To follow on to my earlier links to Brendan Gregg’s blog posts on measuring I/O latency, there is a third one discussing DTrace, and then a very detailed response from Mark Leith showing how to do it with the PERFORMANCE_SCHEMA in MySQL 5.5.

Related posts:

1. Disk latency versus filesystem latency
2. Measuring the popularity of the Percona MySQL build
3. Thoughts on the new …

Shuttleworth opens a can of worms. Fedora 15. IBM commits to Hadoop. And more.

# Mark Shuttleworth shared his thoughts on companies, contributor agreements and free software, prompting responses from Simon Phipps, and Dave Neary.

# The Fedora Project launched Fedora 15.

# IBM launched a new version of its BigInsight software, based on Hadoop, committed $100m to …

OldSQL DBs based on B-trees have some well-known problems and workarounds. TokuDB is a NewSQL storage engines based on Fractal Tree indexing, so the natural question is how InnoDB practice translates into TokuDB. This post gives a quick overview. Enjoy!

FAQ

Q: How do I tune TokuDB?
A: You don’t!

TokuDB has almost no parameters you can set, and many of them are related to diagnostics — things like how often should the information about insertion or query progress be updated. We recommend that parameters be left unchanged.

Q: What do I do when indexes are bigger than memory?
A: Nothing!

TokuDB is based on Fractal Trees, which are designed to index fast, even when indexes are bigger than memory.

Q: What’s …

In the #maria IRC channel the other day I fielded a question someone had about a DATETIME column where they wanted to SELECT by the year. The answer (which is probably obvious to many of you) is to use the YEAR() function like so:

select * from t1 where YEAR(d) = 2011;

(The above assumes the table is named “t1” and the DATETIME column is named “d“.)

In my reply I provided a link to the Date and Time Functions section of the AskMonty Knowledgebase, but when I looked at the entry for the …

To follow on to my earlier links to Brendan Gregg’s blog posts on measuring I/O latency, there is a third one discussing DTrace, and then a very detailed response from Mark Leith showing how to do it with the PERFORMANCE_SCHEMA in MySQL 5.5.

Just a couple of announcements for Percona Live: we’re giving away free tickets through Engine Yard, and we’ve arranged our evening open-bar event.

For a chance to win a free ticket, check out the guest blog post I wrote for Engine Yard on how to make Rails migrations faster and safer.

In other news, Percona Live now has an open-bar evening event, sponsored by Clustrix. It is open to all, not just registered attendees of Percona Live. You can come to Mickey Mantle’s legendary Sports Bar and Restaurant for free drinks and food, networking, prizes, and a chance to win an Unlimited Gold MySQL Support contract from Percona, worth over $15,000!

In case you don’t know, Percona Live is a …

Many organizations consider MySQL as a database because the initial license cost is free (*). Larger organizations that use Oracle and SQL Server also consider implementing MySQL as a means to lower the total cost of software infrastructure due to the initial cost for new software licenses or expensive upgrades due to new hardware.

However free software does not mean that services to support MySQL should be also free or even cheap. Recently a large multi-national customer wanted professional consulting and training for MySQL resources and they seemed shocked that I wanted to charge a reasonable rate for professional services. My MySQL consulting rates are cheaper then industry MySQL peers and also similar skilled resources providing Oracle consulting.

With over 10 years experience in MySQL, many doing consulting and also over 10 years professional experience before MySQL I am more then qualified to provide the best possible …

Last week Oleksandr "Sanja" Byelkin and I pushed Dynamic columns into MariaDB 5.3.

Dynamic columns allows you to store a different set of columns for every row in the table. This allows you to solve things that you can't normally do with a relational database, like the web store problem where you want to store attributes for different things, like a computer (memory, speed, weight) and t-shirt (size, color).

Dynamic columns works by storing the extra columns in a blob and having a small set of functions to manipulate it. The functions exist both in SQL and in the MariaDB client library to allow you to manipulate the data where it suits you best. The client code is BSD to make it easy to port the code to other languages.

You can use a dynamic column in SQL …

Few developers consider, when trying to build robust platforms, all the possible modes of failure. Indeed, it is difficult to consider them all, let alone plan for them, or design tests which exercise particular symptoms.

In this post, I discuss some of the types of failure we can see in real systems.

Complete server failure

Most developers DO consider this. In a "Complete server failure", what generally happens is:

* The server stops processing new requests, completely.
* The server's OS no longer responds to any network request at all (e.g. "ping")
* Processing does not continue within the server
* The contents of memory are immediately and irretrievably lost.

Typically, the server recovers, and when it does so, it is rebooted and restored to full health. All writes which were acknowledged before its failure have been persisted.

This is …

Permalink: http://bit.ly/1vSmnGm



There is no native undo ability inside MySQL. Thus, the common methods to workaround the problem is through creating backup dumps and enabling binary logging, using transactions, and requiring the WHERE clause in Data Manipulation Language commands by using the safe updates option. These methods have drawbacks:

1. Creating backups via mysqldump and using binary logging to revert to a point in time will have the same effect as an undo functionality. However, these are executed via the command line shell. Since these tools are not executed inside MySQL, this method may not be convenient and presents limitations on when it can be used.
2. Transactions allow you to "undo" as long as you have not committed your data manipulation changes. Imagine if you discover data manipulation changes that you wish to …