Presentation is loading. Please wait.

Presentation is loading. Please wait.

Introduction to Precise Indepth for DB2 UDB

Similar presentations


Presentation on theme: "Introduction to Precise Indepth for DB2 UDB"— Presentation transcript:

1 Introduction to Precise Indepth for DB2 UDB

2 What Is Precise/Indepth for DB2 UDB?
Precise/Indepth for IBM ® DB2 ® Universal Database™—hereafter referred to as Precise/Indepth for DB2 UDB—is a comprehensive performance management product for DB2 databases that addresses the needs of DBAs, system administrators, managers, quality assurance specialists and planners.

3 How Can Precise/Indepth for DB2 UDB Help You?
IT staff frequently ask questions like "Why did the application run so slowly last night?" or "Where is the application spending its time?". Precise/Indepth for DB2 UDB can help IT staff answer these questions.

4 How Does Precise/Indepth for DB2 UDB Work?
Precise/Indepth for DB2 UDB employs a client/server architecture to provide comprehensive 24 x 7 performance monitoring, vital for both short- and long-term monitoring of your DB2 database. The Precise/Indepth for DB2 UDB Collector runs on the server machine using a unique low overhead, comprehensive sampling technology. Performance data from your database is correlated and stored in flat history files also located on your server. At regular user defined intervals, data stored in the history files is loaded into the Performance Warehouse database. The Performance Warehouse can be used for long-term evaluations of performance.

5 Interpoint (SAP) Agent (Optional Component)
The Precise/Indepth for DB2 UDB agent software consists of the following: Listener Collector Batch Controller Interpoint (SAP) Agent (Optional Component) Symmetrix® Agent (Optional Component)

6 DB2/UDB OS Indepth for UDB Base Architecture Recent Activity Files
App & DB Snapshots System Calls Precise Client TCP/IP Listener Collector Batch Controller Recent Activity Files Performance Warehouse Summarized History Stored in DB2 Database

7 Precise/Indepth for DB2 UDB Workspaces
Precise/Indepth for DB2 UDB provides six workspaces that enable you to examine the performance of your DB2 databases. Current Activity Workspace Recent Activity Workspace Performance Warehouse Workspace Database Objects Workspace Database Statistics Workspace Statement Workshop Workspace

8

9 Current Activity Workspace
The Current Activity workspace displays almost real-time information on all sessions connected to your DB2 database in the last minute or time slice It can be used to monitor your system's current activities and resource consumption and to identify performance problems as they occur. For example, this workspace enables you to easily identify current locking problems and the sessions involved in the locks.

10 Recent Activity Workspace
The Recent Activity workspace displays detailed information on your DB2 database's short-term historical activity organized into time units. You can define the number of hours of database activity to be displayed in this workspace and the length of each time unit in the Options dialog box. Performance information displayed in this workspace can be used to identify and analyze the cause of a performance problem and is a prime source of input for future tuning decisions. For example, this workspace can help you identify the SQL statement that consumed the most resources in the last day, discover what caused the high consumption and find a solution to the problem.

11 Performance Warehouse Workspace
Performance Warehouse is an integrated feature of Precise/Indepth for DB2 UDB that collects and stores summarized information on database performance throughout the life cycle of the DB2 database. The Performance Warehouse workspace enables you to view this summarized information to assess database performance trends and analyze changes in database activities. Using this information, you can proactively tune your database objects and SQL statements to avoid future problems.

12 Database Objects Workspace
The Database Objects workspace provides a comprehensive browser that can be used to understand the relationships and associations between database schema objects and the SQL statements that access them. For example, this workspace enables you to identify all SQL statements that accessed a specific table.

13 Database Statistics Workspace
The Database Statistics workspace stores database level statistics as well as bufferpools‘ and tablespaces' I/O statistics. You can analyze short-term raw database statistics stored on the historical files or long-term summarized database statistics loaded into Performance Warehouse. You can easily identify the tablespace with most I/O requests, the worst response time per I/O request, the highest I/O wait time, or the worst hit ratio. You can get the same information on the bufferpool level and the database level. Database level statistics provide the following as well: Connections and agents statistics, lock and sort statistics, logging and archiving statistics, executed SQL statements and Units of Work statistics, catalog and package cache statistics and more.

14 Statement Workshop Workspace
The Statement Workshop workspace stores SQL statements in a logical cabinet and folder hierarchy. You can populate it with statements appearing in the Current Activity and Recent Activity workspaces or with manually written statements. You can examine stored statements to locate statements that meet your current requirements. You can also modify a statement and assess the impact of your modifications to the statement’s access plan.

