"Charting the Course... Teradata Database Administration Course Summary

Transcription

1 Course Summary Description This class teaches everything a Teradata DBA needs to know. This class covers everything from Teradata DBA fundamentals, TASM and Viewpoint, and Teradata Internals. Topics Introduction and Good Advice Teradata Architecture Fundamentals the DBA must know The Primary Index is the Axis of all Teradata Systems A DBA s best friend - The Data Dictionary How Teradata Tracks Objects Creating Users and Databases Profiles Roles Access Rights Collect Statistics Locking Protection Features Viewpoint TASM Designing and Implementing Workloads Top SQL Commands Cheat Sheet The Cold, Hard Teradata Facts How Teradata Tracks Objects AMP Worker Tasks Deep Dive Overhead for each Row Compression Data Stored in the Row How Data Rows are Stored in Blocks Disk Cylinders and the Master Index Teradata Virtual Storage (TVS) Teradata Writes and Blocks Access Logging DBQL Query Logging ResUsage Audience This course is designed for Teradata DBAs or anyone who has interest in learning more about how to perform the role of a Teradata DBA. Prerequisites There are no prerequisites for this course. Duration Three to five days

2 Course Outline I. Introduction and Good Advice A. What is Parallel Processing? B. Start Small and Think Big C. Give your Enterprise the Tools they need D. Model the Business with ERwin E. Educate the Business on the Business by Sharing the Model F. Load Your Models and have the SQL Built Automatically G. Five Brilliant Pieces of Teradata (1 of 5) is MPP H. Five Brilliant Pieces (2 of 5) are Tactical Queries I. Five Brilliant Pieces (3 of 5) Is a Traffic System J. Five Brilliant Pieces (4 of 5) Is Viewpoint K. Five Brilliant Pieces (5 of 5) Are Data Processing Options L. Support Large Queries, but Monitor them closely M. Experiment and Improve Loading Data Strategies N. Compress Your Data with Multi-Value Compression O. Separate your Production System from Your Test System II. Teradata Architecture Fundamentals the DBA must know A. Parallel Architecture B. The Teradata Architecture C. All Teradata Tables are spread across ALL AMPS D. Teradata Systems can Add AMPs for Linear Scalability E. AMPs and Parsing Engines (PE s) live inside SMP Nodes F. Each Node is Attached via a Network to a Disk Farm G. Two SMP Nodes Connected Become One MPP System H. There are Many Nodes in a Teradata Cabinet I. This is the Visual You Want to Understand Teradata J. Responsibilities of the DBA III. The Primary Index is the Axis of all Teradata Systems A. The Primary Index is defined when the table is CREATED B. A Unique Primary Index (UPI) C. Primary Index in the WHERE Clause - Single-AMP Retrieve D. A Non-Unique Primary Index (NUPI) E. Primary Index in the WHERE Clause - Single-AMP Retrieve F. A conceptual example of a Multi-Column Primary Index G. Primary Index in the WHERE Clause - Single-AMP Retrieve H. A conceptual example of a Table with NO PRIMARY INDEX I. A Full Table Scan is likely on a table with NO Primary Index J. Table CREATE Examples with four different Primary Indexes K. What happens when you forget the Primary Index? L. Why create a table with No Primary Index (NoPI)? IV. A DBA s best friend - The Data Dictionary A. The Data Dictionary Resides in User DBC B. The DBC.DBCInfoV View C. Querying the Data Dictionary D. Using the Keyword USER E. Restricted Views have an X at the End of their Name F. The V is New with Teradata V12 G. The V and the Restricted X are Now Often Combined H. A Recap of What We Have Learned So Far I. The DBC.DatabasesV View J. The DBC.Users View K. The DBC.Tables View L. Using DBC.Tables to find out about Fallback M. The DBC.Indices View N. The DBC.Columns View O. Clever Queries for the DBC.ColumnsV View

3 P. New V14 - The DBC.PartitioningConstraintsV View Q. The DBC.AccountInfo View R. The DBC.AMPUsage View S. Clearing Out the DBC.AMPUsage Data T. The DBC.AllTempTables U. The DBC.Triggers V. The DBC.All_RI_ChildrenV W. DBC.SessionInfoV Information X. DBC.LogonOffV Y. AllRoleRights, AllRightsV, UserRightsV and UserGrantedRightsV Z. The DBC.Profiles View AA. RoleMembers, RoleInfo, UserRoleRights and ProfileInfoVX, BB. Understanding that Space is based on a Per-AMP Basis CC. Total Space for a Single Database or User DD. Using the Data Dictionary to see the Space for Everyone EE. Finding the Perm Percent Used FF. Finding the Perm Percent Used with a HAVING Clause GG. Finding the Perm Percent Left with a HAVING Clause HH. Creating a Macro for Perm Percent Used with a Dynamic % II. Orphaned Spool Files That Need to be deleted JJ. Finding Table Sizes KK. Finding Skew in the Tables in a Database LL. Finding Skew in a Table MM. Display the Distribution of a Column per AMP NN. Your Users and Databases OO. DBC Tables used in the Collect Statistics Process PP. The DBC Table DBC.Next QQ. DBA Advice - ClearPeakDisk to Reset Peak Space RR. DBA Advice Clean out these Tables Periodically SS. The DBC.AssociationV View TT. The DBC.JournalsV View UU. DBC.Databases2V is for Unresolved Reference Constraints VV. The DBC.All_RI_ChildrenV for Inconsistent RI WW. The DBC.ShowColChecksV View XX. The DBC.ShowTblChecksV View YY. The DBC.PartitioningConstraintsV View ZZ. The DBC.AccessLogV View AAA. The DBC.AccessLogV View for Today s Queries BBB. The DBC.AccessLogV View Denials for Today CCC. DBC.DBQLRulesV DDD. DBC.QryLogV EEE. DBC.QryLogSummaryV FFF. ResUsage Macros GGG. Executing the ResUsage Macro DBC.Resnode HHH. The DBC.IdCol Table V. How Teradata Tracks Objects A. Teradata Assigns each Object a Unique Numeric ID B. The Table ID C. The Table ID in Greater Detail D. Looking at the TableID inside the actual Cylinders E. A More Detailed View of TableID inside the actual Cylinders F. The Blocks Below are All Associated with the Same Table G. Bits, Bytes and More H. Cylinder Sizes VI. Creating Users and Databases A. Creating Users and Databases B. Password Security Meanings C. Now we have Two Users in the Teradata System D. A Grant Statement so others Create a Database or User E. And so the Teradata Hierarchy Begins F. Creating a Database G. Users are Given Passwords While Database are Not

