You are not logged in.
Hello everyone,
this problem is a bit complicated for me to explain, so bear with me.
Since a couple of days I notice nepomukindexer crashing quite often. Basically if I keep file indexing turned on, even after the system is done indexing everything in my home (it shows "indexer is idle"), after an hour or so I notice my system's CPU usage skyrocketing. It's because of systemd-journalctl, which is spamming the log with the message "Process XXXX (nepomukindexer) dumped core." (where XXXX is always changing). If I check with top, there are hundreds of processes called systemd-coredump and they keep getting spawned, forcing me to reboot my system. If I check the journal, I can also read "nepomukindexer[2800]: segfault at 0 ip 00007f3a5dc0c591 sp 00007fffad123b00 error 4 in libc-2.17.so[7f3a5dba2000+1a3000]".
Anyone else is having this issue? Also, how can I check the log of nepomukindexer, to see if there's a particular file that makes it crash? I can't seem to find it anywhere.
Thanks!
Offline
Yes, I am having a very similar issue like gondsman's as well.
I am not 100% sure, but it seems that these high CPU usages start after locking and unlocking the computer.
Offline
I guess I'll have to bring this back up again but I am having the exact same issues.
High CPU usage due to systemd-journald with journal showing segfaults of nepomukindexer such as:
nepomukindexer[13183]: segfault at 560 ip 00007f764b9cf6a6 sp 00007fffe7e8de90 error 4 in libQtGui.so.4.8.5[7f764b5da000+a5e000]
or
systemd-coredump[13187]: Process 13183 (nepomukindexer) dumped core.
Any ideas? Was it solved for the first ones posted?
Thanks
Offline
It's happening to me here too.
nepomukindexer dies and gets respawned. This causes a new process, systemd-coredump (eats 15MB of RAM), to be created. If you wait a while then systemd-journald will process something and the systemd-coredump will terminate, however the new nepomukindexer soon dies. Since nepomukindexer dies and respawns faster than systemd-journald can terminate the systemd-coredumps I end up with heaps of these 15MB processes sitting in the queue, hitting swap.
You can killall systemd-coredump since they are owned by your user once you see it start.
I can also confirm that the indexer only runs when the screen-locker is on
▁▂▃▄▅▆▇█▉▊▋▌▍▎▏▏▎▍▌▋▊▉█▇▆▅▄▃▂▁
Offline
I have the exact same issue as well. It is very annoying systemd-journald eats about 25% of my CPU.
Offline
Same for me. If I untick "Enable Nepomuk File Indexer" my system is back to normal. Hence, the issue is with the file indexer not with the mail indexer.
Offline
Offline
Hello,
same for me. How can I reindex all my data?.
Regards.
Offline
Hello,
same for me. How can I reindex all my data?.
Regards.
qdbus-qt4 org.kde.nepomuk.services.nepomukfileindexer /nepomukfileindexer org.kde.nepomuk.FileIndexer.indexFolder path-to-folder 1 1
Offline