Clearing or resetting the database in MythMusic is not a feature that is available from the user interface. If you run into troubles with your library or database, there is no function to reset, clear or recover the database. MythMusic is a plugin for MythTV that plays….music. It stores your music library in a database and users can update the library by doing a scan through the Utilities/Setup > Music Tools > Scan for New Music menu. I ran into problems when I moved my music collection from its custom location (/home/username/music) back to the default location (/var/lib/mythtv/music).

The library was showing some strange behaviour (such as displaying tracks that were no longer in the library), and there is no way in MythMusic to delete tracks from the library. The only way to update the library is through the “Scan for New Music” menu. After much Googling, I eventually learned that new files are apparently detected when doing a scan by comparing the timestamp on the music files to those already in the database. So, to update the database, I thought I would just set the timestamp of all the files in my Music Library to “now”.

Of course, there is a brief command in Linux to achieve such a thing, (recursively “touching” files with the current time) which I found on another blog to be;

find . -print0 | xargs -r0 touch

After running that command (which took less than a second for 30GB in more than 5000 tracks) and choosing the “Scan For New Music” option in MythTV, my library had been updated and the glitches were gone.

Update 10/08/09:

I read the latest issue of Full Circle Magazine, and there was an article on “Top 5 Screenshot Tools”. Desktop Data Manager was mentioned. Hopeful that the previous bugs that stopped me using it had been fixed, I installed from the .deb on SourceForge.

Unfortunately, I was disappointed again. The application is unusable, it crashes on opening the application. Various other users have reported exactly the same bug on SourceForge. Try again this time next year I guess!

In the meantime, I’ll have a look at Shutter, and I am happily using Screengrab! (a Firefox addon to capture web page screenshots).

While I was studying for some exams, and typing up exam notes and trying to insert screenshots of lecture slides into my revision notes, I found myself wishing for the type of functionality provided by software such as Microsoft’s OneNote.The default GNOME screenshot application is useful under very particular circumstances, but if you just want to copy a screenshot to the clipboard, you’re SOL.

OneNote has an applet running in the system tray, and on pressing the keyboard shortcut, allows the user to select any area on screen and copy it to the clipboard. This image can then be pasted into word processor documents, image editing programs etc etc.

Just a couple of minutes searching located Desktop Data Manager, a program that can do this, and more. Functionality includes;

• clipboard history
• converts HTML / FTP links to images that can be pasted into any app
• customisable keyboard shortcuts
• take screenshots of windows, subwindows, or user-selected area

It really is very useful and far better than the horrible little GNOME app that you get with the default desktop. It seems to have ongoing development, and there are pre-built debs and rpms or you can build from source. There is even a 64-bit deb. Very cool.

Download Desktop Data Manager here.

Update: After using it for a few days, I found it has some issues. It causes repeatable crashes when cutting / pasting files in Nautilus. Also, it crashes repeatedly whenever opening the Desktop Data Download app. I can’t really recommend it any more with these basic bugs, it needs more work.

At work we use Microsoft Windows Sharepoint Services (WSS) for our intranet. It’s pretty good free software if you already have Windows Server and works well in MS land. It integrates really nicely with Office (if you have the latest version), but if you are two versions behind, you lose out on a lot of features.

The architecture of our setup is a little complicated; we have two web servers in a Network Load Balanced web farm, a separate database server, and Kerberos authentication to Active Directory. Everything has been running swimmingly for the last few years, but recently we noticed an annoying problem.

Whenever a user searches the intranet, WSS returns “No results found”. Investigating the Event Viewer on the web server shows an error whenever the search service is meant to be indexing.

This is the infamous Error 2436 (well, infamous if you’ve done any searching for WSS search errors). This error is quite misleading, as it sends you on a wild goose-chase of checking user accounts, permissions, passwords and so on. It turns out that there are many possible fixes for this, but none of them worked in our situation.

Various fixes include;

• DisableLoopbackCheck
• Setting the authentication of the site collection to NTLM instead of Kerberos

However, none of them worked.

Eventually I noticed a Kerberos error in the Event Viewer that roughly corresponded to the error above.

The error shows that there is a missing Service Principal Name (SPN) for the server name shown (pixelized for anonymity). The account that search runs on therefore needs an SPN set for the http service and the corresponding server name shown in the event viewer. The required SPN is in the form “http/intranet.domain.com”.

I found a really great tool called DelegConfig that helps with the arduous task of viewing, deleting and setting SPNs which you can find here;

http://blogs.iis.net/bretb/archive/2008/03/27/How-to-Use-DelegConfig.aspx

A minute later I had set the required SPN, kicked off a new search and I finally had results showing on intranet searches!

It looks like Firefox 3 is setting all kinds of download records. Download your copy and add to the total to help set the record for most software downloaded in a 24-hour period.

If you have installed a few linux kernel upgrades, your GRUB menu will become cluttered, with 2 boot options for each kernel, it can quickly become messy. It is easy to remedy though, just uninstall the kernels you don’t need.

For example, Hardy’s current kernel is 2.6.24-18-generic. I also have the 2.6.24-17 and 2.6.24-16 kernels installed, so I have a total of 6 linux entries in my GRUB menu. I have been using the -18 kernel for a while and know that it works fine, so I just uninstall the 16 and 17 kernels. To do this, search for “linux-image” in Synaptic.

You can see the 16 and 17 kernels listed. Just right-click on each one and select “Mark for Removal”. You will then get prompted to mark additional required changes relating to removing that kernel image.

Finish off by clicking “Apply” in Synaptic to finish off the kernel removal. When you next reboot, your GRUB menu will be nice and clean.