4 H. Teradata Administrator Can CREATE Users I. The Modify User Statement J. A Clever Way to Reset a User Password K. Accounts and their Associated Priorities L. Creating a User with Multiple Account Priorities M. Self-Nicing to change Account Priorities N. Account String Expansion (ASE) O. The DBC.AccountInfo View P. The DBC.AMPUsage View Q. Account String Expansion (ASE) in Action R. Test Run queries Under All Accounts for TeraTom S. The DBC.AMPUsage View VII. Profiles A. Profiles B. Getting Started for Profile Creation C. Creating A Profile and a User D. Password Security E. Password Security Meanings F. Creating A Profile and then Modifying a User G. Quiz What are the Profile Values? H. Answer to Quiz What are the Profile Values? I. Quiz What are the Profile Values After Null? J. Answer to Quiz What Are the Profile Values After Null? K. The DBC.ProfilesVX View L. The DBC.ProfilesV View M. The DBC.AccountInfoVX View N. ProfileInfoVX, RoleMembers, RoleInfo and UserRoleRights O. Teradata Administrator Can CREATE Profiles (1 of 2) P. Teradata Administrator Can CREATE Profiles (2 of 2) Q. Dropping a Profile R. The Effects of Dropping a Profile VIII. Roles A. Roles B. Getting Started for Role Creation C. Create A Role and then Assign that Role It s Access Rights D. Create a User and Assign them a Default Role E. A Role vs. a Profile F. Granting a Role to a Current User G. Active Roles H. Setting Your Active Role to ALL I. Roles and Valid Objects J. Roles and Invalid Commands K. Nesting of Roles L. Nesting of Roles in Action (1 of 3) M. Nesting of Roles in Action (2 of 3) N. Nesting of Roles in Action (3 of 3) O. Quiz What Databases Does Mandy Have Access To? P. Answer What Databases Does Mandy Have Access To? Q. GRANT WITH ADMIN OPTION Command R. REVOKE ADMIN OPTION FOR Command S. RoleMembers, RoleInfo, UserRoleRights and ProfileInfoVX, T. DBC Tables for AllRoleRights, AllRightsV, UserRightsV and UserGrantedRightsV IX. Access Rights A. The Objects That Require Access Rights B. Objects and Available Access Rights C. A Few Examples to Get You Started D. There are Three Types of Access Rights E. There are Three Types of Access Rights F. There are Three Types of Access Rights G. A Dinner Invitation of Access Rights H. One of the Problems with Access Rights I. The Rights for SysDBA and TeraTom J. The GRANT Statement K. Create A Role and then Assign that Role It s Access Rights L. GRANT to PUBLIC M. GRANT To ALL DBC N. GRANT Using the ALL Keyword O. GRANT Database Strategy for Users, Views and Tables P. Inheriting Access Rights Q. GRANT at the Column Level

5 R. GRANT for the Ability to CREATE Secondary Indexes S. Access Rights to CREATE Triggers T. The REVOKE Command U. DBC Tables for AllRoleRights, AllRightsV, UserRightsV and UserGrantedRightsV V. The GIVE Statement W. A DROP User can be Better than a GIVE Statement X. Removing a Level in the Teradata Hierarchy X. Collect Statistics A. The Teradata Parsing Engine (Optimizer) is Cost Based B. The Purpose of Collect Statistics C. When Teradata Collects Statistics it creates a Histogram D. The Interval of the Collect Statistics Histogram E. Histogram Quiz F. Answers to Histogram Quiz G. What to COLLECT STATISTICS On? H. Why Collect Statistics? I. How do you know if Statistics were collected on a Table? J. A Huge Hint that No Statistics Have Been Collected K. The Basic Syntax for COLLECT STATISTICS L. COLLECT STATISTICS Examples for a better Understanding M. The New Teradata V14 Way to Collect Statistics N. COLLECT STATISTICS Directly From another Table O. Where Does Teradata Keep the Collected Statistics? P. The Official Syntax for COLLECT STATISTICS Q. How to Recollect STATISTICS on a Table R. Teradata Always Does a Random AMP Sample S. Random Sample is Kept in the Table Header in FSG Cache T. Multiple Random AMP Samplings U. How a Random AMP gets a Table Row count V. Random AMP Estimates for NUSI Secondary Indexes W. USI Random AMP Samples are Not Considered X. There s No Random AMP Estimate for Non- Indexed Columns Y. A Summary of the PE Plan if No Statistics Were Collected Z. Stale Statistics Detection and Extrapolation AA. Extrapolation for Future Dates BB. How to Copy a Table with Data and the Statistics CC. How to Copy a Table with NO Data and the Statistics DD. When to COLLECT STATISTICS Using only a SAMPLE EE. Examples of COLLECT STATISTICS Using only a SAMPLE FF. Examples of COLLECT STATISTICS for V14 GG. How to Collect Statistics on a PPI Table on the Partition HH. Teradata V12 and V13 Statistics Enhancements II. Teradata V14 Statistics Enhancements JJ. Teradata V14 Summary Statistics KK. Teradata V14 MaxValueLength LL. Teradata V14 MaxIntervals MM. Teradata V14 Sample N Percent NN. Teradata V14.10 Statistics Collection Improvements OO. Teradata V14.10 Statistics Collection Threshold Examples PP. Teradata V14.10 AutoStats feature QQ. Teradata Statistics Wizard XI. Locking A. The Four Major Locks of Teradata B. The Read Lock C. The Read Lock and Joins D. The Write Lock E. The Exclusive Lock

