Db2 for z/OS customization for the IBM zERT Network Analyzer task
About this task
The IBM® zERT Network Analyzer task stores and queries SMF data in a Db2 for z/OS database. Before you can use the task, this database must be created in a suitable Db2 for z/OS subsystem and the connectivity information for the database must be configured in the IBM zERT Network Analyzer.
IBM provides tooling and templates in the SYS1.SAMPLIB data set to help your local Db2 for z/OS database administrator (DBA) create the IBM zERT Network Analyzer database in a manner that conforms to local naming and resource conventions. Collectively, this tooling is called the IBM zERT Network Analyzer database schema tooling and it consists of the following members in SYS1.SAMPLIB:
- The IZUZNADT and IZUZNADA members contain Data Definition Language (DDL) templates for creating
the required database objects using two different approaches.
- IZUZNADT contains a template for creating the required database objects using a fixed schema name (SYSIBM_EZB_ZNADB) and fixed table names that the IBM zERT Network Analyzer depends on. If your local naming conventions allow for these fixed names, then you should use this template.
- IZUZNADA contains a template for creating the required database objects using your own customized schema and/or table names. It also creates aliases for those tables using the fixed schema name and table names that the IBM zERT Network Analyzer depends on. If your local naming conventions require you to change the schema name and/or the table names, you should use this template.
The templates contain all the appropriate DDL commands using variables for a wide variety of resource names and parameter values that the DBA may want to control. The DBA should set values for some or all these variables that will comply with the local Db2 for z/OS conventions and allocation strategies.
Note: The IZUZNADT and IZUZNADA prolog commentary explain each of the variables they include as well as the required variable syntax. - The IZUZNADI member of the SYS1.SAMPLIB data set. IZUZNADI is a sample variable substitution data set in which the DBA specifies the desired values for each of the template variables that are defined in the templates. The IZUZNADI sample specifies the default values for each of the variables, but the DBA can override any of the variables as needed.
- The IZUZNADG DDL generation REXX exec reads the specified template and a customized variable
substitution data set and generates a data set that contains a complete set of customized DDL for
creating a new IBM zERT Network Analyzer database or to update an existing database to the most
current schema level (for applying service).Note: Instructions for using the IZUZNADG exec are available by running the exec with the --HELP parameter.
After running the IZUZNADG exec, your DBA can use the resulting DDL data set in SPUFI or as input to a local JCL stream used to create the new IBM zERT Network Analyzer database in the Db2 for z/OS subsystem.
Procedure
-
Determine the local resource requirements and location for the IBM zERT Network Analyzer database.
Your DBA decides which Db2 for z/OS subsystem contains the IBM zERT Network Analyzer database objects as well as the specific Db2 for z/OS resources (storage pools, buffer pools, and so forth) to be allocated to these objects.
Tip: You should initially deploy the IBM zERT Network Analyzer database and service on a test system. The test system should be a place where you can easily get familiar with the operation of the service and can better understand the Db2 and system resource requirements when running queries against your imported SMF record data. Depending on the number of imported SMF records and the complexity of your queries, you might consider initially limiting query execution to specific times of day or specific systems to minimize system impacts.- Determine the maximum number of query reports that will be open at any given time
Your DBA needs to specify the number of partitions for a subset of tables in the IBM zERT Network Analyzer database called "Query Result Tables." These partitioned tables hold intermediate query results that are displayed through the network analyzer's Report tab. The IBM zERT Network Analyzer assigns one partition in each query result table to each active IBM zERT Network Analyzer query report for as long as that report is open in the web browser. Because of this, you must work with your DBA to determine an appropriate number of partitions to ensure that your database has enough partitions to support your community of IBM zERT Network Analyzer users.
The number of partitions is determined by the <QRTParts> variable in the DDL templates and is controlled by the DBA in the variable substitution data set used with the IZUZNADG exec (see the IZUZNADI member of SYS1.SAMPLIB).
