or files:
      [0] /nz/var/inza/inza-11.2.27.zip
      [1] A compressed file in a different directory or with a non-standard name.
    
    Enter your selection: 0
    .
    .
    .
    Installing NZCM v.11.2.27...
    Do you want to reset the repository configuration? [y]/n y
    .
    .
    .
    Do you want to run the Netezza Analytics cartridge installer now? (y/n): y
    Do you want to do an Express (e) installation or a Custom (c) installation (e/c): e
    
    Confirm that the following components should be installed.
     1) MapReduce: YES
     2) Matrix: YES
     3) IBM Netezza In-database Analytics: YES
     4) Spatial: YES (ESRI)
    Do you want to install the selected components?
       Enter "y" to continue, "x" to exit
       or any other key to be prompted to change your selection: y
  6. After you complete the installation, verify its version by running:
    $  nzcm --inza
    Note: When you are upgrading INZA, and any dependencies are found, follow the recommended steps that are displayed by the installation script. Then, retry the installation. The dependencies, if present, must be backed up first by using nz_ddl* scripts in /nz/support/bin directory.
    Depending on the type of the dependency, take a backup by using:
    • For a function:
      nz_ddl_function  database name function name -schemas ALL > nz_ddl_function_database name
    • For a procedure:
      nz_ddl_procedure  database name function name -schemas ALL > nz_ddl_procedure_database name
    • For a table:
      nz_ddl_table   database name function name -schemas ALL   > nz_ddl_table_database name
    • For a view:
      nz_ddl_view  database name function name -schemas ALL  > nz_ddl_view_database name

    After the dependencies are backed up, you can drop dependencies from one session and restart the installation or resume installation from the current session from which installation started earlier.

    By default, the following INZA cartridges are installed and registered alongside your installation:
    $  nzcm --installed | wc -l  (64 cartridges installed)
    $  nzcm --registered | wc -l (39 cartridges registered)
    
    To view the complete list, run following commands:
    $ nzcm --installed
    $ nzcm --registered

What to do next

After INZA is upgraded, and if you have IBM Fluid Query installed, you must reregister connections and create functions as described in Recompiling UDXs and reregistering IBM Fluid Query.