June 2013 - MariaDB and Galera Synchronous Replication
Hi Visitor, and welcome to the June 2013 issue of the Open Query Newsletter.
In this issue:
- short news
- MariaDB replacing MySQL
- Galera Synchronous Replication
- Training days
- upcoming events
- tip: FULLTEXT indexing or SPHINX
It is custom to introduce Open Query people to you, and this month's choice is Grant Allen. Grant joined us late last year, after a stint at Google! His angle on the move is that the way we work better suits the lifestyle he wants now. Great for us! Grant has also written and co-written a number of books at Apress on a variety of topics, so he is truly a versatile lad! Grant is a joy to work with, and we've very pleased to have caught him.
While we naturally don't hire from clients, we do like to keep you informed about our hiring plans as they're generally aimed at providing you with better service and making sure that noone gets overloaded as our company grows. For this purpose, we're currently seeking an Internal Systems Engineer, and a Sales Account Manager (a geek/geekess with sales clue, not a sales droid - rest assured). If you happen to know someone suitable in your circle of friends, please direct them to our jobs page at http://openquery.com/company/jobs
Following us via the Open Query blog or Twitter (@openquery) is a good way to catch new posts and other topical items we identify. After all, the maximum frequency of this newsletter is only monthly, as to not clutter your inbox from our end.
MariaDB replacing MySQL
After Fedora and other community distros, Red Hat too is moving towards MariaDB in its release of RHEL 7 later this year. The rough background is that Oracle's commitments to the European Union (negotiated when it acquired Sun Microsystems) regarding MySQL will end early in 2014 and it's anybody's guess what effort they'll put in after that. And there are other issues of concern, some of which have been covered already on the Open Query blog.
MariaDB is not new to many of Open Query's clients, most actually have been running MariaDB for several years, and our own enhanced OurDelta builds before that. It's proven. If you need references for this of Australian corporations running MariaDB, ask us.
The MySQL development and community is alive and well, but it's called MariaDB now. We already wrote an item in one of our newsletters last year about the fact that MariaDB (since 5.3) has optimised subqueries. Things like that really help applications and clients.
For more details about the Red Hat announcement which happened at the recent Red Hat Summit, see http://www.itwire.com/business-it-news/open-source/60292-red-hat-ditches-mysql-switches-to-mariadb
Of course, we can assist with upgrading from MySQL to MariaDB. You won't regret it.
Galera Synchronous Replication
Galera is another technology only available in MariaDB. We're all familiar with the class MySQL replication, which work asynchronously. The Galera system, developed by the Finnish company Codership, is available in MariaDB 5.5 (production version) adds synchronous replication to the mix. This truly changes the rules, as now we can now achieve better write scalability: any Galera server is effectively a "master", and they can be written to in parallel. Further, because of the optimised mechanism it's viable to have Galera replication across datacentres. That's right, full active-active. Galera can work in conjunction with classic replication, so backup, reporting and possibly other read-only slaves still have their place.
Naturally, with technology such as this there are new things to learn, and that's why we've updated our two-day Backup, Recovery and Replication Workshop to incorporate Galera. See below for upcoming course days near you!
Last chance to book your training within the current financial year!
We currently have a number of course days scheduled around Australia and New Zealand. MariaDB and its enhancements are of course covered as well as stock MySQL Server. You can register for Open Query course days/modules individually to suit your time, budget and current needs. Your trainers for these days wil be Arjen and Peter.
- Wed 24 Jul 2013: Query Performance Analysis and Tuning (AUD 395 + GST [BYO laptop])
- Thu-Fri 25-26 Jul 2013: Backup, Recovery and Replication Workshop (AUD 975+GST [PC lab])
- Thu-Fri 29-30 Aug 2013: Backup, Recovery and Replication Workshop (AUD 975+GST [PC lab])
- Thu-Fri 24-25 Oct 2013: Backup, Recovery and Replication Workshop (AUD 975+GST [PC lab])
- Thu-Fri 21-22 Nov 2013: Backup, Recovery and Replication Workshop (AUD 975+GST [PC lab])
We will shortly be scheduling additional course days on a number of topics. If you have specific needs, please let us know.
The current full overview is online at http://openquery.com/training/schedule
For bookings and questions, contact us today! All prices excluding GST.
- Brisbane QLD, Wed 10 Jul 2013: SAGE-AU QLD Brisbane chapter meeting. Arjen is doing a talk on the Galera technology.
- Auckland NZ, 21-23 Oct 2013: Open Source Developers Conference. Note that our Auckland training days are adjacent!
- Perth WA, 6-10 Jan 2014: Linux.conf.au 2014. At the University of Western Australia in beautiful Perth.
- Brisbane QLD, fourth Tuesday of the month: Brisbane Web Tech meetup, incorporating MySQL and many other relevant technologies. Arjen co-organises.
Tip: FULLTEXT indexing or SPHINX
You've made it this far, so now to our freebie newsletter tip!
Some applications still use FULLTEXT indexing built in to the MyISAM storage engine, but there are problems. It ties you to the MyISAM storage engine (although InnoDB has added a similar but not identical system recently), but worse than that, the performance for both reads and writes becomes increasingly horrendous when the dataset grows. Really not what you want.
Sphinx is a stand-alone fulltext indexing system that integrates neatly with MySQL and MariaDB:
- it can read new data directy from your tables;
- there are APIs for your applications to talk with;
- Sphinx can also talk the MySQL client protocol;
- MariaDB (like the OurDelta builds before it) includes the SPHINX storage engine so that you can use its search capabilities directly in your queries and join on to other tables to retrieve the data you want.
And it's fast. If you would like more information, or get assistance with either conversion from FULLTEXT or a new implementation, ask us!
Until next time!
Feedback welcome through http://openquery.com/contact
You can also access this issue online: http://openquery.com.au/newsletter/2013-06, other issues can be viewed at http://openquery.com.au/category/newsletter/open-query-newsletter
-- Arjen & the OQ team
We aim to keep our newsletter in plain text, apart from links. If you're reading the Open Query Newsletter online or received it via someone else, you can subscribe for your own copy through http://openquery.com/user/register and http://openquery.com.au/newsletter/confirm/add/