Hello guys,

it seems Nepomuk/Strigi is not configurable for me. I am running Nepomuk-4.4.4 with Strigi-0.7.2. I've configured strigi to index a simple folder with about 150 files, but it tells me about ~70000 files in its index. And it keeps indexing more when I run strigi.

When I start the strigiclient instead I get a window with a complete different set of folders that will be indexed. Allthough I can modify those also (well, remove them) and enter the only folder I want to be indexed, this setting is completely gone when restarting the application. However, the list of inexed files is always empty in strigiclient.

The related files in my home directory (.strigi/* or .local/share/akonadi/* or .config/akonadi/*) have proper permissions.

So, what is strigi indexing, where it is nocifgured and how can I verify this properly?

- Jörg

Can you please provide the content of ~/.kde4/share/config/nepomukstrigirc?

This looks right:

Code: Select all

~ $ cat .kde4/share/config/nepomukstrigirc
[General]
exclude filters=*.vdi,tmp/,target/,.*/,.*,*~,*.part
exclude folders[$e]=
first run=false
folders[$e]=$HOME/Desktop,$HOME/docs
index hidden folders=false
index newly mounted=false


Can you verify that Nepomuk is indexing files outside the directories you have specified? lsof should reveal this information.

When I let strigi run, it tells me constantly that it is "indexing a file in folder", but it does not tell in which folder. After waiting a coubple of minutes I can detect now a constant slow decrease in the number of indexed files. So it seems that strigi tries to "recover" from running it with its default settings. At least lsof does not show any stange activity.

What's the best way to start over with the strigi index? Can I stop it, delete some directory and let strigi build its index from scratch?

If you do not use Nepomuk ( ie. tags in Dolphin or KAddressBook ) then you can simply remove the Nepomuk database with it not running. In this case however, it may just be safe to let it clean itself up ( which is what it is doing )

OK, it seems nepomuk-strigi is idle now and has left 742 files in the index.

Code: Select all

$ find Desktop/ docs/ -type f | wc -l
713


Any idea where the difference might come from? Additionally it claims the storage in use is (still) 1.8 GB. Is it possible to free the unused area?

Unfortunately I don't know where the difference comes from. Strigi stores its database in Nepomuk, so the only way to reset the indexing is to erase the Nepomuk database, which is stored in ~/.kde4/share/apps/nepomuk/

OK, even with Google I did also not find any references how to compact the DB. While 1.8GB is not really relevant for the disk space itself, it gets relevant for frequent backups, since the file is typically changed always.

Therefore I followed your recommendation by deleting anything in .kde4/share/apps/nepomuk/repository/main/data/virtuosobackend/

Nepomuk-Strigi-File-Indexer now tells me it had indexed 740 files and takes 64 MB. Again, we have the difference in number of files (now dropping excluded files form the counter):

Code: Select all

$ find Desktop/ docs/ -type f ! -name ".*" | wc -l
707


It seems only a developer hight say where this difference comes from, maybe strigi is counting archives differently; my directories contain some OOo files, which are technically ZIPs after all.