Oracle Partitioning and Parallel Execution for Performance

Contents

1 Seminar overview
2 Seminar duration and details
3 You will learn the following...
4 Seminar contents
5 Next seminar dates

Seminar overview

Oracle Database’s partitioning options can be used for help in three areas: VLDB information lifecycle manageability, data availability and performance. This seminar will mostly cover partitioning for performance, how to design partitioning for performance and how to write/troubleshoot SQL which should use partitioning for performance.

The second part of this section will cover parallel execution internals, tuning and troubleshooting, both in partitioned and non-partitioned environments and we also touch the topic of configuring parallel execution right on RAC.

Seminar duration and details

Seminar duration: 4 x 4 hour deep dives (4 half-days in a single week)
Venue: Online, virtual seminar at your desk (no travel!)
Audience: Senior DBAs, senior developers, database designers
Skill level: Intermediate to Advanced
Speaker: Tanel Poder
Price: 1196 USD per attendee

Check the next seminar dates and sign up here:

http://blog.tanelpoder.com/seminar/

You will learn the following...

Best practices for partitioning (in different circumstances) and how to avoid common pitfalls
How to control partition pruning and measure whether desired partition pruning is happening
How to achieve partition-wise joins for joining large datasets
Things to be aware of when performing partition exchange loads
Understand how the parallel execution engine and row distribution works
Best practices for parallelism (in different circumstances) and how to avoid common pitfalls
How to troubleshoot parallel execution plans and find where the bottleneck is

Note that this seminar does not aim to cover every single different partitioning combination and teach the syntax of every partition maintenance command. You should already know what partitioning is and how to create partitioned tables. However you will learn the fundamental principles and internals behind partitioning and the Oracle features using it, so you will be able to decide what is best for your application and dataset yourself.

In this seminar we are not going to go deep into Exadata features as not everyone is using Exadata. So there will be a different class for that, which focuses solely on Exadata specific performance and design issues (but it is still based on the fundamentals and internals learned from this Partitioning and Parallel Execution class).

Seminar contents

1) Designing applications and databases with partitioning in mind

Designing for partition pruning
Designing for partition-wise joins
Choosing between local vs. global indexes
Designing for massive data loads
Using partition exchange loads
Partitioned objects, CBO and optimizer statistics collection

2) Troubleshooting partition pruning and partition exchange issues

Measuring which exact partitions are visited during the SQL execution
Why isn’t the desired partition pruning happening?
What causes the ORA-8103: object no longer exists and ORA-1410: invalid ROWID errors when exchanging partitions?

3) Understanding data flow in parallel execution plans

Understanding Data Flow Operators (DFO) and producer-consumer hierarchy
Monitoring long-running parallel queries
Monitoring data distribution and table queues

4) Achieving optimal data distribution in parallel execution plans

Avoiding data broadcasts and many-to-many distribution
Combining multiple rowsource operators onto a single level in parallel execution hierarchy
Parallel partition-wise joins
Best PX distribution for partitioned tables
Parallel execution considerations in RAC

5) Advanced Parallel Execution Troubleshooting

Tracing parallel execution and gathering detailed parallel execution stats
Monitoring IO throughput and detecting IO subsystem saturation

Next seminar dates

Check the next seminar dates and sign up here!

http://blog.tanelpoder.com/seminar/

Comments

Recent site activity

Copyright