Hakan Küçükyılmaz

It has been almost a month since my last WFTO posting.

This time my delay was due to our migration from BitKeeper to Bazaar. You can read about the background and reasoning in our move to Bazaar in

• Lenz Grimmer’s Blog posting: The MySQL source code has moved!

An excellent introduction about using Bazaar with MySQL you will find at

• Daniel Fischer’s article: Getting started with Bazaar for MySQL code

Compared to our last WFTO we have 6 new tests and 2 tests are not failing anymore. This means that we are at a failed/passed ratio of 7.5%. We are making a steady progress:

• 7.5% - 15/198 (this report)
• 8.9% - 17/190 (last report)
• 10.4% - 19/182 (report before last one)

News for these weeks:

• Jim Starkey fixed
  • Bug#37517 falcon -DENGINE conflicts –with-ssl leading to build failure

• Kevin Lewis fixed
  • Bug#36438 Falcon crash in Record::poke, line 823

• Chris Powers fixed
  • Bug#36468 Falcon replication failure, engine reports HA_ERR_END_OF_FILE on slave

• Vladislav Vaintroub fixed
  • Bug#37251 Livelock between UPDATE and DELETE threads
  • Bug#37080 Falcon deadlock on concurrent insert and truncate
  • Bug#36396 Assertion in IO::pread (on closed tablespace file)

• Olav Sandstå fixed
  • Bug#36745 falcon crash on solaris
  • Bug#36703 Unknown symbol EncodedDataStream::decode when linking mysqld with Sun Studio

• Hakan Küçükyılmaz fixed
  • Bug#37078 falcon_bug_26828 fails sometimes on Pushbuild

What about you?

We are interested in you! Where do you use Falcon? What do you do with Falcon? Are there any features you want to see in Falcon? You can test Falcon and get famous by providing valuable bug reports or even test cases for Falcon!

Resources:

• The test language we use is described here: MySQL Test Framework
• List of Falcon related bugs you will find here: http://bugs.mysql.com/saved/FalconBugs
• A preview version of Falcon you will find here: http://forge.mysql.com/wiki/Falcon_Feature_Preview
• Our MySQL Forum dedicated to Falcon related questions is here: http://forums.mysql.com/list.php?133
• To get latest Falcon source code, please see: http://dev.mysql.com/doc/refman/6.0/en/installing-source-tree.html

It has been a while since the last Weekly Falcon Test Overview, also called WFTO.

Our Falcon team was with the Falcon Developers Meeting in London last week. Other than that, I updated my workstation running the WFTO test suite from RHEL4 to RHEL5. This caused some delay in the weekly reporting.

The WFTO test suite now contains RHEL5 64-bit (native), Ubuntu 32-bit (vm), Windows 2003 Server 64-bit (vm), and Mac OS X/PPC 32-bit (native).

We have one new test case falcon_bug_36294 and three tests are not failing anymore.

News for these weeks:

• Jim Starkey fixed
  • Bug#36294 Assertion in Cache::writePage
  • Bug#36367 Falcon DBT2 crash in Cache::ioThread.

• Kevin Lewis fixed
  • Bug#36097 Be sure the highWater index is set to zero when a record is thawed
  • Bug#36636 Move fsync out of the #ifdef HAVE_PREAD scope so that it works in either case.
  • Bug#36603 Adjust age group values to uint32

• Vladislav Vaintroub fixed
  • Bug#36467 Falcon assertion in ha_partition.cc: virtual int ha_partition::extra()

• Christopher Powers fixed
  • Bug#36330 Falcon DBT2 crash in Transaction::needToLock

• Sergey Vojtovich fixed
  • Bug#33607 Falcon allows two mysqld instances to share a tablespace
  • Bug#33404 Falcon ignores partition-level TABLESPACE option

• Hakan Küçükyılmaz fixed
  • Bug#36486 Falcon compilation fails on Solaris 10/x86
  • Bug#36400 Compiling Falcon on Solaris 10/x86 fails with Sun Studio 12

What about you?

We are interested in you! Where do you use Falcon? What do you do with Falcon? Are there any features you want to see in Falcon? You can test Falcon and get famous by providing valuable bug reports or even test cases for Falcon!

Resources:

• The test language we use is described here: MySQL Test Framework
• List of Falcon related bugs you will find here: http://bugs.mysql.com/saved/FalconBugs
• A preview version of Falcon you will find here: http://forge.mysql.com/wiki/Falcon_Feature_Preview
• Our MySQL Forum dedicated to Falcon related questions is here: http://forums.mysql.com/list.php?133
• To get latest Falcon source code, please see: http://dev.mysql.com/doc/refman/6.0/en/installing-source-tree.html

This week we added six new tests to the Falcon test suite. We had six tests, which where labeled as big tests. Big test means here, that the runtime or the resource requirements of that test are higher than usual. As our internal test systems are running steady integration tests on a 24/7 basis, we have to make a trade off and limit the test resource requirements to get an acceptable test throughput. However, from time to time these big tests should be run. In the Falcon team we agreed on running these tests in our Weekly Falcon Test Overview (WFTO) cycles.

