Progressive Computing

I really try to avoid doing any type of windows support. However, when you wife comes to you and asks you to please take a look at her mother’s laptop I really can’t say no.

It turns up that all of a sudden her hard drive has totally filled up. After the usual initial suspects (virus/malware, etc) I tracked it down to Google Earth. Apparently there is a bug out there (that Google can’t seem to replicate), that triggers the caching routines in V7 of GE to go haywire and fill up all usable space with gobs of small files.

I went to the cache tab of the settings pop-up, chose to delete the cache files (all 420GB in my case), and then uninstalled Google Earth. Hopefully problem solved.

Some people are theorizing that if GE is open when the computer goes into hibernation one of the config files gets corrupted, which triggers the caching bug.

You can read the discussions on Google Groups about the problem.

Hope this helps someone.

We have a a client who uses CenturyLink that recently relocated. The new office had the CenturyLink ‘Fiber Optic’ DSL, which required a purchase of a new Actiontec Q1000A DSL Router.

Speedwise it is a great improvement. However we’ve run into a couple snafu’s with the port forwarding. Namely:

• once you get to eight entries in the ‘Application Forwarding’ screen you can’t add any more
• we are unable to delete any of the existing rules from either the ‘Application Forwarding’ or ‘Port Forwarding’ screens

Plus, the web-ui seems to lock up from time to time (the router still works fine though), which requires a reboot to correct the problem.

So now that we’ve maxed out our eight entries, we can’t delete one of the existing/older ones to make a slot for a new one.

The only way I’ve found to remedy the problem is to do a factory reset, but then we’ve got to reconfigure all the wireless and port forwarding entries.

We’ll see what CenturyLink tech support has to say about it….

I’m speechless – I read about this new senate bill today and am wondering what is going through their minds. This bill was originally supposed to increase our email privacy.

Typically I run Debian’s stable distribution (currently code named ‘squeeze’) on all my systems. The great thing about Debian is it’s stability. I’ve been burned by Ubuntu a couple of times, and I really don’t like their custom non-standard windowing system (name Unity), but I digress….

The problem is that this stability has a price – namely that their releases come few and far between, and when you are at the end of their release cycle many of your software packages are outdated. Such was the case I ran into over the last week. On two different software projects I hit a brick wall because the version of Ghostscript in Squeeze (8.7) didn’t have some of the features I needed in one case, and in the other it was puking while processing PDF files created by a new copier/scanner that one of my other clients just started using.

While the upgrade wasn’t as straight forward as some in the past, I was able to follow these instructions without too many problem. The only issues that I ran into were:

At the end of step 4.4.6 I couldn’t install udev due to some dependency conflicts. At this point I was able to reboot to bring the system up under the new kernel. Unfortunately, since I was running the custom NVIDIA Xorg driver, my gdm/xwindows quit working. I stopped gdm (which was looping and consuming resources), and continued the installation in a console (which is what I probably should have been doing in the first place.

Immediately after that, when I ran ‘apt-get dist-upgrade’, dpkg borked with the following errors related to the gnome-watchdog package:

dpkg: error: parsing file '/var/lib/dpkg/status' near line 66623 package 'gnome- watchdog': blank line in value of field 'Description' Log ended: 2012-11-07 06:25:29

which indicated that there was a blank line in a dpkg database file that shouldn’t be there. I had to edit that file, along with /var/lib/dpkg/available to correct the problem, after which I was able to run ‘apt-get dist-upgrade’ again.

After everything was completed I deleted the /etc/X11/xorg.conf file, and I was able to get back into xwindows (since it was able to autoconfigure everything with the NOUVEAU driver), and now my screen and fonts look much better by the way.

And most importantly, it solved both problems I was having with Ghostscript (updated version now 9.05), which for the record were:

• using lout to create a postscript file with internal crosslinks, and used ps2pdf, when you clicked on the links in the PDF file they wouldn’t actually do anything.

• when using ghostscript to burst a PDF created by a newfangled scanner/copier (whose Author tag was ‘LateralData LP’ and Creator and Producer tags were ‘Viewpoint’) it would throw one of the usually cryptic errors that Ghostscript is known for. I’m pretty sure it was due to some new twist on the image compression schemes.

Anyway – hope this helps someone out there in the future.

Introduction

More than a few of my clients have asked about QR Codes, so I thought that I should take the time to explain what they are and how to use them.

In a nutshell, a QR Code is a barcoded image of a web site address. There is no software embedded in it or anything like that. It’s just an image.

Once you have one of these images somewhere, people with smartphones can run their barcode scanning application, and using the camera on their phone they can scan the barcode, which will more than likely redirect them to somewhere on the internet. Check your phone’s app store – there should be more than one available.

For example, you can put a QR Code on the back of your business card, so that if someone takes your card they can later scan it and go right to your web site. If you put the QR Code on your web site then users can scan it with their phone and bookmark it there in case they are browsing the internet on a computer that is not their own. Realtors are now putting them on their yard signs, so visitors can scan the barcode and get more detailed information on the listing on their phone.

