ICE MySQL Query Cache patch

Infobright ICE doesn't currently use the MySQL Query Cache for queries executed by the Infobright engine.
The attached patch enables the MySQL Query Cache for Infobright queries also.

There are currently a bug report about this here.

Patch file for ICE 3.3.1, should also work for earlier version:
queryCache331.patch (1 KB)

ICE 3.3.1 released

Today Infobright 3.3.1 has been released.
The most important change are that MySQL have been upgraded from 5.1.14 to 5.1.40 GA, which fixes a bunch of different bugs, both performance, query syntax, etc., in the MySQL engine.
Because of the MySQL upgrade it is necessary to run the MySQL Updater program, as the data file structure has been changed in MySQL 5.1.40 GA. Upgrade instructions are available here:
Windows Upgrade Instructions
Linux Upgrade Instructions
Tar Upgrade Instructions
There are also a known issue with the 3.3.1 version.

Please note: There is a known issue with 3.3.1. A server crash may occur for queries of the form:
SELECT distinct(a1), max(a2) FROM t1 GROUP BY a3;
Where a1, a2, a3 may or may not be the same. It is caused by the use of max and min and the distinct clause, and may be easily worked around by removing the distinct and adding an additional clause to the GROUP BY.
This issue has been resolved and will be included in the next release. A full list of known and fixed issues is included at the end of this document.

All the information about the new release can be found here and can be downloaded here.
The patches previously posted on this blog, will be updated to work in 3.3.1. The updated patches will be added to the old post.

MySQL and Infobright

At my work place we have used the Infobright Community Edition, http://www.infobright.org/, for a while now, and with great success.

For those who don't know what Infobright is, I recommend that you take a look at this page, What is ICE?, ICE stands for Infobright Community Edition, which is the open source version.

One of Infobrights great strengths are compression, which is top-notch. And for certain analytic queries it performs very, very good.

But there are certainly room for improvements, fx. as Infobright uses its own optimizer, not the default MySQL optimizer, certain queries aren't optimized very well yet, and some aren't supported at all. Unsupported queries will fall back to the standard MySQL optimizer, which doesn't work well with Infobright tables.

Also they are very slow to implement contributed patches, because of this, I've decided to post patches in this blog. Patches which are committed to they're bug tracker, but also patches which are developed specifically for use at my company.