Version: (using KDE 4.3.1) OS: Linux Installed from: Unlisted Binary Package An my machine, when a larger number of files get copied somewhere into my home directory all at once, strigi kind of takes over all of my Laptop's CPU and disk usage for a minute or two in order to index them, and it becomes really cumbersome to do anything with the Computer until strigi is done with the new files. I propose having a throttling setting for strigi in the Nepomuk configuration module, with which you can force it to be less aggressive when indexing (e.g., run with a higher nice value, sleep for a second after every so many files, etc...) Also, maybe the right-click-menu of the strigi system tray icon, where you can currently choose "Suspend strigi indexing", could then be extended to offer the choice between "1) Allow indexing at full speed" "2) Allow slow indexing in the background" "3) Suspend strigi indexing" That way, the user has full control to adapt the operation of strigi to whatever he's currently doing at the computer. In addition to this, the Nepomuk configuration module could then also offer the following settings: "[x] Enable 'slow indexing in the background' mode on startup" "In slow indexing mode, throttle to: [0%----|----------100%]" "[x] Allow automatic temporary switching to full indexing speed while the Computer is idle (i.e. no mouse/keyboard activity for several minutes)" What do you Nepomuk/strigi developers think?
Okay, I only just now realized there are other similar bug reports already (they were just differently phrased... sorry...): https://bugs.kde.org/show_bug.cgi?id=195233 https://bugs.kde.org/show_bug.cgi?id=196402 (with technical discussion about how to realize this throttling) http://article.gmane.org/gmane.comp.kde.nepomuk/396 Still, my feature request mentions more specific ideas for making this accessible through the user interface, so maybe it does have some merit on it's onw... ;-)
I think this is a good idea. It would, however, be great to have some help with it. :P
With KDE SC 4.10, the indexing has been split into multiple parts. It no longer consumes a lot of CPU. Also, the intensive file indexing only happens when you're idle. With the 4.11 release the situation has been improved even more by providing better scheduling code and indexing quietly in the background. I'm marking this bug as FIXED as we now automatically throttle the indexer based on multiple factors - Number of Files - If on Battery