Laptops are the worst for this, since there are more vendors turning out more new crap…. For example, my HP Pavailion dv6-1230us; the HDMI port is tempermental, and the sound card fails to run at all in Ubuntu 9.04. I’m currently upgrading to the Beta 9.10 with high hopes, but I’ll be sure to let you know. Another feature this machine happened to have is LightScribe–a feature I just assumed would not be usable. On a whim, I googled “ubuntu lightscribe” and was thrilled to discover a page detailing the 5 easy steps to get it running on Ubuntu. It’s a pretty stripped, simple interface, but hey–it works!

You can find the debs or rpms pretty easily on the lightscribe site if you need them for another distro.

Caught in the balanace of wanting to get something good and wanting to just get something soon, I followed a recommendation on a forum to try out trac. Trac is an open source, “Integrated Project and SCM Management” system. After digging around in a couple of demos, and reading a few reviews, I decided to jump in and give it a shot. The features that sold me: 1)the entire thing functions like a wiki; there is an explict wiki built in, but then all milestones, tickets, and revisions (see next feature) can be written in wiki formatting and accessed via internal wiki links. 2) Subversion and git integration; you can review the timeline, and see change sets or entire source files right in the browser! 3) plug-in extensibility; I immediately added plugins for finer-grained access control, and account management. I also tried out a time estimation plugin, but decided that it has a ways to go before it is truely useful.

A couple months into it, I’d say it’s been well worth while. The interface has been well done, including the ability to pull ticket reports based on which side of the milestone progress bar you click (very cool), and it’s saved us a lot of headaches by concentrating and connecting our internal documentation with our tickets and our source code. A couple of things I would love to see in the future are out-of-the-box ldap support and the ability to nest tickets. I really can’t complain though, given that it beats out a lot of pricey solutions, and it took next to no time to connect our geographically distributed teams up. Quite a few well-known FOSS projects have adopted it, including Piwik, the Wordpress Plugin Directory, and Browsershots (there is a decent list on the trac site).

The article mentions the 80/20 rule of feature use, and that got me to thinking of something I’d seen in the Piwik plug-ins directory: clickheat. I looked it up out of curiosity, and not ten minutes later I had a copy running, logging clicks on my development system. Our system is resource based (having several hundered individually assigned/authorized resources), and each user is presented with a menu tailored to the features that they have access to. That gives us a pretty good edge on keeping things simple, but even so I know that there are well-intentioned features that can get in the way; as such I’m not about to turn down any intelligence available to me.

Clickheat–like it sounds–logs all the clicks on whatever page you add it to, and then does some geometric math to overlay a heat pattern on a copy of the webpage; the result? A graphical representaion of what users are doing on any given page, overlaid on the actual page that clickheat fetches with javascript.

A "thermal" map of clicks overlaid on the page

To be perfectly candid, it’s not the easiest thing in the world to figure out, but it sure wasn’t bad. You have to do a couple of simple things (like setting a site name in the javascript tag before pasting it into your html, and deciding how to group pages for logging) and the documentation leaves much to be desired, but all in all it was worth a few moments that it took to install. It will be great when this tool matures a little and possibly adds a database back-end (right now it’s all file logging), but in the mean time it will still provide interesting use information to help us reduce the effort for our users to find and access what they really want: that golden 20.

If you already understand the value of analytics, then you likely also know that Google is the modus operandi for the majority of webmins: mostly due to the lack of viable alternatives. Now, I’m all about the innovation, convenience, and especially anti-Microsoftism that Google brings to bear. That being said, they have more information about more people….and “Don’t be evil?” Really?

Enter a new option, Piwik. Piwik offers all of the basics (referrers, search engines with keywords, page counts, etc.), but what’s more, it’s all open source MVC. In other words, anything it lacks, you can add. It has an easy modules and hooks design so that your MVC plug-ins can be displayed in the eye-candy, flash-based graphing widgets–it took me all of five minutes to drop in the GeoIP plugin. The API is very clean for dropping calls into other apps as well. Installation couldn’t be easier, requiring only a database user and php5 with the mysql module.

I’m now using Piwik on pretty much all the pages I host and a few others I have a stake in, or as favors for friends, so I know what you’re looking at =). Don’t worry, I’m just using it to improve things.

The issue was actually a rather simple one, involving only a corrupted ext3 journal in the /var partition; that unfortunetly becomes much more complicated when the offending host is located some 350 miles away in a dark basement’s much darker closet with no known *nix admins to be had. I tried to do a phone walkthough with my very patient and forgiving brother-in-law, but that ended when I failed to assert the gravitous difference between

tune2fs -O ^has_journal /var

and

tune2fs -O ^has_journal / var.

