Stephane's thoughts corner…

Stephane Bailliez's thoughts on everything

I wanted to recycle my old Dell D810 laptop (Pentium M 2GHZ, 2GB RAM, 80GB HD) that was sitting there doing nothing and install Xen on it so that I could have a server for development purposes as well as a media center. For some odd reasons I decided to install the latest Ubuntu 9.10 on it (making sure that it was actually working fine on the laptop)…and have a go at Xen 3.4.2… little did I know how much time I would spend on it. It wasn’t obviously the best match in all aspects.

For the details, I tried to use a vanilla kernel 2.6.31.12. I based the install from Brandon Turner blog entry (keep in mind a few things need to be adapted depending on your system, ). After many failed attempts due to bad grub2 configuration, I managed to load the kernel…a bit.. it crashes with:

Cannot access memory beyond end of bootstrap direct-map area, Unknown interrupt.

A little bit of googling direct me to a post in the xen-users mailing list:

No, it’ll be a 32-bit-specific issue though. I would recommend just running a 64-bit build of Xen, if the CPU supports 64-bit mode. It’ll be something to do with GRUB2 dumping the dom0 kernel and initrd higher in memory than GRUB1, past the end of where 32-bit Xen is set up to be able to access. Probably not a hard fix, although I have no GRUB2 installation to test with.

Ok, so..off we go, I’m downgrading grub2 and revert back to grub-legacy.

After a bit of fiddling, I manage to boot again only to be greeted by another giant stacktrace following:

BUG: unable to handle kernel paging request at c07ab790.

Obviously I can spend weeks on this, debugging and nagging xen-users, but I think I have played enough with patches and alternatives to figure out that best thing to do is to avoid that combination of Ubuntu 9.10 + Xen 3.4.2 on a Pentium M machine. I’ll see what VirtualBox looks like on 9.10 and then decides whether or not to try with an older distrib if it does not do the job.

My estimed friend and former colleague and manager Geir released recently through his company 10gen a schema-free document-oriented datastore called mongodb.

I’m fairly impressed with the velocity at which they were able to release it as well as the apparent quality of the documentation for a product just out of the door.

This is on my todo list to check among other things as I can see a couple of use cases that could fit well this.

If you happen to use java and mysql do yourself a favor and don’t take the 5.1.7 jdbc driver.

I recently opened bug 42970 following a quick profiling to find the reason why 100 test calls of a 3.5KB-long sql query increased from 4.5s to 50s by upgrading the driver from 5.1.6 to 5.1.7.

There is a subtle piece of code in 5.1.7 which tries to find the string ‘ ON DUPLICATE KEY UPDATE ‘ in every statement in a very peculiar way

In my case, as can be seen from the call tree below, 4 PreparedStatement, means 4 calls to containsOnDuplicateKeyInString() and generates more than 12M calls to String.charAt() , 9M calls to Characters.toUpperCase, 4.5M calls to Characters.toLowercase().

  99,9% - 386 s - 4 inv. com.mysql.jdbc.JDBC4PreparedStatement.
    99,9% - 386 s - 4 inv. com.mysql.jdbc.PreparedStatement. (line: 47)
      99,9% - 386 s - 4 inv. com.mysql.jdbc.PreparedStatement$ParseInfo. (line:635)
        99,9% - 386 s - 4 inv. com.mysql.jdbc.PreparedStatement.containsOnDuplicateKeyInString (line: 202)
          99,9% - 386 s - 4 inv. com.mysql.jdbc.StringUtils.indexOfIgnoreCaseRespectMarker(line: 5193)
            99,7% - 386 s - 3 002 inv. com.mysql.jdbc.StringUtils.indexOfIgnoreCase (line:1023)
              18,0% - 69 578 ms - 9 530 899 inv. java.lang.String.charAt (line: 946)
              10,3% - 40 015 ms - 5 506 413 inv. java.lang.Character.toUpperCase (line:946)
              7,4% - 28 716 ms - 4 024 486 inv. java.lang.Character.toLowerCase (line:946)
              6,4% - 24 725 ms - 3 420 156 inv. java.lang.Character.toUpperCase (line:963)
              6,2% - 23 975 ms - 3 420 156 inv. java.lang.String.charAt (line: 963)
              0,8% - 3 228 ms - 450 342 inv. java.lang.Character.toLowerCase (line: 974)
              0,8% - 2 932 ms - 450 342 inv. java.lang.String.charAt (line: 974)

5.1.6 works fine (at least for what we do)…

So Manning released the long awaited Ant in Action 2nd edition as of today.

As you can expect from Steve (Loughran), it’s a thorough description of every bit and pieces that can be automated as part of your build with clear practical examples. Besides Ant, it drives you through the use of Ivy to manage dependencies and SmartFrog to handle deployment.

The book alone is worth it for those chapters and it should be the companion book to every engineer in the java world from my point of view.
Now I wish Steve would do an entire book about deployment with SmartFrog: ‘SmartFrog in Action’ and gives lots of details on how they use it at their HP datacenters (hint hint)

For people interested in, there’s Steve’s presentation Distributed Testing with SmartFrog

NB: For the sake of completeness, in the same category of SmartFrog there’s also Puppet if you feel more comfortable having Ruby than Java.

Congratulations to the Apache DS team for the certification. It was announced during ApacheCon US in Austin this week.

It’s nice to see Alex Karasulu’s baby come to life with such certification after all the hard work he has put into it. He initiated this project on Oct 16th 2002 as LDAPd, so it’s nearly a 4-year project. Well done Alex !

My friend Emmanuel Lecharny also spent an insane amount of time working at night for more than 2 years on Apache DS and decided 2 months ago to dedicate to it full time and in a totally benevolent way. Through Emmanuel’s consulting company IKTEK, Pierre Arnaud Marcelot is also doing amazing work as an intern and was recently voted a committer.

If you need some LDAP expertise in France or around. You can contact Emmanuel @ IKTEK.