This is one close to my heart. I’ve recently written on other storage engines: Where are they now: MySQL Storage Engines, The MERGE storage engine: not dead, just resting…. or forgotten and The MEMORY storage engine. Today, it’s the turn of MySQL Cluster.

Like InnoDB, MySQL Cluster started outside of MySQL. Those of you paying attention at home may notice a correlation between storage engines not written exclusively for MySQL and being at all successful.

NDB (for Network DataBase) started inside Ericsson, originally written in a language called …

There was once a big hooplah about the MySQL Storage Engine Architecture and how it was easy to just slot in some other method of storage instead of the provided ones. Over the years I’ve repeatedly mentioned how this wasn’t really …

One of the things we (Team MariaDB) talked quite a bit about since we released was PBXT. It was a feature differentiation to MySQL as we shipped another storage engine. It was included in MariaDB 5.1, 5.2, and 5.3; however with our release of MariaDB 5.5, PBXT (docs in the Knowledgebase) has been deprecated and not built by default any longer.

The reason behind it is clear: PBXT is currently not under active development. We still include it in the source releases and if you would like to use it, you just have …

In my presentation tomorrow at the MySQL Conference I plan to talk about some aspects of PBXT that I have never spoken about before. Here are the details of the presentation:

Update on the PBXT Storage Engine
10:50am Wednesday, 04/13/2011
Location: Ballroom D

Of course nothing about the engine is really a secret, if you are prepared to read the code. But who does that right? I am pretty sure that not even developers of other engines have spent much time (if any) on that.

But really, there are some gems stuck away in those X 1000 lines of code, and I plan to pick out a few tomorrow and show them to you. So don't miss it! :)

Yesterday I presented PBXT: A Transactional Storage Engine for MySQL at the German Oracle User Group Conference (DOAG) in Nuremberg. A number of people asked for the slides, so here is the link.

The talk was scheduled to be in English, but since I had a German-only audience I presented in German. There was quite a bit of interest, particularly in the Engine Level replication built into PBXT 2.0.

As Ronny observed, this feature can be used effectively for many tasks, including for online backup and maintaining a hot-standby. This all with the addition of a "small" feature:

The Master could initially stream the entire database over to the Slave before actual replication begins. This would also make it extremely easy to setup replication.

A brilliant idea, but a good 3 months work...

Finally I was able to run PBXT 1.0.11 pre-GA in tpcc-like workload, apparently there was bug with did not allow me to get the result earlier, and I am happy to see that PBXT team managed it.

For initial runs I took tpcc 100 warehouses ( about 10GB of data) which fully fits into memory (32 GB on server),
and compared 1 and 16 users in MySQL-5.1.46/PBXT and Percona Server / XtraDB - 5.1.45-rel10.2. As workload is totally memory based it will show how PBXT scales in CPU-bond cases on 16 cores systems.

As storage system it was Intel SSD X25-M card.

While full results and config are on Wiki:
http://www.percona.com/docs/wiki/benchmark:pbxt:tpcc:start

there are graphs for 1 user:

So this is about a SELECT COUNT(*) FROM tblname without a WHERE clause. MyISAM has an optimisation for that since it maintains a rowcount for each table. InnoDB and PBXT can’t do that (at least not easily) because of their multi-versioned nature… different transactions may see a different number of rows for the table table!

So, it’s kinda known but nevertheless often ignored that this operation on InnoDB is costly in terms of time; what InnoDB has to do to figure out the exact number of rows is scan the primary key and just tally. Of course it’s faster if it doesn’t have to read a lot of the blocks from disk (i.e. smaller dataset or a large enough buffer pool).

I was curious about PBXT’s performance on this, and behold it appears to be quite a bit faster! For a table with 50 million rows, PBXT took about 20 minutes whereas the same table in InnoDB took 30 minutes. Interesting! …

With Paul McCullagh’s PBXT storage engine getting integrated into MariaDB 5.1, it’s never been easier to it out. So we have, on a slave off one of our own production systems which gets lots of inserts from our Zabbix monitoring system.

That’s possibly an ideal usage profile, since PBXT is a log based engine (simplistically stated, it indexes its transaction logs, rather than rewriting data from log into index and indexing that) so it should require less disk I/O than say InnoDB. And that means it should be particularly suited to for instance logging, which have lots of inserts on a sustained basis. Note that for short insert burst you may not see a difference with InnoDB because of caching, but sustain it and then you can notice.

Because PBXT has such different/distinct architecture there’s a lot of learning involved. Together …

Here are the slides to my talk yesterday: A Practical Guide to the PBXT Storage Engine.

For anyone who missed my talk, I think it is worth going through the slides, because the are fairly self explanatory.

If there are any questions, please post them as a comment to the blog. I will be glad to answer :)

At this year's User Conference I have some interesting results to present. But more than anything else, my talk will explain how you can really get the most out of the engine. The design of PBXT makes it flexible, but this provides a lot of options. What tools are available to help you make the right decisions? I will explain.

Every design has trade-offs. How does this work out in practice for PBXT? And how can you take advantage of the strengths of the storage engine? I will explain in:

A Practical Guide to the PBXT Storage Engine
Paul McCullagh
2:00pm - 3:00pm Tuesday, 04/13/2010
Ballroom E

Don't miss it! :)