6 F. The Three Levels of Locking G. Locking at the Row Hash Level H. Locking at the Table Level I. Locking at the Database Level J. The Ongoing Battle between Read and Write Locks K. Compatibility between Read Locks L. Why Read Locks Wait on Write Locks M. Why Write Locks Wait on Read Locks N. The Access Lock is Different from the Other Locks O. What is the Purpose of an Access Lock? P. Locking Modifiers - Locking Row, Table or Database Q. All Views should consider the Locking for Access Statement R. What is a Dead Lock or a Deadly Embrace? S. Pseudo Tables are designed to minimize Dead Locks T. Pseudo Tables are referenced in the Explain Plan U. Incompatible Locks Wait on each Other V. The Checksum Lock of Teradata W. The Nowait Option for Locking X. The Automatic Locking for Access Button inside Nexus Y. Viewpoint Lock Viewer Z. Viewpoint Lock Viewer Lets You Configure Your View AA. What is a Host Utility (HUT) Lock? XII. Protection Features A. A List of the Protection Features B. Transient Journal Protects the Transaction Integrity C. The Transient Journal in Action D. A Single Transaction could Involve All AMPs E. The Secret to turning off the Transient Journal F. The Transient Journal s Write Ahead Logging (WAL) G. A Node with 40 AMPs and 40 Dedicated FSG Caches H. The Transient Journal s Write Ahead Logging (WAL) I. Working Example of the Write Ahead Log (WAL) J. The First Step in our Example of the Write Ahead Log (WAL) K. The Second Step in our Example of the Write Ahead Log L. The Third Step in our Example of the Write Ahead Log M. The Fourth Step in our Example of the Write Ahead Log N. The Last Step in our Example of the Write Ahead Log O. Fallback to Protect against an AMP Failure P. Fallback Clusters Q. AMPs in a Cluster are Physically Separated R. The Reason AMPs in a Cluster are Physically Separated S. The Price you pay for Fallback T. How to Create a Table with Fallback U. How to Create a Table with No Fallback V. How to Alter a Table to Add or Drop Fallback W. What is a Virtual Disk? X. Why do AMPs each have Four Physical Disks? Y. Is a Mirror just like Looking into a Mirror? Z. RAID 1 Mirroring Redundant Array of Independent Disks AA. What does RAID Protect? BB. How Does RAID Fail? CC. Do RAID and Fallback have a Connection? DD. What is a Clique? EE. If a Node goes down the AMPs migrate within the Clique? FF. Does Teradata Reset during a Node Failure? GG. Four Node Cliques HH. Migrating AMPs in Four Node Cliques II. The Hot Spare Node JJ. The Hot Spare Node in Action KK. With a Hot Spare a Second Teradata Reset isn t Needed LL. A Node, It s AMPs and their Disks

7 MM. How Cliques are Physically DefinedCliques are cabled so Migrating AMPs can access their Disks NN. A Review of Fallback and Clusters OO. An Example of Fallback and Clusters PP. Quiz 1 How Many Clusters do you see? Quiz 1 Answer How Many Clusters do you see? QQ. Quiz 2 How Many Cliques do you see? RR. Quiz 2 Answer How Many Cliques do you see? SS. Quiz 3 What have we lost? Multiple Choice Answer TT. Quiz 3 Answer What have we lost? Multiple Choice Answer UU. Quiz 4 What have we lost? Multiple Choice Answer VV. Quiz 4 Answer What have we lost? Multiple Choice Answer WW. Quiz 5 What have we lost? Which Answer is False? XX. Quiz 5 Answer What have we lost? Which Answer is False? YY. Quiz 6 What have we lost? Pick Two True Answers ZZ. Quiz 6 Answer What have we lost? Pick Two True Answers AAA. Summary of the facts for Fallback, Clusters, and Cliques BBB. Quiz 7 How Many Virtual Disks (Vdisks) are in this System? CCC. Quiz 7 Answer How Many Virtual Disks are in this System? DDD. Quiz 8 How Many Physical Disks are in this System? EEE. Quiz 8 Answer How Many Physical Disks are in this System? FFF. Quiz 9 How Many Transient Journals in this System? GGG. Quiz 9 Answer How Many Transient Journals in this System? HHH. Quiz 10 How Many Transient Journals are Open? III. Quiz 10 Answer How Many Transient Journals are Open? JJJ. Quiz 11 How Much Space? KKK. Quiz 11 Answers How Much Space? LLL. Quiz 12 How Much Space with Fallback? MMM. Quiz 12 Answers How Much Space with Fallback? NNN. Quiz 13 How Many Disks could we lose with RAID 1? OOO. Quiz 13 Answer How Many Disks could we lose? PPP. Quiz 14 How Many Disk losses could Kill Us? QQQ. Quiz 14 Answer How Many Disk losses could Kill Us? RRR. Quiz 15 How Many AMPs could we lose if Lucky? SSS. Quiz 15 Answer How Many AMPs could we lose if Lucky? TTT. Quiz 16 How Many AMPs could we lose if Unlucky? UUU. Quiz 16 Answer How Many AMPs could we lose Unlucky? VVV. The Permanent Journal WWW. Difference between the Transient and the Permanent Journal XXX. Difference Between the Before and After Permanent Journal YYY. Full System Backup compared to an After Journal ZZZ. How Full System Backups work with the After Journal AAAA. The Many Different Permanent Journal Options BBBB. Where is the Permanent Journal Stored? CCCC. Using Common Sense about Journal Locations DDDD. After Journals are Never stored in the Same Node or Clique EEEE. What is a Dual After Journal? FFFF. What is a Dual Before Journal? GGGG. What is a Journal?