We also see that there are two tests, which are not failing anymore. We disabled two tests, which will not be fixed in any near future:

• falcon_bug_22166.test Test for case-sensitive identifiers.
• falcon_bug_22187.test Test for Foreign Key. Useless as long we don’t have a general FK support from MySQL server side. Please note that Falcon does support FK. It’s just that we at MySQL are working on a general FK support for all storage engines.

News for this week:

• Jim Starkey made a change in commitNoUpdate related to Bug#34602 and Bug#36296. He worked on BLOB performance enhancement, online add column, and online native backup.
• Kevin Lewis pushed his changes for syncPrior code.
• Chris Powers worked on Bug#36269: Thread stalls during DBT2 run.
• Vladislav Vaintroub worked on fixing various Falcon and MySQL related bugs.
• Hakan Kuecuekyilmaz did test case reorganizations.

What about you?

We are interested in you! Where do you use Falcon? What do you do with Falcon? Are there any features you want to see in Falcon? You can test Falcon and get famous by providing valuable bug reports or even test cases for Falcon!

Resources:

• The test language we use is described here: MySQL Test Framework
• List of Falcon related bugs you will find here: http://bugs.mysql.com/saved/FalconBugs
• A preview version of Falcon you will find here: http://forge.mysql.com/wiki/Falcon_Feature_Preview
• Our MySQL Forum dedicated to Falcon related questions is here: http://forums.mysql.com/list.php?133
• To get latest Falcon source code, please see: http://dev.mysql.com/doc/refman/6.0/en/installing-source-tree.html
• Worklog entry about Foreign Key implementation: http://forge.mysql.com/worklog/task.php?id=148

At MySQL we run a whole regression test suite on various platforms after each code “push”. If you are coming from CVS or SVN and alike, then think about a “commit” when we say “push”.

The test language we use is described here: MySQL Test Framework

We at the Falcon team wanted to see our progression in terms of bugs fixed, so for over 18 months we additionally are running Falcon related tests on a weekly basis. With lot’s of teams working on MySQL sometimes there are tests, which fail in rather random fashion. Such tests are usually disabled, so that other teams don’t get distracted. For our Weekly Falcon Test Overview we enable all disabled test, to see where we are.

I will try to do a public weekly report about the status of Falcon along with the Weekly Falcon Test Overview results, bugs we have fixed, and noteworthy news.

Last week we saw this bug fixes for Falcon:

• Jim Starkey fixed couple of bugs.
• Sergey Vojtovich fixed this tablespace related bugs: BUG#33723, BUG#34048, and BUG#34617.
• Vladislav Vaintroub was very busy hunting down a duplicate key bug we saw lately in Falcon.

As we are an Open Source company, our bug system is open to all: http://bugs.mysql.com/

If you are interested in Falcon related bugs, you can use our saved search at: http://bugs.mysql.com/saved/FalconBugs

As a special add-on for fans of numbers and statistics, here is the list of the Falcon test overview status in plain text:

