Thursday, January 28, 2010

MariaDB is Thinking about Fixing MySQL Replication and You Can Help

In case you have not noticed, MariaDB is joining the list of projects thinking about how to improve MySQL replication.   The discussion thread starts here on the maria-developers mailing list.

This discussion was jointly started by Monty Program, Codership, and Continuent (my employer) in an effort to push the state of the art beyond features offered by the current MySQL replication.  Now that things are starting to die down with the Oracle acquisition, we can get back to the job of making the MySQL code base substantially better.  The first step in that effort is to get a discussion going to develop our understanding of the replication problems we think are most important and outline a strategy to solve them. 

Speaking as a developer on Tungsten, my current preference would to be to improve the existing MySQL replication.  I suspect this would also be the preference of most current MySQL users.  However, there are also more radical approaches on the table, for example from our friends at Codership, who are developing an innovative form of multi-master replication based on group communications and transaction certification.  That's a good thing, as we want a range of contrasting ideas that take full advantage of the creativity in the community on the topic of replication.

If you have interest in improving MySQL replication please join the MariaDB project and contribute your thoughts.  It should be an interesting conversation.

9 comments:

hingo said...

Hi Robert

Seems like making MySQL better isn't so popular on Planet MySQL anymore :-)

Or then some people just think you have the wrong employer. Welcome to the club!

Robert Hodges said...

@hingo
If you mean that Oracle will not continue to improve MySQL, I would not be so sure of that. Oracle engineers are among the best in the industry and unlike some I believe their expertise will be applied to MySQL though not necessarily in ways everyone expects or through the same processes as in the past.

As for fixing MySQL--there's definitely interest and it's up to us to help get the focus back on that subject.

Mark Callaghan said...

Henrik,

Get things done and everyone will be thrilled. But this is all talk right now and lots of talk didn't get us much with the Maria storage engine.

There was also a lot of talk about how awful 5.1 is and yet MariaDB put out a 5.1 derived RC without fixing all of the problems mentioned in the original complaint.

I am betting on the work done by the MySQL replication team. They have made amazing progress in fixing bugs and fixing the core to make innovation easier. Jay and the Drizzle team have also made innovation easier.

hingo said...

Robert: Ah, you misunderstood me. I was just referring to the fact that you seem to have ended up in the group of bloggers that automatically gets 5 negative votes on planetmysql for posting what to me seems a technical and helpful post. (5 seems to be more or less the average number to expect.) I was not referring in any way to how MySQL is developed, more a meta commentary on something that many mysql alumni have recently observed on planetmysql.

Mark: As Robert said, the work done in MySQL 5.5 is of course one likely starting point for this work and the MySQL replication team is involved too. (Similarly we will look at Drizzle.) The plan for this project predates 5.5 - on the other hand a MySQL architect who worked at Sun during development of that seems to think it can still be improved upon.

Regarding your comments on 5.1 releases or Maria engine I don't feel that I have been involved in the issues you refer to so I don't have much to say.

Robert Hodges said...

@hingo
Aha. My reader does not show the votes it seems. This seems is a good example of how the whole topic has become unnecessarily politicized. I'm with Mark and the community as a whole that the benchmark needs to be working code. Speaking of which, back to work...

Mark Callaghan said...

Robert - by that metric, Tungsten is doing great with lots of code for people to use in production.

hingo said...

Robert: It's not that big a deal, it's only ~5 people out of millions of MySQL users, but since so few others actually use the voting system as it was intended, it becomes rather visible. I never said anything in public before, but seeing that they decided to drag you into it too I just felt bad for you. Since you never worked at MySQL it's like they're shooting at innocent bystanders.

Robert Hodges said...

Thanks guys. Your comments are very reassuring. (I think.) :)

hingo said...

Oh, and just for the record: I have absolutely nothing bad to say about the MySQL engineering organization. I worked closely with Tomas when he was still head of the Cluster team, and I'm convinced MySQL Server development has never been in better hands than it is now when he is in charge.

Scaling Databases Using Commodity Hardware and Shared-Nothing Design