Vorab ein Disclaimer: nein, ich bin nicht auf der Suche nach einen neuen Job

Ich habe auf der FrOSCon am letzten Wochenende Lukas Chaplin endlich mal persönlich getroffen, nachdem wir schon seit längerem in Online-Kontakt standen. Er betreibt das Job-Portal Linux Lancers (powered by MySQL), auf dem Stellenanzeigen mit dem Schwerpunkt auf Unix-, Linux- und Open Source-Jobs veröffentlicht werden. Dieses Jahr waren sie auf der FrOSCon als Gold-Sponsor und Aussteller vertreten und hatten wohl immer viel zu tun.

Ich finde diese Idee sehr gut – es zeigt, daß die Arbeit an und mit Open Source Software durchaus keine brotlose Kunst ist und Know-How in vielen Bereichen dringend gesucht wird. Mir ist bisher kein weiteres Portal mit diesem Fokus bekannt. Eine Suche nach "MySQL" liefert viele Hits, quer über die Republik verteilt. Kenntnisse und Erfahrung mit Opensource-Technologien zahlen sich aus! Die Website ist noch in einer frühen Betaphase, aber das machen die Inhalte wieder wett. Laut Lukas ist ein Relaunch des Portals in Arbeit, bei dem sowohl die Suchlogik als auch die Darstellung der Suchergebnisse überarbeitet wird.

With OpenSQL Camp and FrOSCon being over for almost a week now, it's time to come up with a short summary. I traveled home on Monday morning and then took Tuesday off, so I had some catching up to do...

As for the past years, FrOSCon rocked again! According to the closing keynote, they had around 1.500 (unique) visitors and I had a great time there. I really enjoyed meeting all the old and new faces of the various Open Source communities. The lineup of speakers was excellent, Jon "maddog" Hall's keynote about "Free and Open Source Software in the Developing World" was quite insightful and inspiring.

Most of the time I was busy with speaking at and running the OpenSQL Camp session track in room "HS6", but I managed to sneak out and listen to a few other talks as well. Additionally, I gave a presentation about ZFS on Linux, which had a good crowd and some interesting discussions.

This year, all of the sessions in the main lecture halls were streamed live, so people not able to attend could at least get a glimpse on what was going on these two days. Many times it was a really tough decision to make – there were way too many good sessions going in parallel. So being able to see the recording afterwards somewhat relieved that pain. The FrOSCon team has already begun to publish the video files – they are separated by day and lecture room and can be obtained from http://froscon.tv/.

And we were really lucky with the weather, too - the weekend was warm and sunny, a perfect setting for the social event, which is one of the highlights of FrOSCon. Impressive: this year, the attendees consumed one cubic meter (1000 liters) of Beer!

Finally, I'd like to express my special gratitude to all the speakers of OpenSQL Camp – thank you very much for your support and for devoting your time to participate in our subconference! It was appreciated.

I tend to switch between Linux and OpenSolaris as my desktop operating system from time to time. To be more flexible in this setup, I store most of my work-related data (e.g. source trees, VirtualBox images) on an external 320GB USB disk drive, using the ZFS file system. While OpenSolaris supports ZFS natively, I can access the file system on Linux using zfs-fuse and I could even mount these file systems on a FreeBSD system, if needed. There aren't that many file systems that allow an easy exchange of data between (Open)Solaris and Linux – the other ones that I am aware of are FAT and UFS, which both don't give me the confidence and flexibility I need.

A while ago, I purchased a second external drive of the same size and now use both of them in a mirrored configuration. This gives me several benefits:

• Redundancy: external disk drives have a higher risk of getting physically damaged, so having a mirrored copy of my data ensures that I won't lose anything important if one of the drives dies.
• Increased performance: ZFS is capable of distributing reads across both devices, thus I get twice the speed of a single USB 2.0 port, if I connect the drives to ports attached to separate USB host controllers.
• Automatic resyncing: When I'm on the road, I only take one half of the mirror with me. In case this drive gets lost/stolen, I still have a second copy of the data at home. ZFS complains about the pool being in degraded state, but continues to work normally. When I return home, I simply attach the second drive again and ZFS automatically resyncs (resilvers) anything that has been modified in the meanwhile.
• I can use snapshots for backup purposes. I have a small Intel Atom based PC (running OpenSolaris) that acts as my central file server (using two 1TB disks in a ZFS mirror) and CUPS print server. If I want to create a backup of my external USB disks, I create ZFS snapshots of the file systems in question and transfer these to the home server using "zfs send/receive". This works both locally by connecting the drives to the server box directly or via SSH over the network. I wrote a small shell script to automate this process. The transfer is done in an incremental fashion – only the differences between the current and the last snapshot are being propagated. To save disk space on the external drives, I usually discard all older snapshots except for the last one or two. On the file server, I maintain snapshots for a longer time period. The snapshots are named by using a simple date/time timestamp (filesystem@YYYY-MM-DD-HH:MM), this makes it easy to go back to a particular snapshot in case I'm looking for something I may have removed by accident, but I still recall when it was last used.

