Hello,

I let baloo index my home folder excluding hidden files and folders.
Then I ran fdupes -r --nohidden on top of my home folder and deleted bunch of duplicates, like 150 000 files.
Baloo responded by removing data about of 900 files (watching the Total files indexed value and it took three days).
I asked balooctl to do the check action what indexed a few (61) new files and not removed any data about those deleted files.
I do not want to purge the index and start from scratch because my home folder is hosted on encrypted LUKS hosted on SW RAID 1 present on rotating hard drives
and it took weeks to finish it and desktop responsibility was close to zero while indexing ran.

Baloo is: kde-frameworks/baloo-5.81.0, KDE is compiled from sources, distribution Gentoo.
Total files indexed: 552 329, the size of the index file is 32,56 GiB.

Could you please suggest how to tell baloo to check for deleted files and to remove their data from the index file?
Using balooctl clear on each of the deleted files would take ages.

Thank you.

I purged the index, ln -s ed baloo folder to a SSD stored folder and started the indexing again.
When at like 2/3 of indexing has been done I made a mistake of removing like 50.000 files in a folder of a httrack mirrored website.
That slowed down indexing rapidly. And Baloo was not able to purge all deleted files from the index file till now, even the count of "Total files indexed" stopped to lesser after two days,
now at 484.141. It purged around 15.000 files from 50.000. Maybe it not indexed all the files in the folder before I deleted them luckilly.

In addition, it seems to me that Baloo does not save Inotify(?) data over session log off events. So it is lost when the machine is rebooted/the session is logged off.
Am I wrong?

I think that I will open a bug report for both.

After files moved between folders or renamed, baloo follows and writes noted changes to the index file and documents them in the .xsession-errors file too.
During one KDE session those files are not re-indexed, 'balooctl status' shows 0 files to be indexed.
But after KDE session log-off and log-on it re-indexes all of them.
Why? Rename in the index file should keep the file data.

I will try baloo again in one year or so.
It now went from "Files remaining to index: 2" to "kf.baloo.engine: dbis is invalid", then to "Checking for unindexed files" after a session log-off and log-on
and then to "Files waiting for content indexing: 454 741" at last.
Index size reported is 40,85 GiB, "Total files indexed: 487 345".
It is fragile too much.
I saw no (disaster) recovery possibilities documented, no index file checker, no index file content exporter.
I have two last versions of the index file available but do not know if the problem resulting in "kf.baloo.engine: dbis is invalid" is there already present or not.