8 HHHH. Creating a Table with Fallback and a Before and After Journal IIII. Does Fallback Affect a Permanent Journal? JJJJ. Permanent Journal Rules KKKK. Example 1: Permanent Journal Scenarios to Test the Rules LLLL. Example 2: Permanent Journal Scenarios to Test the Rules MMMM. Example 3: Permanent Journal Scenarios to Test the Rules NNNN. How to Create Database with a Permanent Journal OOOO. Creating Tables under different Journal Circumstances PPPP. Permanent Journal s Three Main Areas QQQQ. The Current Journal consists of the Active and Saved Areas RRRR. Permanent Journal Commands SSSS. Deleting a Permanent Journal TTTT. Some Great Advice for Maintaining the Permanent Journals UUUU. Recovery Using the Permanent Journals VVVV. The Journals View in DBC (DBC.Journals) WWWW. Archive Recovery Console (ARC) XXXX. Reasons You Might Utilize ARC YYYY. ARC raising the BAR (Backup Archive Restore) ZZZZ. ARC Commands in Alphabetical Order AAAAA. An ARC Example of an Archive and then a Restore XIII. Viewpoint A. Database Administration (DBA) on Teradata B. Viewpoint C. Logging into Viewpoint D. The First Time you Login to Viewpoint E. The Add Content Menu for Monitoring F. The Add Content Menu for TASM G. The Add Content Menu for Tools H. The Add Content Menu for Trend Reporting I. Adding your first Content J. How the Page looks after you add your first Content K. Adding Additional Pages L. The Fundamentals of Viewpoint Pages and Portals M. Adding Multiple Portals to a Single Page N. All Portals to Their Individual Tab O. Portlet Controls P. The Rewind Control Q. Query Monitor Overview R. Query Monitor Details View S. Query Monitor Actions T. My Queries U. Calendar V. Calendar Details W. Add an Event to the Calendar X. Getting a Weekly View Y. System Health Z. System Health Drilldown AA. System Health Preferences BB. ALERT CC. Setting up an ALERT Configuration DD. Setting up an ALERT Configuration Continued EE. SQL Scratchpad FF. SQL Scratchpad Query Results GG. SQL Scratchpad Object Loader HH. Space Usage II. Space Usage Preferences JJ. Investigating Space to See Dynamic Reports KK. Space by Vproc (AMP) LL. How to Obtain a Detailed View of Space MM. Detailed View of Space NN. Adding PERM Space to a Database or User OO. External Content PP. Lock Viewer QQ. Lock Viewer Lets You Configure Your View RR. Canary Queries SS. Multiple Systems Need Multiple Canary Query Portlets TT. What Canary Queries Measure UU. Remote Console VV. Capacity Heatmap

9 WW. Capacity Heatmap Preferences XX. Capacity Heatmap Example YY. Metrics Analysis ZZ. My Queries AAA. My Queries Preferences BBB. Today's Statistics Using System Metrics CCC. Today's Statistics Using Query Metrics DDD. Today's Statistics Using Performance Metrics EEE. Today's Statistics Preferences for System Metrics FFF. Today's Statistics Preferences for Query Metrics GGG. Today's Statistics Preferences for Performance Metrics HHH. Today's Statistics Preferences for Sampling Intervals XIV. TASM A. Three Levels of Workload Management B. Pre-execution, Query Execution, and Postexecution C. What is TASM? D. Query Management compared to Workload Management E. What is the Secret Sauce for Query Management? F. The life of a Query G. What is a Workload? H. Workload Examples I. There are Four Types of Query Rules J. Common Sense Examples of Filters and Throttles K. Performance Period Examples L. The Scoop on Object Throttles M. Load Utility Throttles N. Creating Workloads O. When Creating Workloads the WHO is your Foundation P. After the WHO comes the WHERE Q. After the WHO and the WHERE comes the WHAT R. Exception Actions S. When and How Teradata checks for Exceptions T. DBC.TDWMExceptionLog U. Teradata Workload Analyzer V. Teradata Workload Analyzer W. Pre-execution, Query Execution, and Postexecution X. Why use Priority Scheduler? Y. The Concept of a Resource Partition Z. Resource Partitions AA. The Clever Idea behind Resource Partitioning BB. The Brilliant Idea behind Resource Partitioning CC. The Concept of Resource Partitions and Weights? DD. The Concept of a Workload in a Resource Partition EE. Calculating your CPU Percentage 1 FF. Answers to Calculating your CPU Percentage 1 GG. Calculating your CPU Percentage 2 HH. Answers to Calculating your CPU Percentage 2 II. Calculating your CPU Percentage 3 JJ. Answers to Calculating your CPU Percentage 3 KK. Calculating your CPU Percentage 4 LL. Answers to Calculating your CPU Percentage 4 MM. Calculating your CPU Percentage 5 NN. Answers to Calculating your CPU Percentage 5 OO. Calculating your CPU Percentage 6 PP. Answers to Calculating your CPU Percentage 6 QQ. Calculating your CPU Percentage 7 RR. Answers to Calculating your CPU Percentage 7 SS. Calculating your CPU Percentage 8 TT. Answers to Calculating your CPU Percentage 8 UU. Calculating your CPU Percentage 9