2006-09-01 Failed 30/67 tests, 55.22% were successful.
2006-09-15 Failed 28/75 tests, 62.67% were successful.
2006-09-22 Failed 30/76 tests, 60.53% were successful.
2006-09-29 Failed 26/77 tests, 66.23% were successful.
2006-10-06 Failed 26/79 tests, 67.09% were successful.
2006-10-13 Failed 25/83 tests, 69.88% were successful.
2006-10-20 Failed 26/83 tests, 68.67% were successful.
2006-10-27 Failed 28/88 tests, 68.18% were successful.
2006-11-03 Failed 31/92 tests, 66.30% were successful.
2006-11-10 Failed 31/99 tests, 68.69% were successful.
2006-11-17 Failed 26/101 tests, 74.26% were successful.
2006-11-24 Failed 26/102 tests, 74.51% were successful.
2006-12-01 Failed 25/102 tests, 75.49% were successful.
2006-12-08 Failed 25/105 tests, 76.19% were successful.
2006-12-15 Failed 22/106 tests, 79.25% were successful.
2006-12-22 Failed 23/107 tests, 78.50% were successful.
2006-12-29 Failed 24/107 tests, 77.57% were successful.
2007-01-05 Failed 24/107 tests, 77.57% were successful.
2007-01-12 Failed 25/110 tests, 77.27% were successful.
2007-01-19 Failed 26/110 tests, 76.36% were successful.
2007-01-26 Failed 23/110 tests, 79.09% were successful.
2007-02-02 Failed 22/110 tests, 80.00% were successful.
2007-02-11 Failed 22/110 tests, 80.00% were successful.
2007-02-16 Failed 23/114 tests, 79.82% were successful.
2007-02-23 Failed 27/123 tests, 78.05% were successful.
2007-03-02 Failed 27/123 tests, 78.05% were successful.
2007-03-10 Failed 43/127 tests, 66.14% were successful.
2007-03-17 Failed 30/127 tests, 76.38% were successful.
2007-03-24 Failed 29/129 tests, 77.52% were successful.
2007-04-15 Failed 28/132 tests, 78.79% were successful.
2007-04-20 Failed 32/142 tests, 77.46% were successful.
2007-04-27 Failed 28/145 tests, 80.69% were successful.
2007-05-04 Failed 30/148 tests, 79.73% were successful.
2007-05-11 Failed 27/148 tests, 81.76% were successful.
2007-05-18 Failed 26/149 tests, 82.55% were successful.
2007-05-25 Failed 33/150 tests, 78.00% were successful.
2007-06-01 Failed 32/155 tests, 79.35% were successful.
2007-06-08 Failed 29/155 tests, 81.29% were successful.
2007-07-06 Failed 23/158 tests, 85.44% were successful.
2007-07-13 Failed 19/159 tests, 88.05% were successful.
2007-07-20 Failed 20/159 tests, 87.42% were successful.
2007-07-27 Failed 20/160 tests, 87.50% were successful.
2007-08-03 Failed 20/161 tests, 87.58% were successful.
2007-09-07 Failed 24/170 tests, 85.88% were successful.
2007-09-14 Failed 29/172 tests, 83.14% were successful.
2007-09-21 Failed 28/172 tests, 83.72% were successful.
2008-01-10 Failed 21/185 tests, 88.65% were successful.
2008-01-24 Failed 21/185 tests, 88.65% were successful.
2008-02-01 Failed 22/185 tests, 88.11% were successful.
2008-02-28 Failed 17/191 tests, 91.10% were successful.
2008-03-13 Failed 21/194 tests, 89.18% were successful.
2008-03-20 Failed 21/194 tests, 89.18% were successful.
2008-03-28 Failed 19/189 tests, 89.95% were successful.
2008-04-04 Failed 21/194 tests, 89.18% were successful.

Again some time past by since our last 6.0.4 release of Falcon.

This Feature Preview is intended to allow the community to test some important performance improvements to the Falcon Database. The performance changes are listed below along with key bug fixes.

Several performance improvements were made in this release:

• Supernodes
• Thread signaling

As always you can find out details about our Falcon 6.0.5 Feature Preview at MySQL Forge.

Download is here: http://downloads.mysql.com/forge/falcon_feature_preview/

Your feedback on Falcon is highly welcome. Happy testing!

Dear Falconers,

MySQL 6.0.4-alpha, a new version of the MySQL database system including
the Falcon transactional storage engine (now at beta stage), has been released. The main page for MySQL 6.0 is at:

http://www.mysql.com/mysql60/

If you are new to the Falcon storage engine and need more information,
please read the Falcon Evaluation Guide at:

http://www.mysql.com/why-mysql/white-papers/falcon-getting-started.php

and the Falcon White Paper at:

http://www.mysql.com/why-mysql/white-papers/storage-engines-falcon.php

MySQL 6.0.4-alpha is available in source and binary form for a number
of platforms from our download pages at

http://dev.mysql.com/downloads/mysql/6.0.html

and mirror sites. Note that not all mirror sites may be up to date at
this point in time, so if you can’t find this version on some mirror,
please try again later or choose another download site.

We welcome and appreciate your feedback, bug reports, bug fixes,
patches etc.:

http://forge.mysql.com/wiki/Contributing

The change list is extremely long. Therefore I did not c&p’ed it here. You can view it online at:
http://dev.mysql.com/doc/refman/6.0/en/news-6-0-4.html

Wer hätte das gedacht? Dieter Bohlen und eine Meinung, die ihn wirklich sympathisch macht! In der Johannes B. Kerner Talkshow erzählt er interessante Details über die Musikindustrie, zu dem Zustandekommen von Charts und der Kriminalisierung von Filesharing-Teilnehmern.

Die Zeichen der Zeit hat Herr Bohlen auf jeden Fall erkannt: die Künstler von heute sollten sich auf Live-Auftritte und Konzerte konzentrieren; die Zeiten, in denen man über Konservenmusik (CDs, DVDs und Schallplatten) Geld verdienen konnte, sind eindeutig vorbei.

Das ganze Interview bei Youtube

We started a new category at MySQL Forge called Software preview to allow sneak peeks at products that are shown to the community before their official distribution.

