Benchmark profile: Sysbench (InnoDB back-end): - prepare: 100K inserted rows - simple: 32 threads / 100K read-only selects - complex: 32 threads / 10K read-write transactions mysql-bench (MyISAM back-end): default settings
We can now turn our attention to the widespread MySQL database, using sysbench and mysql-bench. First, we prepare the database by inserting 100K rows into a newly created DB:
All filesystem are very quick to populate the database.
In simple (read only) workload, we can see no real differences.
Even the complex, read-write, transactional DB test results are quite similar between each other.
If sysbench used a InnoDB MySQL DB, with mysql-bench we use the read-oriented MyISAM as database engine:
Here BTRFS is significantly slower then EXT3 and EXT4. Even XFS is somewhat slower then the two EXT-based filesystems.