Bug 493364 - Process seems to hang on some files, indexing content when it should not be
Summary: Process seems to hang on some files, indexing content when it should not be
Status: RESOLVED UNMAINTAINED
Alias: None
Product: frameworks-baloo
Classification: Frameworks and Libraries
Component: Baloo File Daemon (other bugs)
Version First Reported In: 5.115.0
Platform: Other Linux
: NOR normal
Target Milestone: ---
Assignee: baloo-bugs-null
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-09-19 15:54 UTC by pqwoerituytrueiwoq
Modified: 2026-01-31 20:25 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed/Implemented In:
Sentry Crash Report:


Attachments
settings screenshot (69.83 KB, image/png)
2024-09-19 15:54 UTC, pqwoerituytrueiwoq
Details

Note You need to log in before you can comment on or make changes to this bug.
Description pqwoerituytrueiwoq 2024-09-19 15:54:21 UTC
Created attachment 173882 [details]
settings screenshot

SUMMARY

It looks appears to be hanging on some files, ones it should not be trying to index content for, for example
mimetype returned application/octet-stream and kmimetypefinder5 returned status code 1
not this was a 12 byte file and it was just sitting on it

Attempting to load search settings hangs for a while, pausing the serves takes several minutes, maybe 15? i was not counting, i thought the button was broken it was so long

i upgrading from kubuntu 22.04 to 24.04 i noticed my DISK/IO LED was never stopping, baloo was set to do everything thing in my home folder both hidden and content (use case is sometimes i am looking for a config file)

i tired turning off content indexing and rebooting, but it will not stop trying to index content of un-index-able files



STEPS TO REPRODUCE
1. let baloo do it's thing
2. notice disk drive never gets break
3. track usage down to baloo

OBSERVED RESULT
disk usage never stops

EXPECTED RESULT
disk activity should stop

Operating System: Kubuntu 24.04
KDE Plasma Version: 5.27.11
KDE Frameworks Version: 5.115.0
Qt Version: 5.15.13
Kernel Version: 6.8.0-45-generic (64-bit)
Graphics Platform: X11
Processors: 12 × AMD Ryzen 5 3600 6-Core Processor
Memory: 31.3 GiB of RAM
Graphics Processor: AMD Radeon RX 580 Series
Manufacturer: Micro-Star International Co., Ltd.
Product Name: MS-7B79
System Version: 2.0

ADDITIONAL INFORMATION

top - 11:32:51 up 11 min,  1 user,  load average: 1.70, 2.14, 1.46
Tasks:   2 total,   0 running,   2 sleeping,   0 stopped,   0 zombie
%Cpu(s):  1.3 us,  2.0 sy,  1.1 ni, 89.5 id,  6.0 wa,  0.0 hi,  0.0 si,  0.0 st 
MiB Mem :  32034.5 total,  25169.8 free,   4241.8 used,   3162.6 buff/cache     
MiB Swap:  16384.0 total,  16027.2 free,    356.8 used.  27792.7 avail Mem 

    PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND                                                                                                                                                               
   3653 chad      39  19  256.7g 404896 232048 D  15.3   1.2   0:29.46 baloo_file_extr                                                                                                                                                       
   2128 chad      39  19  256.4g  87536  56576 S   0.0   0.3   0:05.41 baloo_file
Comment 1 tagwerk19 2024-09-19 17:56:32 UTC
(In reply to pqwoerituytrueiwoq from comment #0)
> ... upgrading from kubuntu 22.04 to 24.04 i noticed my DISK/IO LED was never stopping ...
There's a very good chance that Baloo is reindexing everything. There's was a patch to fix issues with BTRFS, sometime over a year back:
    https://invent.kde.org/frameworks/baloo/-/merge_requests/131
That changed the way they Baloo tracked the files it indexed, unfortunately after the patch Baloo "did just one more index". There's a description in Bug 475919

> ... baloo was set to do everything thing in my home folder both hidden and content (use case is sometimes i am looking for a config file) ...
You might be indexing a *load* of files you don't want to, everything under .cache for example. Have a look at Bug 434705.