15 Problem Detection and Identification
Precise/Indepth for DB2 UDB enables you to quickly identify your biggest performance problem areas by displaying detailed performance information in easy-to-understand graphs. You can view these graphs on the database, resource consumer and SQL statement level.

16 Viewing Performance Data on the Database Level
At the database level you can see if the database is CPU or I/O bound. You can also find out how much time is spent on different internal waits and how much time is spent as a result of OS resource shortages, locking problems, etc.

17 Viewing Performance Data on the Resource Consumer Level
Precise/Indepth for DB2 UDB enables you to drill down to identify which entities, e.g., users, programs and client machines, consumed the database resources.

18

19 Viewing Performance Data on the SQL Statement Level
After identifying a major resource consuming entity, you can further drill down to the statements executed by the entity. These statements may include long-running SQL statements as well as SQL statements that use few resources but are executed frequently. For each statement, Precise/Indepth for DB2 UDB enables you to access the following information: Entities that executed the statement Resources consumed by the statement to enable you to determine if the statement is CPU or I/O bound, waiting for a lock, etc. Resource consumption patterns to determine if statement execution is affected by excessive use of the same resource by another statement

20

21 Problem Analysis After identifying a problem in your database, you can use Precise/Indepth for DB2 UDB’s analysis features to focus on problematic SQL statements and/or database objects. By analyzing SQL statements and database objects, you can determine whether your performance problems are caused by poorly written SQL statements or by inefficient object design.

22 SQL Statement Analysis
Precise/Indepth for DB2 UDB presents the DB2 access path in detailed steps and displays statistics relevant to each step. It displays the original statement text as well as the “optimized” statement text generated by the DB2 Optimizer.

23 It also highlights the section of the SQL statement associated with each step and displays the database objects accessed by the SQL statement and the statistics collected on these objects. To further simplify viewing the DB2 access path, Precise/Indepth for DB2 UDB marks inefficient steps in the access path so you can easily identify and handle these steps.

24 Precise/Indepth for DB2 UDB uses the explain facility to recommend new indexes for a selected statement. The Recommend Indexes output contains the following two parts: Recommended virtual indexes - their key columns, estimated catalog statistics and a sample of “create index” DDL text. A virtual access plan - as if the recommended virtual indexes actually exist. The virtual access plan is displayed with the same information and formatting as a real access plan.

25

26 Precise/Indepth for DB2 UDB enables you to evaluate a virtual index.
A virtual index means you do not actually have to create an index. You can explain all SQL statements as if the virtual index exists. All you need to do, is to select the desired index key columns and you will get a list of all affected SQL statements. For each affected SQL statement, you will get: Its estimated cost with and without the evaluated index, the improvement percentage and a comparison of the access plan with and without the evaluated index.

27

28 Database Object Analysis
In many cases, you may want to tune a database object rather than a specific statement that accesses the database object. Especially, when the application was not written by you (e.g., ERP and CRM applications) and you cannot change the text of any statements. Precise/Indepth for DB2 UDB uses a unique algorithm to break the total “in DB2 time” of each SQL statement to its accessed objects (i.e., tables and indexes). The contribution of each object to all SQL statements is aggregated on the object and hour level and stored in Performance Warehouse. Precise/Indepth for DB2 UDB enables you to identify top-accessed objects in terms of the total “in DB2 time” that the application spent accessing that object. The top-accessed tables are the best candidates for a design review and/or a creation of new indexes.

29

30 Precise/Indepth for DB2 UDB enables you to filter date range, hour groups and time resolution.
You can drill down from the object level to SQL statements. You can easily identify top-resource consuming SQL statements that access a selected table. You can get all SQL statements that use each of the indexes, which are built on the selected table as well. When triggers are defined on a table, you can drill down from a trigger to all DML statements that execute the trigger event. The drill-down from an index to all SQL statements that use the selected index can tell you whether the selected index is used or not.

31

32 Tablespace and Bufferpool Analysis
Precise/Indepth for DB2 UDB continuously collects tablespace and bufferpool I/O performance data and saves this data in Performance Warehouse. This enables you to identify the tablespace that is accessed most often or which has the worst response time. It also enables you to identify the tablespace or bufferpool with the worst hit ratios. Other collected metrics include I/O counts, I/O time, response time and hit ratios over time and table and index usage over time. You can also access a list of the top accessed tables and indexes sorted by total In DB2 time or I/O time.

33 Buffer Pool I/O Sataistics

34 Buffer Pool Hit Ratios

35 Usage Scenario - Tuning Current Locking Problems

36

37

38

39

40

41

42

43

44

45

46

47

48

49 Questions


Download ppt "Introduction to Precise Indepth for DB2 UDB"

Similar presentations


Ads by Google