tracker-miner-fs-3 never seems to complete.
I remember I used to have this problem back in the 3.2.x era, but recently this has become a thing too. So as it stands now the tracker-miner-fs-3
process has been running over 13 hours now and using 99% CPU across 16 cores (not 99% on one core).
When trying to see what is going on tracker3 status
it will not return and just sits there.
Can't see any errors in user.log
so have no idea what it is stuck doing.
There are no errors in ~/,cache/tracker3/files/errors
though ~/,cache/tracker3/files
is 957Mb in size atm.
For reference, I am on
- Debian 12 (testing)
with
- Gnome 43.3
and
- tracker3-miner-fs version 3.4.3.
There are no mounted network shares.
The configured dconf keys are.
key | value |
---|---|
Ignored directories | ['po', 'CVS', 'core-dumps', 'lost+found', '.snapshots', 'mnt'] |
Ignored directories with content | ['.trackerignore', '.git', '.hg', '.nomedia', '.snapshots'] |
Ignored files | ['~', '.o', '.la', '.lo', '.loT', '.in', '.m4', '.rej', '.gmo', '.orig', '.pc', '.omf', '.aux', '.tmp', '.vmdk', '.vm*', '.nvram', '.part', '.rcore', '.lzo', 'autom4te', 'conftest', 'confstat', 'Makefile', 'SCCS', 'ltmain.sh', 'libtool', 'config.status', 'confdefs.h', 'configure', '##', '~$.doc?', '~*.dot?', '~ *.xls?', '~*.xlt?', '~ *.xlam', '~*.ppt?', '~ *.pot?', '~*.ppam', '~ *.ppsm', '~*.ppsx', '~ *.vsd?', '~*.vss?', '~ *.vst?', '.directory', '.zip'] |
In addtion, querying the daemon seems to hang, so tracker3 status
will just sit there as will tracker info <foo>
I have tried looking at the hold up with perf and so ran
timeout 15 perf record --call-graph dwarf -p `pidof tracker-miner-fs-3`
A pastebin of perf.data of that is here
As it stand the miner has been running for more than 13 hours at this point.
Something to note. Last time this happened (before the last reboot) the metadata in ~/.cache/tracker3/files
was 4.3Gb. Last time I killed the miner, ran tracker3 reset -s
and restarted the miner and it completed leaving only 992Mb of cache left.
Currently (as it is still runnning) that cache is 957Mb
Any ideas how to find why the miner is not completing ?
TIA
Edit: Forgot to attach the perf.data here.