10 VV. Answers to Calculating your CPU Percentage 9 WW. Calculating your CPU Percentage 10 XX. Answers to Calculating your CPU Percentage 10 YY. Calculating your CPU Percentage 11 ZZ. Answers to Calculating your CPU Percentage 11 AAA. Calculating your CPU Percentage 12 BBB. Answers to Calculating your CPU Percentage 12 CCC. Calculating your CPU Percentage 13 DDD. Answers to Calculating your CPU Percentage 13 XV. Designing and Implementing Workloads A. How to Configure Priority Scheduler B. Workload Designer C. The Three Areas of the Workload Designer D. How the Area of Workload Designer are Used E. How to Create a Ruleset F. The First Step to Creating a Ruleset G. Returning to the Workload Designer Main Screen H. A Basic Ruleset in the Working Area I. Editing a Ruleset J. The States K. States Adding Another Planned Environment L. Editing the Name of the NewEnv Planned Environment M. Changing the Planned Environment Name to a Good Name N. Adding a New State O. A New State is Born P. Editing the Default State Name of newstate Q. We Now have a New State named DayState R. Drag a State (DayState) to a Planned Environment (WorkDay) S. Drag a State (DayState) to a Planned Environment (WorkDay) T. Do NOT Forget to Hit Save U. We now Need to Create a Planned Event V. Create a Planned Event to Define the Timeframes W. The Planned Event Screen X. Your Event is now in the Available Events Tab Y. Drag and Drop your Event to your Planned Environment Z. You have a Timeframe for the Planned Environment DayState AA. Creating a Degraded Condition with our Normal Condition BB. Creating a Degraded Condition with our Normal Condition CC. Editing your New Condition to become a Degraded Condition DD. Edit Health Conditions Screen EE. Our New Edited Health Condition named Degraded FF. We Now have a New Degraded Condition GG. Don t Forget to Now Hit the Save Button HH. Let s Define our Degraded Health when CPU hits 100% II. The Available Events Screen JJ. The Create Event Screen KK. The Available Events Screen Now Shows our Event LL. After we Drag and Drop the Event into the Degraded Events MM. Drag the Base State to the Degraded/Always Environment NN. Our Default of BASE is Set in the Degraded Environment OO. Let s Create a New State for our Degraded Workday PP. Let s Create a New State for our Degraded Workday QQ. Let s Create a New State for our Degraded Workday RR. Drag the new state DayDegraded to our Planned Environment SS. Hit Save to keep your Changes TT. Limiting the Sessions

11 UU. When you First come to the Sessions Screen VV. Query Session Limit The General Tab WW. Creating a Session Limit for the WorkDay when DayDegraded XX. Creating a Session Limit for the WorkDay when DayDegraded YY. A Successful Creation of a Session Limit on DayDegraded ZZ. Creating Filters AAA. When you First come to the Filters Screen BBB. Creating a New Filter The General Tab CCC. Creating a New Filter The Classification Tab DDD. Edit Query Characteristics Criteria EEE. Creating a New Filter The General Tab FFF. Edit the Filter for the WorkDay when DayDegraded GGG. Creating a New Filter The Classification Tab HHH. Utilizing Throttles III. Throttles The Throttles Tab JJJ. Throttles New Throttles Tab KKK. Throttles New Throttles Tab LLL. Edit the Throttle for the WorkDay when DayDegraded MMM. Throttles New Throttles Tab NNN. Throttles The Throttles Tab OOO. Let s Go Back and Activate our Rule Set PPP. To Prepare for Activation, Move the Ruleset to the Ready QQQ. The Ruleset is in the Ready and now we can Activate RRR. The Ruleset is in the Ready and now we can Activate SSS. When you First come to the Workloads Screen TTT. Workloads New Workload General Tab UUU. Workloads New Workload Classification Tab VVV. Edit the Request Source Criteria WWW. Workloads New Workload Classification Tab XXX. Workloads Throttles in a Workload YYY. Edit the DayDegraded Settings ZZZ. Workloads Throttles are now Set for the State DayDegraded AAAA. Workloads Service Level Goals BBBB. Service Level Goals for the WorkDay CCCC. Workloads Service Level Goals DDDD. Workloads Exceptions EEEE. Workloads The Exceptions Screen FFFF. Workloads The New Exceptions Screen XVI. The Cold, Hard Teradata Facts A. What is Parallel Processing? B. The Basics of a Single Computer C. Teradata Parallel Processes Data D. Parallel Architecture E. The Teradata Architecture F. All Teradata Tables are spread across ALL AMPS G. Teradata Systems can Add AMPs for Linear Scalability H. Understand that Teradata can scale to incredible size I. AMPs and Parsing Engines (PEs) live inside SMP Nodes J. Each Node is attached via a Network to a Disk Farm K. Two SMP Nodes Connected Become One MPP System L. There are Many Nodes in a Teradata Cabinet M. Inside a Teradata Node N. The Boardless BYNET and the Physical BYNET O. The Parsing Engine P. The AMPs Responsibilities Q. Teradata Parallel Processing R. Each Table has a Primary Index that is Unique or Non-Unique S. The Hash Map Determines which AMP will own the Row

