3 Database Installation Tasks

3.1 Overview

This chapter describes how to install the Oracle Database software. These installation instructions apply to the Oracle Database products.

Before you install the software, you must perform the necessary preinstallation tasks, including transferring the software to your z/OS system and extracting the files. For more information, refer to Chapter 2, "Preinstallation Tasks."

Unlike previous releases, Oracle Universal Installer is not used to install the software. The software is installed by executing z/OS batch jobs. The batch jobs use the standard utility programs, the batch TSO TMP, and the batch UNIX shell BPXBATCH to install the software. Oracle recommends that you use the interactive tools, such as, ISPF and SDSF to edit, submit, and monitor the batch jobs.

This documents assumes that the user performing the installation has the tools to create and edit jobs, submit them for execution, and view the results or listings.

3.2 Installing the Oracle Database Software

Perform the following tasks to install the software:

1. Start the installation by setting up the JCL for JOB1.

   The purpose of JOB1 is to generate the JCL for other jobs. Several configuration parameters are specified in JOB1 to select the correct set of software and to specify the location for the software on the target z/OS system. These parameters are located at the beginning of the JCL for JOB1. You must customize the job contents for these parameters according to your requirements using a text editor.

2. Use any of the following methods to obtain the JCL to customize the parameters for JOB1:

   • Open the Release Notes document (readme.html) in a Web browser on your computer. Copy the section marked JOB1 and create the file job1.jcl. You can either transfer the file to your z/OS system to begin editing or edit the file directly on your computer and then transfer the file to your z/OS system. A bulk transfer technique such as FTP, NFS, or IND$FILE must be used since JOB1 has over 500 lines of data. Since the data is large, it is not practical to cut and paste the data in a terminal emulator screen by screen.

   • On your z/OS system, locate the EBCDIC copy of the Release Notes document (readme-edbcic.html) that was created by unzipping the distribution files. Copy the readme-edbcic.html file into a new text file. Retain only the lines between the markers "### start of JOB1" and "### end of JOB1". Do not include the markers "### start of JOB1" and "### end of JOB1".

3. Customize the JOB statements:

   1. Replace the JOB statement at line 1 in the job1.jcl file with a JOB statement that conforms to the requirements of your z/OS system. Also, provide any job-level statements and job-level JES statements. In general, JOB1 requires minimal resources. However, since JOB1 runs the batch TMP and shell, ensure that any REGION specifications are generous.

   2. Replace the JOB statement located after the line "//JOBCARD DD DATA,DLM='!!'" near line 57 in the job1.jcl file with an appropriate JOB statement for your system. Also, provide any job-level statements and job-level JES statements in this section. This JOB statement is used for each of the jobs generated by JOB1. Since these jobs perform the actual installation, ensure that you provide generous resource parameters.

      JOB1 only generates the jobs. You still have an opportunity to tailor each job before submission.

4. Customize the installation parameters.

   The installation parameters start near line 4 and are similar to the following lines:

   //   SET SSET=A
   //   SET OHFS='/oracle/v10205a'
   //   SET OHLQ=ORACLE.V10G
   //   SET OMLQ=V10205A
   //   SET THLQ=SCOTT.TEMP
   //   SET TZFS='/scratch/scott'
   //*
   //   SET PERMLOC='MGMTCLAS=MCLASS3,STORCLAS=SCLASS3,DATACLAS=DCLASS3'
   //   SET TEMPLOC='UNIT=3390,VOL=SER=TEMPV1'
   //*
   //   SET ZIP1='/scratch/ftp/v10205a_mvs_1of3.zip'
   //   SET UZIP='/usr/local/bin/unzip'
   //   SET JJAR='/usr/lpp/java150_31bit/J5.0/bin/jar'
   

   where:

   SSET is the software set. The value A specifies both server software and client software. The value C specifies only the client software.

   OHFS is the HFS directory for ORACLE_HOME.

   OHLQ is the high-level-qualifier for data sets that correspond to ORACLE_HOME.

   OMLQ is the mid-level-qualifier for data sets that correspond to ORACLE_HOME.

   THLQ is the high-level-qualifier for temporary (installation only) data sets.

   TZFS is the HFS directory for temporary files created in HFS

   PERMLOC and TEMPLOC specify the DD keyword parameters to be used for the allocation of permanent and temporary data sets, respectively. You can use SMS classes or UNIT/VOL keywords for either parameter.

   ZIP1 is the absolute path name to the first (1of3) distribution zip file.

   UZIP is the absolute path name to the unzip command. If there is no unzip command available, then specify UZIP=''.

   JJAR is the absolute path name to the Java jar command. If there is no Java jar command available, then specify JJAR=''.

   The pre-supplied values are just examples. You must edit these values to meet your requirements.

   If you have questions about how these parameters are used, you can examine the lines of JOB1 to find where these parameters (symbols) are used (are substituted). In the JCL you will see &OHFS, &OHLQ, &OMLQ, &THLQ, &TZFS, &PERMLOC, and &TEMPLOC. In shell scripts you will see $OHFS, $OHLQ, $OMLQ, $THLQ, $TZFS, $PERMLOC, $TEMPLOC, $ZIP1, $UZIP, and $JJAR. In JCL, the limitation of 72 character line determines the length of the parameter values.

5. Run JOB1.

   This creates the PDS named &THLQ..V10205A.WRKLIB containing the other jobs.The members are named JOB2, JOB3, JOB4, and so on.

   1. JOB2 handles unzipping the distribution files. This job creates 11 HFS files with names ending in XMIT.

   2. JOB3 copies the HFS XMIT files to XMIT data sets. This creates 11 data sets with names ending in XMIT. These data sets were originally created with the TSO XMIT command.

   3. JOB4 allocates the production data sets and a temporary PAX data set.

   4. JOB5 loads each production data set from its counterpart XMIT data set. If you need to rerun JOB5, then you must rerun JOB4 first.

      The AUTHLOAD PDSE will contain only ORACLE and ORAAMSD modules if you are performing both server and client software installation (that is, if SSET=A). If you are performing only the client software installation (that is, if SSET=C), then the AUTHLOAD PDSE will contain only the ORAAMSD module.

   5. JOB6 creates the Oracle home directory in HFS from the PAX data set.

   6. If additional jobs are created, refer to Oracle Database Release Notes for IBM z/OS on System z.

   7. All steps of JOB1 should terminate with RC=0.

6. Run the generated jobs, JOB2, JOB3, JOB4, and so on in the sequential order.

   In general, the jobs are structured in a way that enables you to rerun or modify the jobs if you encounter problems. In some instances, you may need to go back further when you choose to rerun a job. For example, if you need to rerun JOB5, then you must rerun JOB4 first.

   Step DEL3 of JOB3 and step DEL4 of JOB4 may complete with RC=4 when both JOBS are executed for the first time. This can be safely ignored.

3.3 Required Postinstallation Tasks

You must perform the tasks described in the following sections after completing an installation:

• Downloading and Installing Patches

• Performing System Administrator Tasks

• Configuring Other Oracle Products