I think if you use this with NBD or Sharding or any other write scale application, it would make sense for geographically dispersed nodes.

I will be Southern California next week at two public events in Los Angeles and San Diego. MySQL 5.6 Update on August 1st and MySQL: New Features, Programming Mistakes and Optimization on August 2nd. Both are in Oracle offices, start at 7PM, and there will be Pizza! Please be totally cool dudes and dudettes and RSVP so we can order enough pizza to keep things totally tubular! And there will be MySQL SWAG given away!

Dave Stokes and I tracked down Sarah Novotny Sunday night at the early bird registration for OSCON 2012. We had a few moments to interview her about MySQL Connect and her recent endeavors. I hope you enjoy:

Keith Larson: Thank you for joining us today.

Sarah Novotny: I am glad to be here.

Keith Larson: First just a general question, What were your thoughts when you heard that Oracle was going to do MySQL Connect ?

Sarah Novotny: I was very excited about it. I didn't expect Oracle to give us our own space. Even thought JavaOne had gotten that . JavaOne has existed as it's own conference before, and with the O'Reilly Conference no longer happening,  I didn't expect it and I was surprised. I know the first year after Oracle purchased MySQL, we had MySQL Sunday, which was great. We were told that this was such a big deal and it is really only a transition thing and with 140+ products nobody gets a big name …

I believe I’m nearly done with the SHOW EXPLAIN feature. The idea is that if you’ve got some long-running query $LONG_QUERY running in connection $LONG_QUERY_CONN, you should be able to create another connection, run SHOW EXPLAIN FOR $LONG_QUERY_CONN and see what query plan is being used to run the $LONG_QUERY.

How is this different from just running explain $LONG_QUERY? First, you don’t need to replicate the exact environment that $LONG_QUERY was run in: you don’t need to figure out what values it had for @@optimizer_switch and other settings, what were the contents of its temporary tables, if any, what did InnoDB told the optimizer about the table statistics, etc.

Another, indirect benefit is that we will now be able to produce query’s EXPLAIN at arbitrary point in time, which should make it possible to …

Hi Justin,

I agree that for certain OLAP applications, this might be acceptable. I’ll have to read up on group commit to see how it might help this – I haven’t done any testing w/ 5.6 yet.

With regards to PXC, can you elaborate on how this is mitigated? While some applies could happen in parallel, I expect that each query will still block the client for at least 85ms, which would have a negative impact to most web/OLTP applications. Perhaps this is a good excuse for me to do some testing with PXC.

Another scenario where I think this might be ok is if you have multiple semi-sync slaves, with some of them colocated with the master. Since the protocol only requires one ack, as long as the local slaves are healthy, they would tend to respond first. I’m also curious how semi-sync works with tiered replicas – is the protocol supported in the case where there is a relay slave?

When you're spoiled with Oracle's fabulous query transformation capabilities and its really well-done cost-based optimiser, then you might forget how difficult SQL query tuning used to be in the "old days" or with those less sophisticated databases. Here's a really nice explanation of the various means of implementing an ANTI-JOIN in MySQL: http://explainextended.com/2009/09/18/not-in-vs-not-exists-vs-left-join-is-null-mysql/

The MySQL Market Continues to Grow I am continuing to see the continued growth of MySQL not only in the SMB market but in Fortune 500 accounts.    I recently asked a large customer what their distribution of databases is and they responded with:

2000 Oracle databases 4000 SQL Server databases 6000 MySQL databases

MySQL the Internet Database Platform

Since taking over MySQL, Oracle has added

Here is a perfect post to read during your holidays :-)
There are many ways to get news on the internet : blogs, rss feeds, facebook, linkedin, twitter…
What I want to talk about here is how I use twitter to follow the MySQL Community and how I stay up to date of the latest news.
I use twitter only for focus on the MySQL news and the MySQL community, that’s why I would like to share this experience with you.

Step 1 : Choose your friends

Q: What is the most complicated with Twitter ?
A: Read all the tweets that we receive every day

Of course it can be complicated and that takes a long long time…
That’s why you must choose who is your best friends carefully.

Except for the MySQL rock stars, I advise you to use TweetStats before to blindly follow anyone who tweeted a dark grigri …

[...] Aaron Brown has a gem of a post about performance of MySQL Semi-Synchronous replication over high latency connections. [...]

In a typical organization, all work together to bring out a common good for the outside world. It’s interesting to see how all of these entities blog about technology, and there is more and more interest shown by managerial technologists about the database. This Log Buffer Edition appeases their appetites along with the others in [...]