Future Posting Roadmap: 2007

Posting has been lighter than normal, as I have had the flu and it is footy finals time in Australia.

However I sat down the other day and thought about which areas I hope to cover over the next couple of months.
Given this is a blog and comments are on, if you want something covered which is not on the list feel free to add a comment with a request.



  1. Determine the IO/sec where the slave starts falling behind.
  2. Test mysqlslap running write workload on master and read workload on slave.
  3. Test running mysqlslap 3rd machine and connection to master and slave.
  4. Test running mysqlslap on replication with auto-increment column so show the effect of increment locking on master and slave.
  5. mysqlslap vs MySQL 5.1 row based replication

The end result is a performance report which answers the following questions:

  1. how many concurrent users causes the slave to fall behind
  2. When does network bandwidth become an issue. (test simple scp to get raw bandwidth)
  3. how does reading from the slave affect the slave’s ability to keep up.

MySQL Cluster:

  1. mysqlslap vs MySQL 5.1 Cluster
  2. running ndb Sizing script against normal database
  3. mysqlslap vs MySQL 5.1 Cluster Replication
  4. benchmarking and performance.

MySQL Toolkit and related:

  1. One article per mysqltoolkit tool.
  2. Article on using innotop


  1. Testing mysql partitioning, and maintenance tasks.
  2. How to design your database for your application (using load tables, federated storage)
  3. Using tmpdir as /dev/shm to help with temp table issues.
  4. Point in Time recovery with MySQL. Using slave (start slave until)
  5. Backups using LVM. => Done
  6. innodb_flush_method=O_DIRECT. Reason: see what the difference is, on a standalone instance and replication.


  1. performance
  2. clustering options for redundancy


  1. Using MySql
  2. using OracleXE
  3. using PostgreSQL
  4. using Hibernate

Have Fun