Category Archives: InnoDB

Speaking at OSCON 2010

I am speaking in a few days at OSCON 2010 in Portland, Oregon. My talk, “MySQL Bottleneck Hunters – How Schooner Increased MySQL Performance by 8x” is about how Schooner optimized MySQL to run on modern hardware with flash memory. Come on by if you’re at OSCON!

Here’s the abstract:

MySQL Bottleneck Hunters – How Schooner Increased MySQL Performance by 8x — Thursday, July 22, 2010 at 11:30am

MySQL users have an insatiable need for speed, capacity, and availability, all at a reasonable cost. This session will provide technical overview of the approach that Schooner engineering took to optimize MySQL Enterprise and InnoDB with flash memory, multi-core processors, and DRAM to achieve an 8x improvement in performance relative to existing systems.

This session will provide a crash course in the performance tuning tools and techniques that Schooner used to radically improve MySQL performance on commodity hardware. It will also offer an overview of the results that were achieved, and compare these results to other commonly deployed MySQL server architectures, including stock MySQL on SSDs and Fusion-io, using the industry-standard DBT2 benchmark.

MySQL Cluster splits from MySQL — not pluggable?

Kaj just announced that as of 5.1.25, MySQL Cluster will no longer be included in the normal MySQL packages. Instead, a new branch is being created for MySQL Cluster releases, where the first release is informally called “6.2.15” but the releases are really named “mysql-5.1.23-ndb-6.2.15”. This new branch is based on the MySQL Cluster Carrier Grade Edition of MySQL.

Overall, this seems like a good idea — the needs for releases, schedules, pressures, etc., are at odds between MySQL Cluster and MySQL’s core. I am, however, baffled by the decision of how to release the new product: coupled with MySQL as a single monolithic package with compiled-in storage engine. After all, 5.1 has long been touted to have this amazing new pluggable storage engine architecture. Why not use it?

With Oracle/InnoDB recently announcing that they are decoupling from MySQL and releasing their storage engine as a plugin, this would make so much sense. The only thing I can think of is that MySQL and/or the MySQL Cluster team must think that the pluggable storage engine concept is not workable in its current state or easy enough to use… which I would agree1 with absolutely. However, having a team within MySQL really pushing a product and using the pluggable interface to make their releases would help dramatically to make the interface really usable for the rest of the world.

Why not do it? Let’s hear it…

1 My short-list of gripes, by no way complete: The error message interface sucks, no way to compile a plugin outside of the MySQL source (or even symlink it into the source tree), plugins are tied to an exact version of MySQL server, and no reasonable way to manage plugins in an OS context (RPMs, .debs, etc).

Bravo Oracle: InnoDB Plugin 1.0 released

Yesterday, Oracle‘s Ken Jacobs and Heikki Tuuri, creator of InnoDB, have announced the immediate release of InnoDB Plugin 1.0 for MySQL 5.1. I’ve already downloaded it and played around with it a bit. I haven’t had time to do any performance benchmarks or similar just yet. Those will come in due time.

This release is the beginning of two exciting things: InnoDB is now officially decoupled from MySQL release-wise, and lots of new features have been added to this new release. I will come to what the decoupling means in a moment, but first, the major new features in this release of InnoDB (from my perspective, and with my commentary):

  • Fast in-place index management — The ability to add and drop indexes without rebuilding the entire table in place. This isn’t a complete implementation of the long-awaited “online ALTER TABLE“, as that is mostly a MySQL problem rather than an InnoDB problem.
  • Compression of data and indexes — This should allow data size on disk to be reduced substantially.
  • Storage of entire BLOB, TEXT, and VARCHAR data off-page — This can allow more efficient PRIMARY KEY indexes (where the data is stored in InnoDB because of the index clustering) on tables with BLOB, TEXT, or large VARCHAR columns
  • New information_schema tables — InnoDB is now providing a lot more visibility into what is going on internally. I’m hopeful to extend this even further.
  • InnoDB-specific “strict mode” — Don’t allow InnoDB to fudge things internally, forcing it to error in circumstances it can’t handle, rather than just give a warning or silently continue.

All of the above features look excellent, but one of the more interesting aspects of this announcement is the fact that MySQL and InnoDB are now decoupled for release. That is, Oracle can make a release of InnoDB without having to wait for MySQL to make their own release. While this will make it slightly more difficult to describe what version of MySQL/InnoDB you’re using (especially without a way to find that out from MySQL), it has the potential to make the release process much quicker and more efficient.

I’m quite hopeful, and from what I have seen in the recent past with the collaboration between Heikki, Ken, and Yasufumi (an outside contributor) on fixing InnoDB performance bugs, quite positive, that Oracle will be a lot more accepting of outside patches and code contributions to the InnoDB codebase than MySQL has been recently. I’ve got a lot of ideas for new features targeted at manageability that I’d like to get implemented. I’d love to hear Oracle’s comments on how they will accept patches to InnoDB now, and what they see the community interaction looking like in the future. Ken, any comment?

The only negative aspects I can see with this announcement are:

  • Oracle spent a couple of years working on this in silence, away from the community. Most everyone was surprised by this release, as we haven’t seen anything from Oracle about InnoDB in a long time, and to some extent we were sitting in the corner hopeful that Oracle really doesn’t plan on killing the project. While I understand that perfectly well from a business standpoint, I am hopeful that working for long periods away from the community can be minimized in the future, so that we can all stay more involved.
  • It is now a bit harder to get a new MySQL/InnoDB installation up and running, as the newest InnoDB is not part of the MySQL packages anymore. I’m sure this can be cleaned up with some smart RPM (etc.) packaging.

Overall, I am excited about this announcement, and quite happy that Oracle is making some serious contributions and commitments to maintaining InnoDB. Thanks for all your hard work, Ken and Heikki and the rest of the InnoDB team! Let me know if there’s anything I can do for you!