12 T. A Unique Primary Index Spreads the Data Evenly U. The AMP Adds a Uniqueness Value to Create the Row-ID V. Each AMP Sorts Their Tables by the Row-ID W. A Non-Unique Primary Index Skews the Data X. Comparing the Same Table with Different Primary Indexes Y. Unique Primary Index Queries are a Single AMP Retrieve Z. A Non-Unique Primary Index is also a Single AMP Retrieve AA. Teradata has a No Primary Index Table called a NoPI Table BB. There are Normal Tables and then there are Partitioned Tables CC. A Visual of One Year of Data with Range_N Per Month DD. Partitioning is designed to eliminate the Full Table Scan EE. A Partition # and Row-ID = Row Key FF. An AMP Stores its Rows Sorted in only Two Different Ways GG. AMPs Moves Their Data Blocks into Memory to Read/Write HH. The Most Taxing thing for an AMP is Moving Blocks into Memory II. Rows are Stored in Data Blocks which are stored in Cylinders JJ. Rows for an AMP Stored Inside a Data Block in a Cylinder KK. An AMP s Master Index is Used to Find the Right Cylinder LL. The Row Reference Array (RRA) Does the Binary Search MM. A Block Splits into Two Blocks at Maximum Block Size NN. Data Blocks Maximum Block Size has Changed (V14.10) OO. The New Block Split with Teradata V14.10 PP. The Block Split with Even More Detail in Teradata V14.10 QQ. There is One Master Index and Thousands of Cylinder Indexes RR. Each Table has a 48-bit TableID XVII. How Teradata Tracks Objects A. Teradata Assigns each Object a Unique Numeric ID B. The Table ID C. The Table ID in Greater Detail D. Looking at the TableID inside the actual Cylinders E. A More Detailed View of TableID inside the actual Cylinders F. The Blocks Below are All Associated with the Same Table G. Bits, Bytes, and More H. Cylinder Sizes XVIII. AMP Worker Tasks A. Teradata is a Message Passing System B. The Parsing Engine Parses the SQL and comes up with a Plan C. What is an AMP Worker Task (AWT)? D. Each AMP has 80 AMP Worker Tasks (AWTs) E. Each Query Takes Up One or More AMP Worker Tasks F. An All-AMP Query Usually Won t Use More Than 4 AWTs G. A Live Example of AWTs in Action H. There are 24 AWTs Reserved for Internal Work I. How Utilities Use AWTs J. Monitoring AMP Worker Tasks with ResAMPCpuByGroup XIX. Deep Dive Overhead for each Row A. Why Go Deep inside the Overhead of a Row? B. A Row Layout in Teradata C. Row Length D. Row ID E. How The Row Hash is created for Each Row

13 F. Unique Primary Indexes have Even Distribution G. The AMP adds a Uniqueness Value to Its Rows H. The Row-Hash is 32-bits and so is the Uniqueness Value I. Non-Unique Primary Indexes have Skewed Data J. Flag Byte K. Presence Byte L. Presence Byte is used to show Null Values in each Row M. A Close-up look at the Presence Byte for Nulls N. An Extreme example to look at the Presence Byte for Nulls O. Quiz How Many Presence Bits used for these Columns? P. Answer How Many Presence Bits used for these Columns? Q. Quiz How Many Presence Bits used with NOT NULL? R. Answer How Many Presence Bits used with NOT NULL? S. Quiz Which bit will be Set to a One? T. Answer Which bit will be Set to a One? U. Quiz How Many Presence Bits Needed Now? V. Answer How Many Presence Bits Needed Now? W. What Happens when we need more than One Presence Byte? X. An Example that must use a 2nd Presence Byte Y. An Example that must use a 2nd Presence Byte Z. Quiz Answer the Presence Bit/Byte Questions? AA. Answers to the Answer the Presence Bit/Byte Questions? BB. Quiz - How Many Nullable Columns are Possible? CC. Answer- How Many Nullable Columns are Possible? XX. Compression A. Important Information about Compression B. Presence Bytes are also used for Compression C. Why One Byte (8 bits) can represent up to 255 Values D. Answers to One Byte (8 bits) can represent up to 255 Values E. Now that you Understand that 8 Bits can Represent F. A Compression Example that Compresses Two Values G. A Compression Example that Compresses Three Values H. Quiz Name that Compression Value I. The Next Important Concept in Compression J. uiz Can you Fill in the Compression Values? K. Answer Can you Fill in the Compression Values? L. The Last Major Concept in Compression M. Quiz Using One Presence Byte for Multiple Columns N. Answer Using One Presence Byte for Multiple Columns O. Quiz How Many Presence Bytes are Needed? P. Answer How Many Presence Bytes are Needed? Q. Advanced Quiz Fill in the Presence Bits? R. Answer to Advanced Quiz Fill in the Presence Bits? S. The Cost Vs. the Savings T. The Cost List of Compression U. A Deeper Dive Into NULL Values V. Quiz - How Much Space Did We Just Save? W. Answer - How Much Space Did We Just Save? X. Advanced Quiz - How Much Space Did We Just Save? Y. Advanced Quiz - How Much Space Did We Just Save?