> It looks appears to be hanging on some files, ones it should not be trying to index content for, for example
mimetype returned application/octet-stream and kmimetypefinder5 returned status code 1
not this was a 12 byte file and it was just sitting on it
Not sure here. Did you start with indexing content and then switch to no content? Baloo might have make itself a list of files it needed to index before that change.

> i tired turning off content indexing and rebooting, but it will not stop trying to index content of un-index-able files
That also suggests it is working through a backlog of "things to index"... And the fact that baloo_file_extractor appears in "top"

> Attempting to load search settings hangs for a while, pausing the serves takes several minutes, maybe 15? i was not counting, i thought the button was broken it was so long
I don't know on that one. It feels like Baloo is completely overloaded.

I would kill Baloo - try a "systemctl --user stop kde-baloo" as that works immediately, then purge the index with "balooctl purge" (maybe "balooctl6 purge"), add in .cache (as per Bug 434705) as exclusions, reboot and let Baloo reindex. It should be a lot better behaved without all the backlog.
Comment 2 pqwoerituytrueiwoq 2024-09-19 18:24:52 UTC
(In reply to tagwerk19 from comment #1)

> There's a very good chance that Baloo is reindexing everything. There's was a patch to fix issues with BTRFS, sometime over a year back:
Using EXT4 partitions

> You might be indexing a *load* of files you don't want to, everything under .cache for example. Have a look at Bug 434705.
That is why i turned off content
> Did you start with indexing content and then switch to no content? Baloo might have make itself a list of files it needed to index before that change.
yes, i suspect the same
> Baloo might have make itself a list of files it needed to index before that change.
when i hit play and it started to run is shows alot of files going by then stopped on this one, it should not be hanging on such a small file regardless of what it is doing
-----
$cat -v /path/to/file
B
B
^@^@^@^@^@^@^FM-M
-----

well i think this is a contributing factor (note this drive is 1TB 7200RPM HDD)

$ls -lah .local/share/baloo/
total 5.4G
drwxrwxr-x  2 chad chad 4.0K Sep 19 11:22 .
drwxrwxr-x 62 chad chad 4.0K Sep 19 11:55 ..
-rw-rw-r--  1 chad chad 5.4G Sep 19 11:33 index
-rw-rw-r--  1 chad chad 8.0K Sep 19 11:58 index-lock

maybe there should be a max index size setting or something

i did notice on other thing, but i feel that would fall under feature request, will make a separate issue for that
Comment 3 pqwoerituytrueiwoq 2024-09-19 18:30:54 UTC
To add to that last part i just checked by my backup, this file size was 4.5G before i upgraded to 24.04 on 9/2/24
Comment 4 tagwerk19 2024-09-19 18:33:28 UTC
Sorry, should have been quoted thus:

> It looks appears to be hanging on some files, ones it should not be trying to index content for, for example
> mimetype returned application/octet-stream and kmimetypefinder5 returned status code 1
> not this was a 12 byte file and it was just sitting on it
Not sure here. Did you start with indexing content and then switch to no content? Baloo might have make itself a list of files it needed to index before that change.
Comment 5 pqwoerituytrueiwoq 2024-09-19 18:38:37 UTC
(In reply to tagwerk19 from comment #4)
> Sorry, should have been quoted thus:
> 
> > It looks appears to be hanging on some files, ones it should not be trying to index content for, for example
> > mimetype returned application/octet-stream and kmimetypefinder5 returned status code 1
> > not this was a 12 byte file and it was just sitting on it
> Not sure here. Did you start with indexing content and then switch to no
> content? Baloo might have make itself a list of files it needed to index
> before that change.
that is what i did, i figured it had to finish the one file i was stopped on
----
i have wiped out the index added exclusions, re-enabled content and the LMBD index file is under 0.25GB
Comment 6 tagwerk19 2024-09-19 18:46:50 UTC
(In reply to pqwoerituytrueiwoq from comment #2)
> (In reply to tagwerk19 from comment #1)
> 
> > There's a very good chance that Baloo is reindexing everything. There's was a patch to fix issues with BTRFS, sometime over a year back:
> Using EXT4 partitions
The patch is "generic", it switches Baloo from using the device number to using the FSID of the disc. 

