This blog post is part two in what is now a continuing series on the Star Schema Benchmark. In my previous blog post I compared MySQL 5.5.30 to MySQL 5.6.10, both with default settings using only the InnoDB storage engine. In my testing I discovered that innodb_old_blocks_time had an effect on performance of the benchmark. There was [...]
MySQL and Percona Server in LinkBench benchmark
May 8, 2013 By 17 Comments
Around month ago Facebook has announced the Linkbench benchmark that models the social graph OLTP workload. Sources, along with a very nice description of how to setup and run this benchmark, can be found here. We decided to run this benchmark for MySQL Server 5.5.30, 5.6.11 and Percona Server 5.5.30 and check how these servers [...]
Benchmarking Percona Server TokuDB vs InnoDB
May 7, 2013 By 32 Comments
After compiling Percona Server with TokuDB, of course I wanted to compare InnoDB performance vs TokuDB. I have a particular workload I’m interested in testing – it is an insert-intensive workload (which is TokuDB’s strong suit) with some roll-up aggregation, which should produce updates in-place (I will use INSERT .. ON DUPLICATE KEY UPDATE statements [...]
Rotating MySQL slow logs safely
April 18, 2013 By 7 Comments
This blog post is part two of two. Like part one, published Wednesday, this is a cross-post from Groupon’s engineering blog. Thanks again to Kyle Oppenheim at Groupon. And one more reminder that I’ll be at the Percona Live MySQL Conference and Expo next week in Santa Clara, California so look for me there. You [...]
MySQL 5.6 vs MySQL 5.5 and the Star Schema Benchmark
March 11, 2013 By 7 Comments
So far most of the benchmarks posted about MySQL 5.6 use the sysbench OLTP workload. I wanted to test a set of queries which, unlike sysbench, utilize joins. I also wanted an easily reproducible set of data which is more rich than the simple sysbench table. The Star Schema Benchmark (SSB) seems ideal for this. [...]
MySQL 5.5 and MySQL 5.6 default variable values differences
February 21, 2013 By 8 Comments
As the part of analyzing surprising MySQL 5.5 vs MySQL 5.6 performance results I’ve been looking at changes to default variable values. To do that I’ve loaded the values from MySQL 5.5.30 and MySQL 5.6.10 to the different tables and ran the query:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 | mysql [localhost] {msandbox} (test) > select var55.variable_name,left(var55.variable_value,40) value55, left(var56.variable_value,40) var56 from var55 left join var56 on var55.variable_name=var56.variable_name where var55.variable_value!=var56.variable_value; +---------------------------------------------------+------------------------------------------+------------------------------------------+ | variable_name | value55 | var56 | +---------------------------------------------------+------------------------------------------+------------------------------------------+ | PERFORMANCE_SCHEMA | OFF | ON | | PID_FILE | /mnt/data/sandboxes/msb_5_5_30/data/mysq | /mnt/data/sandboxes/msb_5_6_10/data/mysq | | CHARACTER_SETS_DIR | /mnt/nfs/dist/mysql-5.5.30-linux2.6-x86_ | /mnt/nfs/dist/mysql-5.6.10-linux-glibc2. | | PERFORMANCE_SCHEMA_MAX_COND_INSTANCES | 1000 | 836 | | PERFORMANCE_SCHEMA_MAX_MUTEX_INSTANCES | 1000000 | 3282 | | OLD_PASSWORDS | OFF | 0 | | INNODB_STATS_ON_METADATA | ON | OFF | | PERFORMANCE_SCHEMA_EVENTS_WAITS_HISTORY_SIZE | 10 | 5 | | PERFORMANCE_SCHEMA_EVENTS_WAITS_HISTORY_LONG_SIZE | 10000 | 100 | | PERFORMANCE_SCHEMA_MAX_RWLOCK_INSTANCES | 1000000 | 1724 | | PERFORMANCE_SCHEMA_MAX_TABLE_HANDLES | 100000 | 2223 | | INNODB_LOG_FILE_SIZE | 5242880 | 50331648 | | BASEDIR | /mnt/nfs/dist/5.5.30 | /mnt/nfs/dist/5.6.10 | | BACK_LOG | 50 | 80 | | OPEN_FILES_LIMIT | 1024 | 5000 | | INNODB_AUTOEXTEND_INCREMENT | 8 | 64 | | MAX_CONNECT_ERRORS | 10 | 100 | | SORT_BUFFER_SIZE | 2097152 | 262144 | | LC_MESSAGES_DIR | /mnt/nfs/dist/mysql-5.5.30-linux2.6-x86_ | /mnt/nfs/dist/mysql-5.6.10-linux-glibc2. | | MAX_ALLOWED_PACKET | 1048576 | 4194304 | | JOIN_BUFFER_SIZE | 131072 | 262144 | | TMPDIR | /mnt/data/sandboxes/msb_5_5_30/tmp | /mnt/data/sandboxes/msb_5_6_10/tmp | | TABLE_OPEN_CACHE | 400 | 2000 | | INNODB_VERSION | 5.5.30 | 1.2.10 | | INNODB_BUFFER_POOL_INSTANCES | 1 | 8 | | QUERY_CACHE_SIZE | 0 | 1048576 | | SLOW_QUERY_LOG_FILE | /mnt/data/sandboxes/msb_5_5_30/data/dpe0 | /mnt/data/sandboxes/msb_5_6_10/data/dpe0 | | TABLE_DEFINITION_CACHE | 400 | 1400 | | PORT | 5530 | 5610 | | QUERY_CACHE_TYPE | ON | OFF | | REPORT_PORT | 5530 | 5610 | | PERFORMANCE_SCHEMA_MAX_FILE_INSTANCES | 10000 | 1556 | | SQL_MODE | | NO_ENGINE_SUBSTITUTION | | INNODB_OLD_BLOCKS_TIME | 0 | 1000 | | LOG_ERROR | /mnt/data/sandboxes/msb_5_5_30/data/msan | /mnt/data/sandboxes/msb_5_6_10/data/msan | | VERSION_COMPILE_OS | linux2.6 | linux-glibc2.5 | | THREAD_CACHE_SIZE | 0 | 9 | | PLUGIN_DIR | /mnt/nfs/dist/5.5.30/lib/plugin/ | /mnt/nfs/dist/5.6.10/lib/plugin/ | | SYNC_RELAY_LOG | 0 | 10000 | | GENERAL_LOG_FILE | /mnt/data/sandboxes/msb_5_5_30/data/dpe0 | /mnt/data/sandboxes/msb_5_6_10/data/dpe0 | | PERFORMANCE_SCHEMA_MAX_TABLE_INSTANCES | 50000 | 445 | | SYNC_RELAY_LOG_INFO | 0 | 10000 | | SLAVE_LOAD_TMPDIR | /mnt/data/sandboxes/msb_5_5_30/tmp | /mnt/data/sandboxes/msb_5_6_10/tmp | | SECURE_AUTH | OFF | ON | | VERSION | 5.5.30 | 5.6.10 | | INNODB_CONCURRENCY_TICKETS | 500 | 5000 | | INNODB_PURGE_THREADS | 0 | 1 | | INNODB_OPEN_FILES | 300 | 2000 | | INNODB_DATA_FILE_PATH | ibdata1:10M:autoextend | ibdata1:12M:autoextend | | INNODB_PURGE_BATCH_SIZE | 20 | 300 | | PERFORMANCE_SCHEMA_MAX_THREAD_INSTANCES | 1000 | 224 | | SOCKET | /tmp/mysql_sandbox5530.sock | /tmp/mysql_sandbox5610.sock | | INNODB_FILE_PER_TABLE | OFF | ON | | SYNC_MASTER_INFO | 0 | 10000 | | DATADIR | /mnt/data/sandboxes/msb_5_5_30/data/ | /mnt/data/sandboxes/msb_5_6_10/data/ | | OPTIMIZER_SWITCH | index_merge=on,index_merge_union=on,inde | index_merge=on,index_merge_union=on,inde | +---------------------------------------------------+------------------------------------------+------------------------------------------+ 56 rows in set (0.05 sec) |
Lets go over to see what are the most important changes [...]
Is there room for more MySQL IO Optimization?
January 3, 2013 By 11 Comments
I prefer to run MySQL with innodb_flush_method=O_DIRECT in most cases – it makes sure there is no overhead of double buffering and I can save the limited amount of file system cache I would normally have on database server for those things which need to be cached — system files, binary log, FRM files, MySQL [...]
Review of MySQL 5.6 Defaults Changes
October 8, 2012 By 10 Comments
James Day just posted the great summary of defaults changes in MySQL 5.6 compared to MySQL 5.5 In general there are a lot of good changes and many defaults are now computed instead of hardcoded. Though some of changes are rather puzzling for me. Lets go over them: back_log = 50 + ( max_connections / [...]
Write contentions on the query cache
September 5, 2012 By 5 Comments
While doing a performance audit for a customer a few weeks ago, I tried to improve the response time of their top slow query according to pt-query-digest‘s report. This query was run very frequently and had very unstable performance: during the time data was collected, response time varied from 50µs to 1s. When I ran [...]
A case for MariaDB’s Hash Joins
May 31, 2012 By Leave a Comment
MariaDB 5.3/5.5 has introduced a new join type “Hash Joins” which is an implementation of a Classic Block-based Hash Join Algorithm. In this post we will see what the Hash Join is, how it works and for what types of queries would it be the right choice. I will show the results of executing benchmarks [...]