If you’re not a sysamdin, the short story is that the mis-placement of a single space made the machine completely unbootable. So thanks to FEDEX, we’re back in business little the worse for the wear.

Aside from hosting a few blogs and our primary DNS, that server was not doing much anyway, and I was in the process of backing up the data to decommission it when the error occurred. I never would have guessed it a year ago, but we are now 100% Ubuntu as that was our last Fedora based server. We just upgraded most of our desktops at work to the new Beta release of Hardy Heron, and I’m happy as cake. The NVidia drivers for my GeForce 6150 LE are finally solid in Compiz-Fusion after endless screen locks (the mouse would move, but no one was home) and the intel video cards are doing their own work so that the processor is freed up for snappy performance. The background art is the best yet.

I apologize to the faithful few who keep checking in hoping for personal news and continue to get assualted with tech jargon.. I have plans to split my personal and professional blogs apart–stop laughing, I know that keeping has been more than a task for me so far. Thanks for sticking with me anyway. Blessings!

It takes all of about 10 minutes to set up, and about 12-72 hours to download the packages. Once it’s done, your updates and installs will happen in seconds, since the downloads are only limited by your local LAN. With my sources enabled (all of the official and community repos on the ubuntu mirror plus the wineHQ one) it was a 37 gig download. Assuming you have a running web server, all it takes is:

sudo apt-get install apt-mirror

After the install, open up your mirror.list and replace the lines that start with deb with the similar ones in your sources.list. here is mine:

sudo vim /etc/apt/mirror.list

############# config ##################
#
# set base_path /var/spool/apt-mirror
#
# if you change the base path you must create the directories below with write privlages
#
# set mirror_path $base_path/mirror
# set skel_path $base_path/skel
# set var_path $base_path/var
# set cleanscript $var_path/clean.sh
# set defaultarch
set nthreads 20
set tilde 0
#
############# end config ##############

deb http://us.archive.ubuntu.com/ubuntu/ feisty main restricted
deb-src http://us.archive.ubuntu.com/ubuntu/ feisty main restricted

deb http://us.archive.ubuntu.com/ubuntu/ feisty-updates main restricted
deb-src http://us.archive.ubuntu.com/ubuntu/ feisty-updates main restricted

deb http://us.archive.ubuntu.com/ubuntu/ feisty universe
deb-src http://us.archive.ubuntu.com/ubuntu/ feisty universe

deb http://us.archive.ubuntu.com/ubuntu/ feisty multiverse
deb-src http://us.archive.ubuntu.com/ubuntu/ feisty multiverse

deb http://security.ubuntu.com/ubuntu feisty-security main restricted
deb-src http://security.ubuntu.com/ubuntu feisty-security main restricted
deb http://security.ubuntu.com/ubuntu feisty-security universe
deb-src http://security.ubuntu.com/ubuntu feisty-security universe
deb http://security.ubuntu.com/ubuntu feisty-security multiverse
deb-src http://security.ubuntu.com/ubuntu feisty-security multiverse

clean http://us.archive.ubuntu.com/ubuntu

After that you just need to run the mirror updating command…
sudo su - apt-mirror -c apt-mirror
…and wait for several hours. Or days.

Once it has completed the first run, you can go into /etc/cron.d/apt-mirror and uncomment the line in there to make it run on it’s own every day.

With the files in hand, I opened up my httpd.conf to replace my self-signed certs with the real deal. I made the changes and restarted the server, only to find that the old certificate was still being used. At first I thought that the issue was in reloading the .conf file, but that proved not to be true after a reboot. I started googling to see what could be wrong, only to learn that name-based virtual hosting and SSL do not mix. Apparently the SSL handshake occurs before the http headers with the hostname (which is how Apache knows what page to serve you) so that you can only use one SSL cert per IP address.

I made myself another latte and sat down to figure the damage on switching everything over to ip-based hosting. Running Linux makes it trivial to use multiple ip addresses on one network card (I don’t know of any limit), but I was afraid I would run short on leased ips. After scanning half a dozen sites with bits and pieces about ip virtual on Apache, it dawned on me that I might be able to run both name-based and IP-based on the same daemon. I’m sure why I assumed I couldn’t (maybe because almost every howto contains something to the effect of “Name-Based vs. IP-Based” in the title), but the revelation was a cheering thought.

After backing up my old httpd.conf, I changed the virtual server in question to a IP based one (leaving the others in tact). A quick change in the DNS A records and I was ready to test it out. To my delight, everything worked as it should. Anyone who can read this–and comprehend it–knows that it’s never that easy, but lucky for me it was.

The power and flexibility of the LAMP stack continues to amaze me. No matter how messy the situation is, this winning combination will deliver in spades. It is no wonder that so many organizations use it.