We’re rapidly pushing toward an excellent Ubuntu 10.04 LTS release, and we have made a few improvements in the way your Encrypted Home’s metadata is stored.

If you configured your Encrypted Home with Ubuntu 9.10 (Karmic) or Ubuntu 10.04 (Lucid), then no action is required, — you may stop reading here.

If you’re not sure, and you want to check if you need to read this article, take a look at your /var/lib/ecryptfs directory. If that directory is empty, or it does not exist, you may stop reading here. If that directory has contents, then you may want to continue reading…

Ubuntu 9.04 (Jaunty) Encrypted Home installations stored eCryptfs metadata is located in /var/lib/ecryptfs/$USER. This information is absolutely required to mount your Encrypted Home Directory. Actually, everything in here can be re-created if you wrote down your randomly generated mount passphrase!

Big fat reminder here … please be absolutely certain that you have recorded your mount passphrase, on a piece of paper, stored somewhere safely, separate from your computer! You can retrieve your randomly generated passphrase by running the ecryptfs-unwrap-passphrase utility.

For Ubuntu 9.10 (Karmic), new installs actually put this metadata in /home/.ecryptfs/$USER. This is far more convenient for users who put all of /home on its own partition, or for users who just simply backup all of /home.

I’ve previously written about how to move your metadata out of /var/lib/ecryptfs. Particularly if you’re planning a Lucid upgrade of a system that was originally installed with Jaunty’s Encrypted Home Directory, I strongly recommend that you follow these instructions:




Ubuntu Lucid Lynx is in Feature Freeze, and Alpha3 is right around the corner, releasing next week.

If you’re running Ubuntu 9.04, 9.10, or 10.04, it’s trivial to testdrive Lucid in a virtual machine, without modifying your current installation!

If you’re already running Lucid, congrads! All you need to do is add the testdrive ppa, and install testdrive, and either kvm or virtualbox-ose.

Just pop open a terminal and run:

  sudo add-apt-repository ppa:testdrive/ppa && \
sudo apt-get update && \
sudo apt-get install testdrive

Then you can either run testdrive from the command line, or use the menu, Applications -> System Tools -> Testdrive.

Then, a menu will pop up, with a menu of Lucid daily images. If you have previously downloaded any of these, you should see a timestamp of the cached file. If you run testdrive everyday, your cache will stay up-to-date, and the incremental download will be much faster!

Select one of the offered ISOs, or enter a URL to another one of your choosing, and you should be off and running. Help us make Lucid Lynx the best Ubuntu to date!


Introducing LinuxSearch.org!

February 18, 2010

I have created a handful of Google Custom Searches over the past couple of years. These are really handy ways of narrowing your search to a particular field, or set of websites.

When I first joined the Ubuntu Community, I found it difficult to comprehensively search Launchpad, and the Forums, and all the wiki’s, etc. So I created the Ubuntu Developer Documentation Search.

The leaders of the Ubuntu Documentation Team found it more effective than the search they were previously using, so I created the search used at the top of the official documentation site, help.ubuntu.com.

During the course of my daily development activities as a Canonical Ubuntu Server Developer, Ubuntu Community Member, and Upstream Free Software Maintainer, it is often useful to search for patches, bugs, or documentation in other Linux distributions, such as Debian or Fedora. While I know the available Ubuntu resources very well, I may not necessarily have Gentoo’s bugtracker or Arch’s forums at my fingertips.

Thus, I created LinuxSearch.org!

For each of the seven Linux distributions where I most often find useful developer documentation:

I identified the websites providing their:

  • Documentation, Wikis, Forums, Mailing Lists, Package details/sources, and Bug trackers

Let’s walk through an example…

  1. Search for a term at LinuxSearch.org, such as “ext4 corruption“. The default results are for the entire web, without narrowing the focus. Perhaps you find what you’re looking for.
  2. Next, you can focus your search by choosing one of the 7 distributions, perhaps Fedora. Now, you are only looking at results from Fedora-related websites.
  3. Finally, you can refine your search even further, by only looking at entries from, say Fedora’s Documentation.
  4. Now, if you want to broaden your search once again, you can click any of the links across the top, similar to Google’s home page, searching the Web, Images, Videos, Maps, News, or Shopping. And there’s one bonus that Google doesn’t provide across their top bar — Wikipedia 😉

Or just give it a shot yourself.

There’s also support for a browser plugin. If you click “Install the Browser Search Plugin”, you can add LinuxSearch.org to your search engine tool bar (ctrl-k).

I’m using this tool every day, and it’s very much helping me tread through volumes of disparate Linux documentation resources, track down patches, and correlate bugs. I hope you might find it useful too!


p.s. If there are resources or websites that I missed (as I’m not an expert in any of these distributions besides Ubuntu, please leave a comment below!)