From my experience, ZFS is a very solid and reliable solution, providing impressive functionality with a very user-friendly UI (you only need use two commands, zfs and zpool).

If you want to learn more about ZFS and how to use it in practice, consider attending my upcoming talk (in German) about this subject at FrOSCon 2010 in Sankt Augustin, Germany!

If you wonder why there hasn't been an update from me for quite a while — I just returned from two months of paternal leave, in which I actually managed to stay away from the PC most of the time. In the meanwhile, I've officially become an Oracle employee and there is a lot of administrative things to take care of... But it feels good to be back!

During my absence, Giuseppe and Felix kicked off the Call for Papers for this year's European OpenSQL Camp, which will again take place in parallel to FrOSCon in St. Augustin (Germany) on August 21st/22nd. We've received a number of great submissions, now we would like to ask our community about your favourites!

Basically it's "one vote per person per session" and you can cast your votes in two ways, either by twittering @opensqlcamp or via the opensqlcamp mailing list. The procedure is outlined in more detail on this wiki page.

As we need to finalize the schedule and inform the speakers, the voting period will close this coming Sunday, 18th of July. So don't hesitate, cast your votes now! Based on your feedback we will compile the session schedule for this year's camp. Thanks for your help!

As I already wrote, I will be speaking at the MySQL Conference & Expo in Santa Clara in two weeks and I am excited to be there again. This year's conference is going to be interesting for a number of reasons, but most importantly I think that the schedule looks great! This is going to be a "drinking from the firehose of MySQL knowledge" event. Afterwards, I'll be on parental leave in May and June, so I likely will miss a lot of great conferences – these months are usually quite packed, as our Open Source Events Calendar can confirm. I just received a notice that my talk submission to OSCON has been rejected, which currently leaves me with two more speaking engagements in the upcoming weeks:

On April 24th, I'll be at the Grazer Linuxtage in Graz, Austria. The schedule has not been published yet, but I've been asked to give a keynote on the subject of working in a virtual company and a more technical session about MySQL HA solutions. Linuxtage is said to be the second largest Opensource event in Austria – they had 28 different sessions and around 450 visitors last year. I haven't been to an event in Austria for a while, so I look forward to being there!

Even though I'm technically on leave at that time, I will attend the amoocon in Rostock, Germany in June (4-6). While last year's focus at this event was on opensource telephony (Asterisk, VoIP et al), they decided to broaden the scope for this year's event: "It is a boutique conference where we create an environment to give every attendee a fair chance to actually speak to each speaker. So you can tank knowledge and new ideas without the bullshit-bingo." I really enjoyed my stay there last year and look forward to talking about "A look into a MySQL DBA's toolchest" (for those who won't make it to my talk about this at the MySQL conference) and "Why you should be using a DVCS". I noticed that Monty Widenius will be there as well, speaking about "MariaDB release 5.1 - What is it and what to expect from it." and "MySQL & MariaDB history". The organizers are also looking for a speaker from the PostgreSQL camp, so this is going to be an interesting event for me. In addition to that, Rostock is a pretty nice city and the baltic sea is nearby. The organizers have limited the number of attendees to 100 people and the ticket price is slowly increasing every second day – so make your reservations now!

Berkeley DB (BDB) is undoubtedly the workhorse among the opensource embedded database engines. It started as a university project in the mid-eighties and was further developed by Sleepycat Software, until it got acquired by Oracle in February 2006.

I had the impression that BDB had lost a lot of its popularity among opensource developers to SQLite in recent times, which has evolved into becoming the default choice for developers looking for an embedded data store. I'd assume primarily because the code is not released under any particular license, but put in the public domain (which makes it very attractive for embedding it into one's code), and also because it's lightweight, supports SQL and has interfaces to a number of languages.

Of course, SQLite has its limitations and use cases (as every product), so it may not be suited for some particular application. As the SQLite developers put it: "SQLite is not designed to replace Oracle. It is designed to replace fopen().".

Yesterday, Oracle announced a new version of BDB. One of the notable features of this release is the introduction of a new SQL API, based on SQLite. According to Gregory Burd, Product Manager for Berkeley DB at Oracle, they did so by including a version of SQLite which uses Berkeley DB for storage (replacing btree.c). I think this is a very smart move – instead of introducing a new API, developers can now easily switch to a different storage backend in case they are experiencing issues with the default SQLite implementation. So now MySQL isn't the only database with different storage backends anymore

I am curious to learn more about how the BDB implementation compares against the original (both feature- and performance-wise).

Oh, and this is actually not the first time someone put an SQL interface in front of Berkeley DB – BDB was the first transaction-safe storage engine that provided page-level locking for MySQL in version 3.23.15 (released in May 2000). The InnoDB storage engine was added some time afterwards (MySQL 3.23.34a, released in March 2001).