If you followed my recent video tutorial on enabling Ubuntu/GNOME’s Tracker search tool, you should now be enjoying the ability to search the files on your system.

During a recent upgrade to Ubuntu 9.04, which was recently released, I found that the Tracker search tool index had become corrupted. I tried to tell Tracker to rebuild the index, but experienced some weird issues.

In this quick tip, I’ll show you how to avoid this and other issues by deleting Tracker’s index and cache manually, so that it can start ‘from scratch’. Note that if you are having issues with tracker, it is worth trying to rebuild the index from the graphical interface first. This tip is a last resort if you need to clear everything out from scratch and start it again.

Tracker’s Files

According to the documentation, Tracker stores its files in the following locations:

• Configuration Files – ~/.config/tracker/tracker.cfg
• Data Files – ~/.local/share/tracker
• Index Cache – ~/.cache/tracker

If you’re happy with the settings you have, you may want to leave the configuration file intact, and just wipe out the data files and index cache.

Quit Tracker and Delete the Files

First of all, you need to go ahead and quit the Tracker application in the system tray.

Next, you must quit all processes of both trackerd and tracker-indexer, so open up a terminal and run these commands:

And finally, carefully delete the files listed above (excluding the config file, if you don’t think you need to clear that out as well).

So to recap, go ahead and delete the folders:

• ~/.local/share/tracker
• ~/.cache/tracker

and if you wish, the config files:

• ~/.config/tracker/tracker.cfg
• ~/.config/tracker/tracker-applet.cfg

To restart Tracker, I recommend that you log out and log back in so that trackerd and the system tray applet both restart properly. You should then be able to reconfigure it from the beginning as detailed in the original tutorial.

And that should be it – Tracker should rebuild itself from scratch. Useful to know for if things go wrong!

Locating files that you know exist, but you can't remember where they are can be frustrating. There are a handful of GUI-based search tools for Linux and Unix, but if you prefer to work in the command line, or don't have one of these special apps installed, you can often fall back on command line tool locate.

The locate command works by querying a database of your files. It won't be very useful, unfortunately, unless this database is kept up to date. What I recommend you do is set up a cron job to run the /usr/bin/updatedb command (as root) every day/night so that the locate tool becomes useful.

Some distributions have taken the liberty of scheduling updatedb for you already, so you might find this step unnecessary and you can get straight to the searching.

Once you have the database in place, locating a file is as simple as this:

$ locate alostfile

If there are no results, you will be just sent back to a prompt, and you will need to widen or refine your search. Any results will be spat back out as a list of the paths to the files that it has found.

This method of searching works best when you do actually know the filename, or part of it, and as far as I'm aware doesn't look in the contents of files to search.

Nevertheless, if you have lost that important file and don't have an alternative desktop search program already installed, give locate a try.