A few days ago I wrote about MySQL performance implications of InnoDB isolation modes and I touched briefly upon the bizarre performance regression I found with InnoDB handling a large amount of versions for a single row. Today I wanted to look a bit deeper into the problem, which I also filed as a bug.

First I validated in which conditions the problem happens. It seems to happen only in REPEATABLE-READ isolation mode and only in case there is some hot rows which get many row versions during a benchmark run. For example the problem does NOT happen if I run sysbench with “uniform” distribution.

In terms of concurrent selects it also seems to require some very special conditions – you need to have the connection to let some …

Over the past few months I’ve written a couple of posts about dangerous debt of InnoDB Transactional History and about the fact MVCC can be the cause of severe MySQL performance issues. In this post I will cover a related topic – InnoDB Transaction Isolation Modes, their relationship with MVCC (multi-version concurrency control) and how they impact MySQL performance.

The MySQL Manual provides a decent description of transaction isolation modes supported by MySQL – I will not repeat it here but rather focus on performance implications.

SERIALIZABLE – This is the strongest …

Query caching is one of the prominent features in MySQL and a vital part of query optimization. It is important to know how it works as it has the potential to cause significant performance improvements – or a slowdown – of your workload.

The MySQL query cache is a global one shared among the sessions. It caches the select query along with the result set, which enables the identical selects to execute faster as the data fetches from the in memory. It is important to have everything identical, no new comments, spaces, or most significantly differences in the WHERE clause. Basically when you trigger a select query, if it is available in the cache; it fetches from there or it considers the query as a new one and will go to the parser.

Even though it has some nice advantages, the MySQL query cache has its own downsides too. Well, let’s think about this: If you are frequently updating the table, you are then invalidating …

Some time ago, Peter Zaitsev posted a blog titled “How well does your table fits in innodb buffer pool?” He used some special INFORMATION_SCHEMA tables developed for Percona Server 5.1 to report how much of each InnoDB table and index resides in your buffer pool.

As Peter pointed out, you can use this view into the buffer pool to watch a buffer pool warm up with pages as you run queries. You can also use it for capacity planning. If you expect some tables need to be fully loaded in the buffer pool to be used efficiently, but the buffer pool isn’t large enough to hold them, then it’s time to increase the size of the buffer pool.

The problem, however, was that system tables change from version to version. Specifically, the INNODB_BUFFER_POOL_PAGES_INDEX table no longer exists in Percona Server 5.6, and the …

It has been a while since I have looked at InnoDB crash recovery. A lot has change in the last few years – we have serious crash recovery performance improvements in MySQL 5.5 and MySQL 5.6, we have solid state drives raising as typical high performance IO subsystem and we also have the ability to set much larger log files and often have a much larger InnoDB Buffer Pool to work with.

First let me revisit the challenge with have with InnoDB configuration. For write-intensive workloads it is extremely important to size innodb_log_file_size for good performance, however the longer log file size you have the longer you might have to wait for InnoDB to complete crash recovery, which impacts your recovery strategy.

How much can innodb_log_file_size impact performance? Massively! Doing intensive writes to a database that well fits in memory, I’d say there’s a 10x …

A few years ago Peter Zaitsev, in a post titled “To UUID or not to UUID,” wrote: “There is timestamp based part in UUID which has similar properties to auto_increment and which could be used to have values generated at same point in time physically local in BTREE index.”

For this post I’ve rearranged the timestamp part of UUID (Universal Unique Identifier) and did some benchmarks.

Many people store UUID as char (36) and use as row identity value (PRIMARY KEY) because it is unique across every table, every database and every server and allow easy merging of records from different databases. But here comes the problem, using it as PRIMARY KEY causes the problems described below.

Problems with UUID

• UUID has 36 characters which makes it bulky.
• InnoDB stores data in …

I believe InnoDB storage engine architecture is great for a lot of online workloads, however, there are no silver bullets in technology and all design choices have their trade offs. In this blog post I’m going to talk about one important InnoDB limitation that you should consider.

InnoDB is a multiversion concurrency control (MVCC) storage engine which means many versions of the single row can exist at the same time. In fact there can be a huge amount of such row versions. Depending on the isolation mode you have chosen, InnoDB might have to keep all row versions going back to the earliest active read view, but at the very least it will have to keep all versions going back to the start of SELECT query which is currently running.

In most cases this is not a big deal – if you have many short transactions happening you will have only a few row versions to deal with. If you just use the system for reporting queries but do not …

Often enough I find MySQL benchmark results where the difference between results is 1% or even less and some conclusions are drawn. Now it is not that 1% is not important – especially when you’re developing the product you should care about those 1% improvements or regressions because they tend to add up. However with such a small difference it is very important to understand whenever this is for real or it is just the natural variance for your baseline test.

Take a look at this graph:
Click the image for a larger view)

This is the result for a simple in-memory, read-only “select by primary key” SysBench benchmark on dedicated physical hardware that is otherwise idle, simple 1 socket system. I tried to stabilize it as much as possible, for example …

Starting with MySQL 5.6 there is an INNODB_METRICS table available in INFORMATION_SCHEMA which contains some additional information than provided in the SHOW GLOBAL STATUS output – yet might be more lightweight than PERFORMANCE_SCHEMA.

Too bad INNODB_METRICS was designed during the Oracle-Sun split under MySQL leadership and so it covers only InnoDB counters. I think this would be a great replacement to all counters that are currently provided though SHOW STATUS – it captures more information such as providing MIN/MAX counts for variables as well as providing the type of the counter (whenever it is current or commutative) as well as human readable comment – describing what such counter means.

The examples of data you can get only from the INNODB_METRICS table includes information about InnoDB Page Splits and merging (which can …

I visited MongoDB Day in London on November 6. Here are a few observations:

App-Developer Centric. It is interesting to see how much MongoDB is about developers; the ops side is something which is a necessary evil developers have to deal with. The ops topics covered in principle that there are no topics about choices of operating systems or hardware for MongoDB beyond flash and more memory.

Development Stacks. Being application centric there was good coverage of the MongoDB-powered stacks – MEAN and METEOR specifically got attention. Especially the METEOR presentation by Henrik Ingo was cool – real-time view synchronization between the Web browser (or mobile app) and database as well as the same language for server-side and client-side development is a really great concept. Though …