14 Z. Using the DBC Tables in a Compression Experiment AA. A Compression Test BB. A Compression Test CC. We then moved all Eight Tables to another Database DD. Compression Reports with Nexus and SmartCompress EE. We Then Created Two Global Temporary Tables FF. We Then Created and Executed our Macro GG. Report Comparing Compressed and NonCompressed Tables XXI. Data Stored in the Row A. The Varchar Offset B. An Example of a Varchar Offset C. An Example of Two Varchar Columns D. The Fixed Columns E. An Example with Multiple Fixed Columns F. Compressible Columns G. An Example with Fixed Columns and A Compression Column H. An Example with A Fixed Column and a Compressed Varchar I. VARCHAR Columns J. An Example of a Fixed Column and a Varchar K. Teradata s Maximum Row Size XXII. How Data Rows are Stored in Blocks A. Why Go Deep inside Data Blocks? B. In The Beginning a Table is created C. Every AMP has the Exact Same Tables D. Rows are Stored in Blocks E. Each Table Header and Data Block have the Same TableID F. AMPs Moves Their Data Blocks into Memory to Read/Write G. AMPs can Read/Write their Rows once they are in FSG Cache H. Every Data Block Starts with a Data Block Header I. Every Data Block Ends with a Data Block Trailer J. Each Block has a Row Reference Array (RRA) K. The Row Reference Array (RRA) is in Descending Order L. A Binary Search is always done through the RRA M. A Binary Search is a quick Search among thousands of Rows N. The Ref Array Pointer in the Row Layout in Teradata O. How Blocks of Data Begin in Teradata P. How Blocks of Data Grow in Teradata Q. Did You Notice the Row Reference Array (RRA)? R. A Great Picture of a Single AMP s Data Block with Details S. Data Blocks Grow until they Reach Maximum Block Size T. The Block Split U. The Block Split with Even More Detail V. The Block Split Showing Two Blocks with Greater Detail W. Blocks Continue to Split as Tables Grow Larger X. Reminder - Data Blocks Maximum Block Size has Changed (V14.10) Y. Reminder - The New Block Split with Teradata V14.10 Z. Reminder - The Block Split with Even More Detail in Teradata V14.10 AA. Reminder - Teradata V14.10 Block Split Defaults BB. FYI Some Advanced Information about Data Block Headers XXIII. Disk Cylinders and the Master Index A. Disks have Cylinders which hold Data Blocks B. Rows are Stored in Data Blocks which are stored in Cylinders C. A Real World View of Rows inside a Data Block in a Cylinder D. A Top down View of Cylinders E. There are Hot, Warm, and Cold Cylinders

15 Cylinders are used for Perm, Spool, Temp, and Journals F. Synchronized Scan (Sync Scan) G. EXPLAIN Using a Synchronized Scan H. Intelligent Memory (Teradata V14.10) I. Teradata V14.10 Intelligent Memory Gives Data a Temperature J. Data deemed VeryHot stays in each AMP's Intelligent Memory K. Intelligent Memory Stays in Memory L. Each AMP has Their Own Master Index M. Each Cylinder on an AMP has a Cylinder Index N. Quiz What Two Things Does and AMP Read? O. Answer What Two Things Does and AMP Read? P. Quiz How Many Row Reference Arrays do you See? Q. Answer How Many Row Reference Arrays do you See? R. Quiz How Many Row Reference Arrays are there Now? S. Answer How Many Row Reference Arrays do you See? T. Quiz How Many Row Reference Arrays in Total? U. Answer How Many Row Reference Arrays in Total? V. Quiz How Many Cylinder Indexes are Here? W. Answer How Many Cylinder Indexes are Here? X. A More Detailed Illustration of the Master Index Y. A Real-World View of the Master Index Z. An Even More Realistic View of an AMP s Master Index AA. The Cylinder Index BB. An Even More Realistic View of a Cylinder Index CC. How a Query using the Primary Index works DD. How the AMPs Do a Full Table Scan EE. How An AMP Reads Using a Primary Index FF. Teradata Assigns each Object a Unique Numeric ID GG. The Table ID HH. The Table ID in Greater Detail II. Looking at the TableID inside the actual Cylinders JJ. A More Detailed View of TableID inside the actual Cylinders KK. An Even More Realistic View of a Cylinder Index LL. Bits, Bytes, and More MM. Cylinder Sizes NN. How TVS Monitors and Migrates Tables OO. How TVS Monitors and Migrates Partitioned (PPI) Tables PP. A Summary of the Master and Cylinder Index XXIV. Teradata Virtual Storage (TVS) A. Solid State Drives (SSD) Vs. Hard Disk Drives (HDD) B. Teradata Uses Two Types of Disks C. Traditional Teradata Without Teradata Virtual Storage (TVS) D. Teradata With TVS in a Conceptual Diagram E. What TVS is Responsible for Doing F. The Benefits of Teradata Virtual Storage (TVS) G. What is a Clique? H. If a Node goes down the AMPs migrate within the Clique? I. Review of a Clique J. Teradata Virtual Storage (TVS) Manages within a Clique K. Before TVS vs. TVS Today with Teradata V13.10 L. TVS Operates in Two Different Modes M. TVS Knows the Disks and Which Cylinders are the Fastest N. A Concept Called Recency O. Data Placement and Migration P. Review - Intelligent Memory (Teradata V14.10) Q. Review - Teradata V14.10 Intelligent Memory Gives Data a Temperature