You will find things like GIS functions and Online Backup there. The latest addition to the preview page is Falcon Feature Preview. Some time has passed since our last release of  Falcon in MySQL 6.0.3, which we released November, 16^th 2007. Since that version was cloned off we fixed over 20 Bugs and added new features like:

• Pool of Asynchronous I/O Threads
• Direct I/O
• Pool of Gopher Threads
• A new Thread Scheduler

You can download binaries for Linux x86-64 and Windows 32-bit. You can also download the source to build MySQL 6.0.4 Feature Preview on your own. Check it out at:

Falcon Feature Preview

Please note that Falcon now builds on Mac/Intel, Solaris/x86, and FreeBSD as well.

As always your feedback on Falcon is highly welcome. Happy testing!

I was always wondering whether developing MySQL with Eclipse/CDT will work or not. I tried to import our code base more than a year ago, but it did not worked out that well.

Eclipse/CDT got a lot better and I tried to import the MySQL code base again. To my positive surprise this time it worked.

You can try it for yourself: I wrote up a small howto covering installation, configuration, and basic debugging of MySQL with Eclipse/CDT at MySQL Forge.

You can find the document here: Eclipse/CDT on Linux and Mac OS X

I am looking forward to get your feedback. There are still some open points I would like to figure out:

• How to ignore SCCS/ directories of Bitkeeper in “Project View” and for code lookup?
• How to generate Makefile without running full BUILD/compile-xyz script. Use BUILD/compile-xyz -c (notice the -c option). Thanks to Timour Katchaounov who figured this out.
• Starting debug triggers a “make all” every time, which costs too much time. How to avoid “make all” if no files where changed?

We had our first Laptopbattle in Stuttgart last Saturday. It was the fifth Laptopbattle in Germany so far. We put in a lot of energy and time to organize this battle, so our expectations where quite high.

After six months of organization we had eight contestants for this great event:

* nq, Munich
* Wesen, Karslruhe
* tylth, Stuttgart
* tranqui, Munich
* KABELTON, Graz/Austria
* Hollaender, Mannheim
* Spleen, Mannheim
* Ter Roshak, Munich

Two months ago I was about to give up the battle, because we had only two contestants and no press coverage. I don’t know exactly how it happened, but we found the remaining 6 contestants and so I continued to push Stuttgart’s first Laptopbattle.

For the visuals we had two VJ’s Sebastian and Julia. And for the after party we got
Kenny C from London. One of our judges from Berlin could not make it, so we where left with two judges: Karl and Tom. They decided to pick a third judge from the crowd.

The event started a bit delayed as our technician and judge from Berlin could not make it. After an hour or so fighting with the sound setup we started sometime around 11:30pm.

Then the hardest part begun: the battle! As organizer I have contact to each contestant and after each round I was happy and sad at once. I was happy for the guy who made it to next round and I was sad about the guy who got kicked out. The other sad part as an organizer is that one cannot really enjoy the event, because there are so much thing one has to care about and look after.

Despite the great stress I can recall to be impressed by the great visuals and nq’s first round. nq started with some texture and no beats at all. The people expected an ambient track or something, but suddenly nq kicked in a great bass, which also kicked the crowd. Also Wesen’s show by controlling Ableton Live with his wii controller while dancing in the crowd was great.

The result:
In the finals we had nq vs. Hollaender. The judges could not decide who to win, so we made it an audience decision. It wasn’t that unfair as nq and Hollaender had no close friends with them at the battle. Finally the audience chose Hollaender to get the first place. Second came nq and KABELTON made third.

Why we are doing Laptopbattles, somebody could ask. We do it to push electronic music to the next level. Here in Germany we still suffer that techno and electronic music is connected to acts like Scooter. In Stuttgart we had a lot of people never heard of artists performing live on their laptops and they got curious and came to see what it is about. We also to bring people together and to build contacts. While organizing this battles I learned a number of new things around electronic music like netlabels, em411, the STFU, various new artists and I am sure that there is a lot more learn and discover.

The battle is a frame to have a nice event and to present and represent electronic music and the people behind it. The battle format is also good for attracting sponsors and media. I still have to work on sponsors and media.

The future:
In Stuttgart I got the impression that the Laptopbattle as an event works. We had > 200 people joining the party. However, I still could not figure out how to break even with the financial part of the event. Either I have to organize it on a much lower quality like in a bar or I have to look for partners and expand the Laptopbattle in Germany. My secret goal is to lead the Laptopbattle in Germany to the German finals with the winners of eight different cities. So far we have three cities: Munich, Mannheim, and Stuttgart. For next year I plan Frankfurt.

http://laptopbattle-stuttgart.de/?lang=en
http://www.alx-soft.de/main/music/overview/071110_laptopbattle_stuttgart/