I like to be able to search for stuff on my computer – who doesn’t? Well , I don’t enjoy searching, but I do want the ability to do so in case I loose track of something in the perfect hierarchical world that is my directory tree. From the CLI, one can invoke various search commands to do the dirty work. Even ls can find stuff for you if you grep it right. But, alas; search for me is better suited to the gui, so about two weeks ago, I got the strigi daemon indexing so I can actually do the aforementioned searches – after just a bit of drive crawling, I could find media, email messages, etc., as expected. I had some problems displaying found items when clicking on the links to them in the search results data, but at least I could see where the items are and so just go to them manually. Out of the box, strigi seemed to be working fine.
Now, I’ve got around 500GB of disk space between 4 drives with more than half of that currently free. My root partition is mounted on a 30GB disk and then media and other data are stored on other larger disks. 30GB for / is plenty of room, or so I thought. Last week I started getting error messages that my home directory was running out of free space. I dumped the trash, cleaned up some temp files, and figured that would be the end of it. The message popped up again a few more times telling me that I was quickly running out of free space, so I decided to see what was eating a drive that should be more than adequate for my root partition.
I use Konqueror for my file browser. I have 11 different views that I can choose from, one of which is the File Size View (see the screenshot). After cranking away at / for about 5 minutes, the disk hoarding culprit was soon found: the ~/.strigi directory â€“ it was using 24GB of a 30GB disk â€“ no wonder I was running out of room. I uninstalled strigi and removed the clucene directory, with 98 .cfs files ranging in size from 1.6GB to just 98k, from within .strigi and soon had my 24GB back.
I’m not that hard up for disk space, but 24GB is just too much storage to dedicate to a search daemon that claims to be lightweight. I am glad that I never “upgraded” to KDE 4.x as strigi is an integral part of the desktop and I would not have been able to just stop using it and delete the index. On the other hand, I have to hope that the version of strigi used in KDE 4.x is a bit more resource friendly.