The Ubuntu Server Meeting Minutes for 2010-02-10 have been posted at:



So yesterday was Valentine’s Day, of course. Let’s hope you treated your sweetheart extra sweet. Kim bought me a bottle of 14 year old Oban single malt Scotch whisky. Mmm, peaty.

But February 14, 2010 has been marked on my calendar for the last 6 months as the date of the the Austin Marathon! A couple of brilliant billboards advertised the race in Austin over the last few weeks. One said Love Austin, Run Austin. Another one said Love Hurts. 🙂

I’ve trained with Steve Sisson’s Rogue Training Systems (and Runtex University) in the past. These guys have excellent programs for new runners. They can absolutely get anyone who can run/walk 3 miles to completing a marathon with about 6 months of hard work.

My brother-in-law Josh and I trained for and ran the 2006 Freescale Austin Marathon with Rogue. We decided back in September to do the 2010 Austin Marathon, training ourselves and our friends with the techniques we learned in the past. Here’s our team, at 6am morning of the race, looking sleepy, but fresh…

My personal best marathon time was 4:08 in the 2006 Austin Marathon. I was hoping to beat that time, but simply enjoying the race was the real goal.

For the first time, I actually ran the race with my camera/phone (Palm Pre). This was pretty cool, as I was able to communicate with my wife, and help coordinate the places where she’d meet us along the way. I was also able to snap some pictures along the way.

The race started at 7am, with a temperature of about 40F (4C), a bit chilly for Austin, but still pleasant running conditions. Clear, clear, clear, without a cloud anywhere to be seen.

Some people ran in costume, such as Batman…

Of course, I ran in my I’m Running Ubuntu t-shirt. Let me just say that Ubuntu is live in strong in Austin! I must have gotten 30+ Ubuntu catcalls over the course of the run!

Josh and I first saw our support crew (Kim and Gerri) having run uphill for about 2 miles, but still feeling great!

After another mile or so uphill, we hooked around, and headed back downhill toward downtown Austin.
We were supposed to see Kim again around mile 7-8, but with the roads blocked, she couldn’t make it. About an hour into the race, the temperature had increased a bit, and it was time to peel off a layer. Miles 1-3 were uphill, but 4-8 were nice and downhill.

Miles 8-13 were the toughest of the race from a topography perspective, with lots of rolling hills, up and down. It was getting a bit warmer, and the first hints of exhaustion started setting in, usually on the uphills 😉

Moreover, the density of runners split in half between miles 10 and 11, as the Half Marathoners motored on toward the finish line, leaving us Marathoners to a few more hours of fun.

Up until the 13 mile mark, three of us (Josh, Steph, and myself) ran together, chatting it up. Unfortunately, that wasn’t to be the rest of the race. Josh had some severe leg cramps, dropping behind right around the 13 mile mark. Steph and I picked up the pace a bit (per our race plan) for the next 3 miles or so, but I could tell that I wouldn’t be able to hold that for the rest of the race, unfortunately. So I wished her luck, and we each slipped into our own pace, just trying to make the finish line.

Just after mile 16, I had to smile when I saw a gal with an Easy button 🙂

I think miles 16 and 17 were the hardest two of the race for me. These were along Great Northern Drive, a 2 mile, perfectly straight, perfectly flat, perfectly boring stretch that any technical runner in Austin probably knows. It eats your psyche.

Thankfully, Kim and Gerri were cheering me on at mile 17, and boy did I need it!
I was already a bit off my goal time, so I decided to just enjoy the race. I stopped for a good 60 seconds or so to wish my wife a happy Valentines Day, thank her for being there, and find out how my friends ahead of me were doing.

It was nice to get a bit of a recharge, and I was able to ride that for a mile or two. But around mile 20, I started to feel fatigued. For each of my previous 3 marathons, I ran at least one 22 mile long run before the race. I missed my team’s 22 miler (while I was trekking through New Zealand).

Fortunately, I made a random friend. It’s a great thing to do during a race. Find someone else to chat with. I met a guy named Ed. This was Ed’s 53rd marathon, and 3rd in the last two months! Holy smokes. Ed was also a class of ’82 Aggie (myself being a class of ’01 Aggie). We Howdy’d and Gig’Em’d and Whooped our way through the next few miles. And I tried to learn as much as I could from a guy who had run 52 previous marathons.

I lost Ed during one water stop, and started struggling a bit, until I made it to mile 24, where Kim, Gerri, Steph, Josh who had to drop out of the race due to leg cramps 😦 and my newest nephew Jackson!

Again, I was well out of contention for breaking my best time, so I stopped running, and spent a minute thanking everyone for their support. They encouraged me to hit the road, which I did for the best miles 25 and 26 of any marathon I’ve ever run.

The last two miles were downhill, through the University of Texas. Here, there was a radar speed sign, which correctly identified me as running 6 miles per hour 😉

A few blocks later, then past National Museum of Texas, the Texas History Museum (great exhibits, if you’re ever in Austin looking for something interesting to do)!

