Procedures to validate the correctness of a Merge/Modify Accum Change and Accum Log functions

Products

• Merge/Modify for DB2 for z/OS

Issue/Introduction

What are recommended procedures to validate the correctness of Merge/Modify for Db2 for z/OS (PMM)
Accumulate Change and Accumlate Log functions. 

Environment

Release : 20.0
Component : Merge/Modify for Db2 for z/OS

Resolution

Testing procedures for the Change Accumulation and Log Accumulation components of Merge/Modify (PMM).

During the log apply phase, PMM validates thoroughly that the log record makes sense given the current state
of the data page. If it does not, the PFR0997E message is returned and the process terminates.

The following describes the most used testing scenarios

Change Accumulation
The standard change accumulation (ACCUMULATE CHANGES syntax) test case scheme we use is the following:

Create a testing tablespace
Load tablespace with initial data using data generator
Create a full image copy of the tablespace with IBM COPY
Execute testing DML activity (INSERT, UPDATE, DELETE) on the testing tablespace
Execute PMM to create a change accumulation dataset for the testing object
Execute testing DML activity (INSERT, UPDATE, DELETE) on the testing tablespace
Run the following recoveries to the current point:
IBM Recover to recover the data to the current point. SELECT all data from the object to dataset A.
Fast Recover for Db2 for z/OS (PFR) with USE-CHANGE-ACCUM YES. SELECT all data from the object to dataset B.
Fast Recover for Db2 for z/OS (PFR) with USE-CHANGE-ACCUM NO. SELECT all data from the object to dataset C
Compare datasets A, B and C to verify that they hold the same data.

Many instances of the test case scheme are executed, varying mostly the following:

Tablespace types (segmented, partitioned, PBR, PBG)
Page sizes, RBA formats, compression
DML activity types
Db2 Log record versions (DB2 12, DB2 13)
Output of the change accumulation dataset to different devices – tape, disk

Log Accumulation
The standard change accumulation (ACCUMULATE syntax) test case scheme we use is the following:

Create a testing tablespace
Load tablespace with initial data using data generator
Create a full image copy of the tablespace with IBM COPY
Execute testing DML activity (INSERT, UPDATE, DELETE) on the testing tablespace
Execute PMM to create a change accumulation dataset for the testing object  
Execute testing DML activity (INSERT, UPDATE, DELETE) on the testing tablespace
Execute PMM to create a change accumulation dataset for the testing object  
Execute testing DML activity (INSERT, UPDATE, DELETE) on the testing tablespace
Execute PMM to create a change accumulation dataset for the testing object  
Execute testing DML activity (INSERT, UPDATE, DELETE) on the testing tablespace
Create a full image copy C1 using Log Accumulation
Create a full image copy C2 using IBM COPY
Run the following recoveries:
Recover with IBM Recover or PFR using the C1 copy. SELECT all data from the tablespace to dataset A
Recover with IBM Recover or PFR using the C2 copy. SELECT all data from the tablespace do dataset B
Compare datasets A and B to verify that they hold the same data
Use IBM Unload to unload data from the full image copy created with log accumulation

Many instances of the test case scheme are executed, varying mostly the following:

Tablespace types (segmented, partitioned, PBR, PBG)
Page sizes, RBA formats, compression
DML activity types
Db2 Log record versions (DB2 12, DB2 13)
Image copy target devices – tape, disk

Additional Information

Suggested Uses for Change Accumulation

Creating Image Copies Using Log Accumulation