> -----
> $cat -v /path/to/file
> B
> B
> ^@^@^@^@^@^@^FM-M
> -----
Cannot tell much from that :-)

May not be a good idea to attach the file to the issue but you could do a "od -c /path/to/file"
Comment 7 tagwerk19 2024-09-19 18:48:00 UTC
(In reply to pqwoerituytrueiwoq from comment #5)
> ... index file is under 0.25GB ...
Far, far safer :-)
Comment 8 pqwoerituytrueiwoq 2024-09-19 18:48:33 UTC
output of od:
0000000 005102 005102 000000 000000 000000 146406
0000014
Comment 9 pqwoerituytrueiwoq 2024-09-19 18:49:17 UTC
oh you wanted with -c... sorry
0000000   B  \n   B  \n  \0  \0  \0  \0  \0  \0 006 315
0000014
Comment 10 tagwerk19 2024-09-19 21:58:21 UTC
(In reply to pqwoerituytrueiwoq from comment #9)
> oh you wanted with -c... sorry
> 0000000   B  \n   B  \n  \0  \0  \0  \0  \0  \0 006 315
> 0000014
No problem... 

For me Baloo treats the binary as "application/octet-stream", although it would depend on your mime type database. Baloo treats these as data.

You can see what Baloo thinks of the file with:
    balooshow -x /path/to/file
Could be that Baloo hung on it previously was bad luck.
Comment 11 pqwoerituytrueiwoq 2024-09-20 02:58:18 UTC
maybe it hung on what ever it was doing after it fed that line to the UI, that file lives next to a SQLite file (8192 bytes)
both kmimetypefinder5 and mimetype see it as application/vnd.sqlite3

looks to be data related to html5 local storage that it got hung on
Comment 12 Stefan Brüns 2026-01-31 02:57:23 UTC
Frameworks Version: 5.115.0
^ No longer maintained, and 2 years old already. Please use a recent version.
Comment 13 pqwoerituytrueiwoq 2026-01-31 14:54:08 UTC
> No longer maintained, and 2 years old already.

Reported: 	2024-09-19

The difference between Sep 19, 2024 and Jan 31, 2026 is 1 year 4 months 12 days
Comment 14 tagwerk19 2026-01-31 16:02:47 UTC
(In reply to pqwoerituytrueiwoq from comment #13)
> > No longer maintained, and 2 years old already.
> 
> Reported: 	2024-09-19
Maybe the better question is whether you are still having trouble, whether you've stumbled across the issue again.

I look back at the thread and it looks as if I didn't get a similar failure - maybe my setup was different, maybe I didn't do exactly the same as you. If we can replicate the issue it is easier to see whether the issue has been resolved, if not we do rather depend on "common sense" troubleshooting steps and hope for a bit of luck.

The version number can be troublesome as some distributions pick up new releases very quickly, others stick with older, and you can read that as "tried and tested", releases.
Comment 15 pqwoerituytrueiwoq 2026-01-31 16:12:32 UTC
> Maybe the better question is whether you are still having trouble, whether you've stumbled across the issue again.

yes...

I have not had a issue since adding exclusions for indexing such as ~/.local/share/baloo
Comment 16 Stefan Brüns 2026-01-31 20:25:25 UTC
(In reply to pqwoerituytrueiwoq from comment #13)
> > No longer maintained, and 2 years old already.
> 
> Reported: 	2024-09-19
> 
> The difference between Sep 19, 2024 and Jan 31, 2026 is 1 year 4 months 12
> days

Frameworks 5.115 was released 10th of February 2024, so it is 1 year, 11 months and 20 days.