MySQLConf2009MondayNotes
[edit] Monday, April 20
[edit] 8:30 am - 12 noon
(3h 30 min, 1 break)
[edit] Camp: "Hackfest"
Mark Callaghan. Pick a MySQL feature to add or extend (or a bug to fix) and Mark will walk you through the important basics and help one or more groups hack away!
[edit] Build and Release Management for Database Engineers
Gregory Haase (Bill Me Later, Inc.)
[edit] Practical MySQL Plugin Development
Roland Bouman (XCDSQL Solutions / Strukton Rail), Sergei Golubchik (MySQL)
[edit] MySQL Cluster Tutorial
Stewart Smith (Sun Microsystems)
[edit] Replication Tutorial
Mats Kindahl (MySQL), Lars Thalmann (MySQL)
[edit] How to Avoid the Three Stooges of Data Warehousing - Practical Lessons for Open Source BI Implementations
David Lutz (Infobright)
[edit] MySQL DBA Certification Tutorial, Part 1
Kai Voigt (Sun Microsystems)
[edit] The Revised Memcached Tutorial
Alan Kasindorf (Six Apart), Brian Aker (Sun Microsystems)
[edit] MySQL benchmarking and Monitoring tools
Tom Hanlon (MySQL)
[edit] 1:30 pm - 5:00 pm
(3h 30 min, 1 break)
[edit] Camp: "Ask a Guru"
Drop-in sessions. Come with your questions, and leading MySQL experts will provide help free of charge. Or just come to listen.
[edit] Partitioning in MySQL 5.1
Giuseppe Maxia (Sun Microsystems Inc), Sarah Sproehnle (MySQL)
[edit] Introduction to Data Modeling with MySQL Workbench
Michael G. Zinner (MySQL), Max Mether (MySQL)
[edit] Scale Up, Scale Out, and High Availability: Solutions and Combinations
MC Brown (Sun Microsystems)
CPU load -- scaling up won't help if the DBs aren't loaded
RAM -- always add more
- cache rows
- cache indexes
- check key_reads
- check key_buffer_size
- key_reads / key_read_requests = miss rate
- is mySQL set to use all available memory?
Disks -- spread the load
- RAID
Partitioning -
- splitting the data across disks
- resolves disk contention
Multiple instances
- if you have lots of CPU, this may be the solution
- use virtualization (solaris containers?)
Scaling out
- Reducing read/write load
- More machines
- Spread the data
- duplicate data
- segregate/partition data
Replication
- Difficult to seed
- Synchronization
- Lag
- Statement and row-based
- Chaining replication -- saves network interface on master machine
Clustering
- Management node
- SQL node
- Data nodes
- Synchronous distributed writes
- Needs lots of network bandwidth
- Mostly for high-read, low-write
- Add more nodes for complicated queries
- Add data nodes for more storage
- NDB (wtf?) is the engine
- Limits on column types
Memcache
- IS THE SOLUTION TO ABSOLUTELY EVERYTHING
- ALL HAIL MEMCACHE
DRBD
- Device Level replication
- Replicates actual blocks of a disk, only information that has really changed
- Very efficient use of network
- Easy synchronization
- Mainly for High Availability
ZFS
- We were not interested
Sharding
- SHARDING IS EVEN MORE THE SOLUTION TO EVERYTHING
- SHARDING WILL HAVE YOUR BABIES
[edit] MapReduce + SQL Using Open Source Software and Tools
Christophe Bisciglia (Cloudera, Inc), Jeff Hammerbacher (Cloudera, Inc), Aaron Kimball (Cloudera, Inc), Tom White (Cloudera, Inc)
[edit] MySQL DBA Certification Tutorial, Part 2
Kai Voigt (Sun Microsystems)
[edit] SQL Antipatterns Strike Back
Bill Karwin (Karwin Software Solutions)
The database is the foundation of any innovative project, but you can’t innovate if you’re making the same mistakes of yesteryear.
Bill Karwin has supported SQL developers since 1994. Some blunders occur so frequently that we call them antipatterns: brittle, buggy solutions and create security or scalability risks in your projects. In this tutorial, Bill describes SQL techniques that should be called antipatterns, illustrates their weaknesses, and offers better solutions.
[edit] Understanding How MySQL Works by Understanding Metadata
Sheeri K. Cabral (The Pythian Group), Patrick Galbraith (Lycos Inc.) PDF of slides
[edit] Blog posts
[edit] "MySQL Conference 2009 Daily summary: Monday"
- Shlomi Noach