PRODUKTE FALLSTUDIEN LEISTUNGEN JOBS KONTAKT IMPRESSUM MEETINGS
DE  |  ENGLISH       
DB2 - Produkte System / z/OS Anwendungsentwicklung Rechenzentrum

 
ULT for DB2 z/OS -  The DB2 Log Tracker

Benefits
  • Let DBAs quickly and easily identify, isolate and restore unwanted changes
  • Help auditors locate updates to sensitive data
  • Provide information about those updates, including who made them and when
Features
  • Identifies log records by various attributes as transaction, user, group, job
  • Presents DB2 log data decompressed
  • Prints DB2 log data and reports
  • Generates Undo and Redo SQL

What users say about ULT:
   ... boosts performance and lowers costs ...
   ... it is easier to handle and executes faster ...
   ... it saved us money and time ...
   ... getting updates made to certain tables requires to submit a single job.

ULT for DB2 on z/OS is a program that analyzes DB2 Archived Logs. It was created in order to assist in auditing changes and in the cumbersome task of backing out erroneous updates. ULT reports data modifications to a set of specified tables within a given period of time. Output is provided as binary before and after images of changed rows.

DB2's logging facility is optimized for recovery purposes, but it’s condensed record layout is difficult to analyze and to map. A DSN1LOGP report demonstrates this problem; the data printed does not contain complete records, some columns are in DB2 internal format, sometimes one SQL statement causes up to three DB2 log records, and even worse, the data may be compressed, and therefore unreadable without additional programming effort. It is obviously a real challenge to isolate, reconstruct and output comprehensive data changes from DB2's Archived Logs.

It is also rather difficult to identify log records for specific transactions, e.g. for all transactions of a user, a group of users or for a group of jobs.

ULT makes it easy to utilize DB2 log data for auditing or correction purposes. The user simply specifies what is be extracted in terms of
  • the table/s to be monitored,
  • start and end time of the log analysis,
  • scope rules i.e. selection criteria for the 'units of recovery' to be considered.

ULT table selection allows DB2 LIKE patterns:
DTL.DTLTAB
DTL.%ABC
DTL%.MONTH__

The first line selects a specific table DTLTAB with creator DTL. The second line selects all tables of creator DTL, with names ending with ABC. The third line selects all tables having names beginning with MONTH and with two additional characters, whose creator begin with the letters DTL.

The following attributes may be used as scope rules:

USER DB2 authorization ID,
[USER = HUGO, USER = SYS*]
CORRID DB2 Correlation ID,
PLAN DB2 Plan name,
[PLAN = MYAPP]
CONNID DB2 Connection ID
CONNTYPE DB2 Connection Type.

Based on user specifications ULT provides clear and easily understandable log data records for the specified transaction/s and time period.

Workflow
A ULT execution cycle consists of three steps, which may be executed either as separate jobs or combined within one job: The first step extracts the table definitions of the tables to be monitored from DB2 Catalog, the second step extracts the relevant records from DB2's log datasets, and the third step formats the extracted log records.


Data Capture Changes Option
In order to save resources and to accelerate the monitor process tables to be monitored by ULT should be defined with the 'Data Capture Changes Option' (DCC). DCC is either invoked by the DATA CAPTURE CHANGES clause of CREATE TABLE or for existing tables by an ALTER TABLE ... DATA CAPTURE CHANGES. When tables are defined with the DCC, the entire 'before image' of an updated row is captured on the log. This additional information can represent an increase in log data compared to tables that are not defined with the DATA CAPTURE CHANGES option. However, for tables which don't have DCC switched on ULT calculates the impact of DCC, i.e.  reports the increase in log data that would occur if DATA CAPTURE CHANGES is invoked. Usually the increase is around or below 1%.

Compressed Tablespaces
ULT fully supports compressed tablespaces and is able to expand compressed records. For REORGs of compressed tablespaces the KEEPDICTIONARY clause should be specified. This enables ULT to extract and report data changes that took place before the REORG as well as after it. ULT receives the compression dictionary from the tablespace partitions, hence it always works with the last or most recent version of the dictionary. Hence REORGs executed within ULT monitor period which do not keep the dictionary can cause unpredictable results.

ULT Output
  • ULT reports who changed which record at what time.
  • ULT outputs 'undo' and 'redo' data which is applicable as input for data correction after execution of faulty application programs.
Output is obtained in the following formats:
As sequential dataset, variable length record format, all fields converted,  easy to use for COBOL or other application programs.
As Sequential dataset, fixed length record format, SQL text, all fields converted to printable characters, the text is ready to use for dynamic SQL processors (SPUFI, DSNTEP2).
ULT for DB2 z/OS
 
Go to Top