To calculate the number of partitions you need, consider the number of users that will be using the IBM zERT Network Analyzer as well as the number of reports each user might have open at any given time (a single user can have multiple reports open at one time, with each one in its own web browser tab). Multiply those two numbers together to determine the maximum possible number of open reports. You should create at least that many partitions for each query result table. You might also want to increase this value by an appropriate percentage to ensure there is some room for growth over time.
The following equation summarizes the above calculation.NumOpenReports = MaxNumberOfUsers * MaxNumberOfReportsPerUser <QRTParts> = NumOpenReports + ( NumOpenReports * ExtraSpace% ) - Decide whether to use a separate database for the query result tables
Two variables in the DDL templates control whether the query result tables are created in the same database as the persistent IBM zERT Network Analyzer tables, or in
a different
database.- <database>
- specifies the name of the database that contains the IBM zERT Network Analyzer's persistent tables (all tables except the query result tables). When you are creating a brand new IBM zERT Network Analyzer database, set this variable to any valid database name. By default, the IZUZNADI sample sets this to 'ZNADB'.
- <QRTDatabase>
- specifies the name of the database that contains the query result tables. You may choose to store the query result tables in the same database as the persistent tables or in a different database. To use the same database, set the <QRTDatabase> and <database> variables to the same value. To use different databases, specify different values. By default, the IZUZNADI sample sets this to 'ZNAQRDB', placing the Query Result Tables in their own database.
- Determine how much table space to allocate in the database
The amount of Db2 for z/OS table storage required by IBM zERT Network Analyzer varies, but you can use the following guidelines to estimate the table storage required in your environment:
- Start with an allocation of 20 MB to hold the core security session data and operational data related to data management and user-built queries. If you are collecting zERT data from an unusually large number of unique security sessions across the z/OS systems, you might eventually need to add to this amount over time.
- Estimate the space that is required to store the maximum number of SMF Type 119 zERT Summary
(subtype 12) records that will be represented in the database at one time and add that amount to the
initial 20 MB. To do so, see details in Table 1.
Table 1. How to estimate the space for SMF Type 119 zERT Summary (subtype 12) records Step Task Formula Example 1 Estimate the number of unique security sessions that typically exist across the set of z/OS systems from which you are collecting zERT data. Use local procedures to estimate this value. Assume that 10,000 unique security sessions typically exist across all of the zERT-monitored systems. UniqueSessions = 10,000
2 Determine the maximum number of SMF intervals to be represented in the IBM zERT Network Analyzer database at a single time. MaxIntervals = ((1440 / SMFInterval) * Days)
where:
1440 = the number of minutes in a day;
SMFInterval = the SMF interval in minutes as defined in your SMFPRMxx parmlib member. Note that if you use different SMF intervals across the zERT-monitored z/OS systems, use the average interval length across the different systems here;
Days = the maximum number of days' worth of SMF data that is stored in the database at a single time.
Assume the average SMF interval is set to 20 minutes and you plan to store 30 days of SMF data in the zERT Network Analyzer database: MaxIntervals = (1440 / 20) * 30 = 72 * 30 = 2160
3 Determine how many SMF 119 subtype 12 records are collected over the Days value and imported into the database. MaxRecords = UniqueSessions * MaxIntervals MaxRecords = 10,000 * 2160 = 21,600,000 4 Determine how much DASD storage is required to store the data for the maximum number of records. Each SMF 119 subtype 12 record consumes about 500 bytes of DASD storage above and beyond the base space allocation of 20 MB. DASDSpaceMB = 20 + ((MaxRecords * 500) / 1,048,576) DASDSpaceMB = 20 + ((21,600,000 * 500) / 1,048,576) = 20 + 10,299 = 10,319MB or 10GB According to the example shown in the table, you need to allocate a total of 10 GB of space in the storage group that is used for your IBM zERT Network Analyzer database's table spaces.
- Determine the maximum number of query reports that will be open at any given time