Wikimedia Foundation is expanding Wikipedia to multimedia with Sun Open Storage Solution and MySQL Database:

Wikipedia receives between 25,000 and 60,000 page requests per second, depending on the time of day. Wikimedia needed to update its infrastructure to handle this huge volume of traffic and ensure that its systems were reliable, highly available, and easily scalable. It also wanted to expand its upload file limit from 20 MB to 100 MB to accommodate rich media (audio and video) content, but before it could do that it needed to expand its storage capacity.

…

My presentation at LOSUG on tuesday went down like a house on fire - I think it would be safe to say that the phrase for the evening was ‘It’s a cache!’.

For that to make sense, you need to look at the slides, which are now available here.

Attendance was great, but it seems the last minute change of day meant that some people missed the session. We had 151 people register, and about 80 turned up on the night.

This is a simple mysql circular replication implementation on a single machine (just a proof of concept) which can easily be done on a broader scale. Just be aware of the cons of circular replication which mainly gets down to: once a node freaks out or stops for one reason or the other, it’s a bitch and you need to take care of IMMEDIATELY.

Download Sandbox from https://launchpad.net/mysql-sandbox/+download
Download MySQL from http://dev.mysql.com/downloads

Copy the downloaded software onto the your *nix box onto any folder of your preference called $BASEDIR

run:

cd /$BASEDIR

gunzip mysql_sandbox_X.X.XX.tar.gz
tar -xf mysql_sandbox_X.X.XX.tar

ln -s mysql_sandbox_X.X.XX sandbox

time /$BASEDIR/sandbox/make_replication_sandbox –circular=4 –topology=circular /$BASEDIR/mysql-5.1.30-linux-x86_64-glibc23.tar.gz


user@hostname $ time …

Next Tuesday (27th Jan), I’ll be speaking at the London OpenSolaris User Group again. For those that follow the LOSUG meetings, we normally the third thursday of the month, but due to the overwhelming popularity of the event this month (more 100 registrations so far) we have had to push the event back to the last Tuesday of the month.

This month, I’ll be talking about the DTrace probes that we have added into MySQL and demonstrating their use and functionality. Along the way I’ll also cover some of the internals of MySQL and how it works (and how they relate to the DTrace probes we’ve added), how to use the probes to analyze and diagnose your queries, and how I’ve already used the DTrace probes to provide information up to the Query Analysis functionality within …

Installing MySQL is quite an easy thing to do, especially when done using pkgs, dmgs or exes. It gets just a tad more time consuming and brain intensive when installing a .tar.gz binary package. It is when you’re compiling MySQL source directly that you’ll need some planning and playing, but the latter is only done in particular cases such as when you’ll need a particular engine not shipped with a pre-compiled package etc.

Today we’re going through the steps required for a typical MySQL installation from a .tar.gz package on a *nix based platform, including the download, installation, configuration and securing.

Steps involved:
1. Download MySQL binary tarball from mysql.com
2. Create a folder structure where the installation will be held.
3. Install the package
4. Secure the installation

Step 1: Download MySQL

Go to http://dev.mysql.com/downloads/ and choose the particular …

After getting sysbench running properly with a scalable memory allocator (see last post), I can now return to what I was originally testing - what memory allocator is best for the 5.1 server (mysqld).

This stems out of studies I have made of some patches that have been released by Google. You can read about the work Google has been doing here.

I decided I wanted to test a number of configurations based on the MySQL community source, 5.1.28-rc, namely:

• The baseline - no Google SMP patch, default memory allocator (5.1.28-rc)
• With Google SMP patch, mem0pool enabled, no custom malloc (pool)
• With Google SMP patch, mem0pool enabled, linked with mtmalloc (pool-mtmalloc)
• With Google SMP patch, mem0pool disabled, linked with tcmalloc (TCMalloc)
• With Google SMP patch, mem0pool …

After getting sysbench running properly with a scalable memory allocator (see last post), I can now return to what I was originally testing - what memory allocator is best for the 5.1 server (mysqld).

This stems out of studies I have made of some patches that have been released by Google. You can read about the work Google has been doing here.

I decided I wanted to test a number of configurations based on the MySQL community source, 5.1.28-rc, namely:

• The baseline - no Google SMP patch, default memory allocator (5.1.28-rc)
• With Google SMP patch, mem0pool enabled, no custom malloc (pool)
• With Google SMP patch, mem0pool enabled, linked with mtmalloc (pool-mtmalloc)
• With Google SMP patch, mem0pool disabled, linked with tcmalloc (TCMalloc)
• With Google SMP patch, mem0pool …

After getting sysbench running properly with a scalable memory allocator (see last post), I can now return to what I was originally testing - what memory allocator is best for the 5.1 server (mysqld).

This stems out of studies I have made of some patches that have been released by Google. You can read about the work Google has been doing here.

I decided I wanted to test a number of configurations based on the MySQL community source, 5.1.28-rc, namely:

• The baseline - no Google SMP patch, default memory allocator (5.1.28-rc)
• With Google SMP patch, mem0pool enabled, no custom malloc (pool)
• With Google SMP patch, mem0pool enabled, linked with mtmalloc (pool-mtmalloc)
• With Google SMP patch, mem0pool disabled, linked with tcmalloc (TCMalloc)
• With Google SMP patch, mem0pool …

This Thursday (December 18th), Martin "MC" Brown will talk about using DTrace with MySQL. MC is the Solaris (and, naturally, DTrace and ZFS) expert on the Sun Database Group documentation team. He's helped the openSolaris team port MySQL to openSolaris.

Note that we'll be using a new session address / Dimdim URL:

http://webmeeting.dimdim.com/portal/JoinForm.action?confKey=mysqluniversity

You can bookmark this address, since it will remain valid for all future MySQL University sessions. Remember, though, that the meeting room will open only 15 minutes before the session starts.

Dimdim is the conferencing system we're using for MySQL …

My mind is playing "Suffering Succotash..."

I have been working on MySQL performance for a while now, and the team I am in have discovered that SysBench could do with a couple of tweaks for Solaris.

Sidebar - sysbench is a simple "OLTP" benchmark which can test multiple databases, including MySQL. Find out all about it here , but go to the download page to get the latest version.

To simulate multiple users sending requests to a database, sysbench uses multiple threads. This leads to two issues we have identified with SysBench on Solaris, namely:

• The implementation of random() is explicitly identified as unsafe in multi-threaded applications on Solaris. My team has found this is a real issue, with occasional core-dumps happening to our …