And just past the Texas History museum stands the enormous Texas State Capitol in its gleaming pink granite. The finish is near!

After a half lap around the capitol building, the finish line stood two blocks down, at Congress and 10th Avenue. A huge crowd awaited on the south side of the capitol, cheering all the way to the finish. Four hours and thirty minutes later, I had covered twenty six point two miles on foot. I had just completed my fourth marathon. Not my fastest time, but not my slowest time either. But I was proud, happy, and healthy.

And finally, back to Josh’s house where we started oh so long ago that morning for the post race banquet and beers!

What’s next? New Orleans in two weeks, maybe 😉


Byobu Icon Contest!

February 9, 2010

You might have noticed that I took a crack at designing an icon for Byobu a few months back.

While I consider myself a decent hacker, and a reasonable photographer, I’m pretty bad at vector graphics, sadly.

So I’m reaching out for your help! Have you always wanted to contribute to an open source project? Are you pretty good at vector graphics? Here’s your chance 😉

So here are the rules…

  1. Submissions must be in SVG (scalable vector graphics) format. You might want to use the Inkscape package to do your drawing.
  2. Images must be square in overall dimension.
  3. Images must scale well, being recognizable as a 16×16 pixel icon, while also looking sharp at 1200×1200.
  4. Submissions must be made via bzr and Launchpad. You should branch lp:byobu, add your SVG file, commit, and push to Launchpad. These commands might help.
    bzr branch lp:byobu
    cd byobu/icons
    cp /path/to/your/foo.svg .

    bzr add foo.svg
    bzr commit
    bzr push lp:~yourname/byobu/byobu-icon-contest
  5. Submissions must also be contributed under the GPLv3 (same license as Byobu). You can find the GPLv3 on your Ubuntu system in /usr/share/common-licenses/GPL-3.
  6. I really like the idea behind the icon that I did (a three-panel folding screen), it’s just kind of ugly 😉 You are welcome to start with that .SVG and simply improve it. Or start from scratch with a design of your own!
  7. When you’re satisfied with your submission, please use Launchpad’s “Merge Proposal” feature.

You’re also welcome to link to your submissions in the comments below for feedback from other readers!

Assuming we get some good submissions, I’d like to select a few finalists around Feb 25, and then a winner by March 1st. What does the winner receive? Well, besides bragging rights of having designed Byobu‘s logo, I’m going to print a few Byobu t-shirts with the new logo to show off my favorite free software project, and the winner will receive two of these shirts in whatever sizes he or she wants.

Good luck!


I just uploaded a qemu-kvm package that enables KSM by default on Ubuntu Lucid.

KSM is a bacronym, for Kernel SamePage Merging. Previously KSM stood for Kernel Shared Memory. KSM is a new feature of KVM, which can provide more efficient memory utilization on virtualization hosts. Basically, the host kernel tracks identical pages in memory, and stores only one copy when possible. If you’re running several basically identical virtual machines, then you will likely have some identical pages in memory.

Ubuntu inherited these features from upstream with the merge of the Linux 2.6.32 kernel and the qemu-kvm 0.12.2 package. Fedora 12 shipped with the KSM kernel pieces backported to their kernel.

You can disable KSM, if you like, by editing /etc/default/qemu-kvm and then restarting qemu-kvm with sudo restart qemu-kvm.

I did a bit of very rough testing of KSM in a test deployment of Ubuntu Enterprise Cloud. I had 1 Eucalyptus Node, a simple laptop, with a dual-core 2.4GHz and 4GB of memory. I registered a single Ubuntu 9.10 64-bit server image, and started instances with 256MB of memory apiece.

With KSM disabled:

  • Running 0 VMs, the Node’s memory utilization was steady around 12%
  • Running 1 VM, the Node’s memory utilization was steady around 18%
  • Running 14VMs, the Node’s memory utilization spiked and stabilized at 88%

With KSM enabled:

  • Running 0 VMs, the Node’s memory utilization was steady around 12%
  • Running 1 VM, the Node’s memory utilization was steady around 18%
  • Running 14VMs, the Node’s memory utilization was steady at 60%
    • with 18,000 – 20,000 pages shared

It looks to me that KSM “saved” me about 28% of my host’s memory, which is a little over a gigabyte.

If you want to try out KSM on your Lucid host:

  1. Upgrade to the latest qemu-kvm package
  2. Make sure that /sys/kernel/mm/ksm/run is set to 1 (the qemu-kvm upstart job will do this for you now)
  3. Launch two Ubuntu LiveCD ISOs
    • testdrive -u ./lucid-desktop-amd64.iso
    • testdrive -u ./lucid-desktop-amd64.iso
  4. Once they’re running, you can see how many pages are shared with
    • watch cat /sys/kernel/mm/ksm/pages_shared