How to Create A QR Code

The easiest way to create a QR Code is to use one of the free web sites, such as this one. Just type in the URL you want the QR Code to redirect to, choose the size of the image, and then click the create button. Once the image is on your screen you can right click on it and choose ‘Save Image As…’ to save the image to you computer. You can then use it in your word processor or pass it along to your web designer.

Advanced Uses

There are other uses for QR Codes, such as instead of a web site address you can encode an address book entry, so when someone scans it with a smartphone it adds your info to their address book. But for 90% of the people out there, it’s just a way to easily publish a web site address.

In preparation for a new project that needs some NoSQL components, I was installing and configuring Redis on our main server. Since we’re running Debian and installing this particular package from source, I needed a startup script and came across this one. The nice thing about it is that it uses the redis-cli to connect to the server instead of using signals to get the server to shut down.

Redis 2.6 has just been release, which include LUA scripting.

If you were running Redis 2.4 and are upgrading to 2.6 there are a number of parameters in the config file that are no longer supported, so be prepared for some tweaking.

One of our clients decided to switch over from their traditional phone system to an Asterisk voip platform. We initially ran into problems with too much jitter on the CenturyLink business DSL line (which was causing a lot of echo and drop outs), so we built a dedicated Asterisk server for inside the office and added a four port FXO card to plug in two copper lines to it.

After doing some research I decided to give the Rhino 4 port card a try. I had thought that it would compatible with the Digium cards, but that was not the case. It took a little bit of poking and prodding, but the steps needed to get this up running under Debian Wheezy wound up being:

• install asterisk and dahdi-linux via apt-get as normal
• apt-get source dahdi-linux
• cd to the directory and run ‘make’
• download the latest rhino modules
• unpack them
• in the Makefile set DAHDI_DIR= the directory where you unpacked the dahdi-linux package
• make
• make install
• run dahdi-config

Briefly, during the build of the dahdi-linux source there is a file created (I think it was version.h) that the rhino package needs in order to build. Once those modules are installed you will more than likely need to reboot the server. After that you should be able to follow the documentation and configure asterisk as normal.

When we originally migrated our sites up to the Linode data center in Dallas last year I ran across an issue with PHP displaying dates properly. Even though everything was set correctly on the server (hardware clock set to UTC, locale set properly to America/Phoenix, and the date command showing the proper time), the PHP date functions were still using the wrong time zone. The fix for this is to either set the time zone in php.ini, or to call the date_set_time_zone function.

Rescently when we spun up a new WordPress instance we noticed that the above fixes weren’t correcting the problem. It turns out that there is a setting in WordPress that overrides everything, and out of the box it is set to UTC. To fix it just adjust line 37 in wp-settings.php to the correct value.

// Set default timezone in PHP 5.
if ( function_exists( 'date_default_timezone_set' ) )
    date_default_timezone_set( 'America/Phoenix' );

While setting up a new access point for a client I came across a problem that in retrospect I think I ran into a few years ago (but must have had the information paged out of active memory).

I set up a Linksys E1500 as an access point (they already had an actiontec router) using WPA2/WPA Personal encryption. After getting the access point on their network with a fixed unique IP address I tested connectivity with my Droid phone along with my laptop using both Linux and WinXP with no problems.

A half an hour later the client called back saying they couldn’t connect with their laptop running Vista. Sure enough, when I got back on site and poked at the laptop I couldnt get it to connect using various configuration methods.  The laptop would connect to the network, pick up an IP addtess and other network confiuration info via DHCP, but could not ping or connect to anything else.

After trial and error (and noticing that the clients home WIFI network was using WEP), it turns out the only way to get everything working was to switch to WEP. After that everything worked like a charm.

I guess I will rack this up as yet another issue that drove me nuts with Vista

Today I was cutting one of our client’s development wordpress sites over to production and ran into a couple of tweaks that were needed.

Mainly, I wanted to get Munin hooked up to the new server so we could monitor system load, disk activity, mysql activity, and the apache statistics. However, when I went to set up the Apache and Mysql plugins there were a few issues that needed to be tweaked.

First, the .htaccess file for WordPress was causing the /server-status/ url to get hijacked by WP prior to getting processed by the Apache Status Module. Based on this post I needed to adjust the .htaccess file to ignore the status requests, and wound up with:

# BEGIN WordPress
<IfModule mod_rewrite.c>
RewriteEngine On
RewriteBase /
RewriteRule ^index\.php$ - [L] 
RewriteCond %{REQUEST_URI} !=/server-status/?auto
RewriteCond %{REQUEST_URI} !=/server-status
RewriteCond %{REQUEST_URI} !=/server-status/
RewriteCond %{REQUEST_FILENAME} !-f 
RewriteCond %{REQUEST_FILENAME} !-d 
RewriteRule . /index.php [L] 
</IfModule>
# END WordPress

Next, when the munin apache modules where run an error like the following would occur:

Can’t locate object method “new” via package “LWP::UserAgent”

the solution based on this post was to install the libwww-perl package, which provides those methods. This was able to get the Munin Apache Plugins working.

As for Mysql – based on this post I had to install the libcache-cache-perl and libipc-sharelite-perl packages to meet the missing dependencies.

Now it’s just a matter of waiting for the hits to come in and stare at the graphs for a while.