June 19, 2013

Post: MySQL 5.5.8 and Percona Server: being adaptive

used innodb_log_file_size=4G (innodb_log_files_in_group=2) for Percona Server. That gave 8GB in total size for log files (MySQLquery_cache_size=0 skip-name-resolve table_cache=10000 [mysql] socket=/tmp/mysql.sock (post edited by

Post: Replaying database load with Percona Playback

mysql-host=127.0.0.1 –mysql-user=root –mysql-schema=my_db –query-log-file=slow.log Executed 429467 queriesIn the previous report we see that 324 queries returned different row numbers from 429467 queries inupdates using real data and real queries

Post: Using any general purpose computer as a special purpose SIMD computer

by the native database interface (MySQL): mysql> select word, md5(word), md5(reverse(word)), count(*) from words2 group by 1,2,3 order byGROUP BY 1,2,3 ORDER BY NULL ) The other important optimization combines results from multiple queries together. This query

Post: Quickly finding unused indexes (and estimating their size)

… to consider dropping indexes in).  Let’s keep using views so this dynamically updates as our schema changes over time: mysql> create view all… ‘sakila%’ group by t.table_schema, t.table_name, i.index_name; Now I can query this view to see my indexes: mysql> select… ‘sakila%’ group by t.table_schema, t.table_name, i.index_name; Now we can can see index sizing information in the all…

Post: Percona XtraDB Cluster: Failure Scenarios with only 2 nodes

updates even in the case of split brain? This is a dangerous setting in multi-master setup, but should simplify things inin set (0.00 sec) percona2 mysql> insert into percona values (0,’percona2′,’daniel’); Query OK, 1 row affected (0.02 sec) percona2 mysql

Post: Filtered MySQL Replication

In quite a lot of cases I’ve seen such filtering being broken by developer forgetting this restriction and running one time update queryMySQL 5.0 way to break such replication is also using multi-table update

Post: MySQL 4 to MySQL 5 Upgrade performance regressions

usegroup commit would trigger in MySQL 4.0 and large number of transactions in total so serializing them would make MySQLqueries in “SHOW PROCESSLIST” stuck in “end” stage or have “commit” queries inin case there are away to many updates outside of transactions in auto_commit mode. Wrap them in

Post: Heikki Tuuri Innodb answers - Part I

by query where clause when it was not resolved by index HT: That is already implemented in…one of little known new features in MySQL 5.1 Q15: How …Group commit still busted in 5.0/5.1? HT: Yes. However, work has been done on it by Inaam. Q19: INNODB uses… portion of the questions. UPDATE: Second Part …

Post: PHP Large result sets and summary tables.

by default mysql_query uses mysql_store_result C library call and buffers all result set inusing temporary” in EXPLAIN) this was not issue for given case as GROUP BY in

Post: MySQL: Followup on UNION for query optimization, Query profiling

Using where | +—-+————-+——–+——-+—————+——+———+——+——+————-+ 1 row in set (0.00 sec) So as you see there are really different types of ranges in MySQL. ININ can’t: Lets say we want to show people in appropriate age group sorting by