16 R. Review - Data deemed VeryHot stays in each AMP's Intelligent Memory S. Review - Intelligent Memory Stays in Memory XXV. Teradata Writes and Blocks A. A Teradata Write B. Teradata Insert (Option 1 of 3) has enough space for the Insert C. Teradata Insert (Option 2 of 3) is a Defragment to make Space D. Teradata Insert (Option 3 of 3) is to Get a Bigger Block E. Checksum Determines if a New Block is Needed F. A Reminder of How Rows are Sorted with Block Utilities G. A Reminder of How Rows are Sorted with SQL Inserts H. When a Block Reaches Maximum Size, it Splits into Two I. A Block Split Always Sorts the Rows Perfectly Once Again J. In Teradata V14.10 the Maximum Block Size is 1 Megabyte K. Cylinder Sizes L. Teradata V14 Large Cylinders M. Quiz on Block Split N. Answer - Quiz on Block Split O. Quiz How many Items are in the Picture? P. Answer to Quiz How many Items are in the Picture? Q. Quiz on Teradata V14.10 Block Split R. Answer - Quiz on Block Split S. Blocking Terms for Teradata V14 and Below T. Blocking Terms for Teradata V14.10 and Beyond U. Block Sizes and Filling of Cylinders V. Space Fragmentation (1 of 2) W. Space Fragmentation (2 of 2) X. What is a Defrag? (1 of 2) Y. What is a Defrag? (2 of 2) Z. What Happens When a Cylinder is Full? AA. What is a Mini-Cylpack? (1 of 2) BB. What is a Mini-Cylpack? (2 of 2) CC. What is a Mini-Cylpack Vs. a Pack Disk? DD. A Pack Disk Picture EE. New Teradata Auto Cylinder Pack Feature FF. A Pack Disk Honors the Free Space Percent GG. Free Space Percent HH. Three Examples of the Free Space Percent II. Simpler Terms for our Free Space Percent Examples JJ. The Free Space Percent can be set in Three Ways KK. Two Table Create Examples of Min and Max Block Size LL. The Same Table Creates using KILOBYTES (KBYTES) MM. Why Would I Want Bigger or Smaller Block Sizes? NN. How Does Teradata Manage Space? OO. How Can Many Big Blocks Become Many Small Blocks? PP. Merge Datablocks (13.10 Feature) QQ. Merge Datablocks Details RR. Setting Merge Datablocks in DBS Control or at Table Level SS. How Have Customers Previously Handled Shrinking Blocks? XXVI. Access Logging A. Access Logging B. Security for the DBA C. The Tables and Views Associated with Access Logging D. Begin Logging Options E. Begin Logging, View Rules, See Log Data, and End Logging F. Begin Logging Examples G. The DBC.AccessLogV View H. The DBC.AccessLogV View for Today s Queries I. The DBC.AccessLogV View Denials For Today J. Controlling the Log Files

17 XXVII. DBQL Query Logging A. DBQL Query Logging B. The Tables and Views Associated with DBQL C. There are Seven Major Tables to Store DBQL Entries D. The Views for the Major DBQL Tables E. Begin Query Logging Default Information F. Begin Query Logging WITH Options G. Begin Query Logging Examples H. Begin Query Logging LIMIT Options I. Begin Query Logging LIMIT Examples J. SUMMARY and THRESHOLD have Additional Options K. Begin Query Logging with Additional Options Examples L. Begin and End Query Logging Examples M. Replace Query Logging Statement N. An Inside Look at the View DBC.DBQLRulesV O. The Columns in the View DBC.DBQLRulesV P. Begin Logging, View Rules, See Log Data and End Logging Q. DBC.DBQLRulesV R. DBC.QryLogV S. DBC.QryLogSummaryV T. Chapter 28 ResUsage U. ResUsage V. Major Tables to Store ResUsage Entries W. The ResUsage Views X. ResUsage Macro Information Y. ResUsage Macros Z. Executing the ResUsage Macro DBC.Resnode AA. DBC.Resnode Major Column Explanation BB. ResAMPCpuByGroup CC. ResCpuByAMP DD. ResCpuByGroup EE. ResCpuByNode FF. ResCpuByPE GG. ResHostByGroup HH. ResHostByLink II. ResHostTotal JJ. ResHostTotalDay KK. ResHostTotalHour LL. ResIvprMigrate MM. ResIvprMigrateHour NN. ResLdvByGroup OO. ResLdvByNode PP. ResMemByGroup QQ. ResMemMgmtByNode RR. ResNetByGroup SS. ResNetByNode TT. ResPeCpuByGroup UU. ResPeCpuByGroup VV. ResScpuDayTotal WW. ResScpuSec XX. ResSvprDetailReadTotal YY. ResSvprPreReadBySec ZZ. ResSvprQLenAvg AAA. ResSvprQLenAvgByVproc BBB. ResSvprQLenAvgByVproc CCC. ResSvprQLenMaxHour DDD. ResSvprReadByVprocSec EEE. ResSvprReadByVprocSec FFF. ResSvprReadTotal GGG. ResSvprReadTotal HHH. ResSvprWriteTotal III. ResSvprWriteTotalHour JJJ. ResSyncScan