Background

InnoDB gathers statistics for the data in user tables, which are used by the MySQL optimizer to choose the best query plan. For a long time the imprecision and instability of these statistics have been creating problems for users.

The problem is that these statistics are recalculated at any of the following events:

* When the table is opened

* When the table has changed a lot (1/16th of the table has been updated/deleted or inserted)

* When ANALYZE TABLE is run

* When SHOW TABLE STATUS or SHOW INDEX FROM is run

* When InnoDB Monitor is turned ON

* others

and so their recalculation must be quick an unnoticeable. Thus the quick algorithm just picks 8 random pages and could give a wildly varying results.

…

One of the most important things a database user or DBA wants to know is what columns, indexes etc. a table has. There are a few ways to find these things out, such as show tables. However, to really reveal all of the detailed metadata information in an InnoDB database, many have tried the “innodb_table_monitor” to peek into internal data dictionary. By creating an “innodb_table_monitor” table, InnoDB will print out the contents of metadata periodically. Unfortunately, it is printed out in an unstructured text for each table, and to find out what you need you would have to either carefully scan the output or have your own parser to do some additional analysis on the result if you want to display them systematically.

Well, in the MySQL 5.6 release, “innodb_table_monitor” can become history, and you will no longer need to search the text output to find out system metadata information. Instead, you can query the InnoDB internal …

In MySQL 5.6.2 we have introduced a new background thread named the page_cleaner in InnoDB. Adaptive flushing of modified buffer pool pages in the main background thread, async flushing by a foreground query thread if the log files are near to wrapping around, idle flushing and shutdown flushing are now moved to this thread, leaving only the last resort sync flushing in foreground query threads. We’ve also added counters for these activities.

As page_cleaner is all about the flushing of dirty pages to disk it’ll do you a world of good if you can go through this post where I have explained different types of flushing that happen inside InnoDB and the conditions that trigger flushing. The page_cleaner thread is only concerned with flush_list flushing (this may change in future releases). So let us dig a bit deeper into why …

Data Dictionary Language (DDL) operations have traditionally been slow in MySQL. Any change to the table definition would be implemented by creating a copy of the table and index layout with the requested changes, copying the table contents row by row, and finally renaming tables and dropping the original table.

The InnoDB Plugin for MySQL 5.1 implements a more efficient interface for creating and dropping indexes. Indexes can be created or dropped without rebuilding the entire table.

Speeding up Bulk INSERT, UPDATE and DELETE Operations

Normally, InnoDB would update all indexes of a table when rows are inserted or deleted. If you update an indexed column, InnoDB would have to delete the old value and insert the new value in the corresponding index. If you update a primary key column, the row would be deleted and inserted in every index of the table. This can be slow even despite the …

Purpose

What does purge exactly do and why is it needed? If you have ever wondered then read on. It is really a type of garbage collector. When a user issues a DML like “DELETE FROM t WHERE c = 1;”, InnoDB doesn’t remove the matching record. This is what happens under the hood:

1. It marks the record as deleted by setting a bit in the control bits of the record.
2. Stores the before image of the modified columns to the UNDO log
3. Updates the system columns DB_TRX_ID and DB_ROLL_PTR in the clustered index record. DB_TRX_ID identifies the transaction that made the last change, and DB_ROLL_PTR points to the new UNDO log record. This UNDO log record contains the old values of DB_TRX_ID and DB_ROLL_PTR, possibly pointing to an older transaction and undo log entry.

From …

“The least expensive query is the query you never run.”

Data access is expensive for your application. It often requires CPU, network and disk access, all of which can take a lot of time. Using less computing resources, particularly in the cloud, results in decreased overall operational costs, so caches provide real value by avoiding using those resources. You need an efficient and reliable cache in order to achieve the desired result. Your end users also care about response times because this affects their work productivity or their enjoyment of your service. This post describes some of the most common cache methods for MySQL.

Popular cache methods

The MySQL query cache

When the query cache is enabled, MySQL examines each query to see if the contents have been stored in the query cache. If the results have been cached they are used instead of actually running the query.. This improves the response time …

Nowadays MySQL Databases are encapsulated into many mission critical software solutions. Lots of companies host one or many MySQL databases in their data center, sometimes even without knowing it except when the MySQL service is no more available. In order to increase this service availability it is mandatory to have a monitoring solution. Regardles of you are using MySQL Server, MySQL replication or cluster the Oracle/MySQL monitoring solution is called MySQL Enterprise Monitor.

Isolation is an important part of ACID properties that guarantee that transactions are processed in a reliable manner. But there are four different levels of isolation available and you have to understand each one of them to be able to select the correct one for your needs. This post intends on explaining the four levels together with their effects on locking and performance.

This week we conclude the interview with Tokutek's co-founder and Chief Scientist, Martin Farach-Colton. Part 1 of the interview. TokuDB is a transactional, fully ACID-compliant storage engine that uses fractal trees for data and indexes, instead of MySQL's standard B-tree implementation. Combined with TokuDB's message-based architecture, TokuDB data and indexes do not fragment, have completely online column and index addition and removal and unlike InnoDB and XtraDB, do not fall apart when indexes no longer fit into memory.

read more

I got access to our 12-core Intel server, so I was able to do some better benchmarks to test the different group commit thread scheduling methods:

This graph shows queries-per-second as a function of number of parallel connections, for three test runs:

1. Baseline MariaDB, without group commit.
2. MariaDB with group commit, using the simple thread scheduling, where the serial part of the group commit algorithm is done by each thread signalling the next one.
3. MariaDB with group commit and optimised thread scheduling, where the first thread does the serial group commit processing for all transactions at once, in a single thread.

(see the previous post linked above for a more detailed explanation of the two thread scheduling algorithms.)

This test was run on a 12-core server with hyper-threading, memory is …