I am experiencing a weird, reproducible crash loop during startup on my Arch Linux system running KDE Plasma 6 (Wayland).
The Symptoms:
-
First Login Attempt: I type my password at the SDDM login screen. The system immediately freezes on a black screen for roughly 30 seconds to a minute.
-
Session Reset: After the timeout, it kicks me back to the SDDM login screen and prompts me to enter my password again.
-
Second Login Attempt: The second login succeeds and takes me to the desktop, but I am immediately greeted by a
drkonqi-coredump-launchercrash notification. -
Disk Exhaustion: Right at this moment, my root (
/) partition hits 100% capacity, causing other autostart applications to crash. -
Self-Recovery: If I wait a few minutes, the disk space automatically frees itself back up (returning to its normal baseline), and the system functions completely normally until the next reboot.
What seems to be happening: It looks like an underlying core session component (like kwin_wayland or kaccess) crashes on the first login attempt. Because it dumps core, it triggers systemd-coredump or drkonqi, which fills the root drive with massive dump files, locking up the system. Once systemd automatically vacuums or processes the cores, the space recovers.
Has anyone encountered this specific initialization race condition or cascade failure on Plasma 6? Which logs should I check next to pinpoint the initial culprit?
Additional Context (`drkonqi` Meta-Crash Log):
When the root partition gets choked by the initial dump file, `drkonqi` itself runs out of resources to draw its GUI and aborts, throwing this stack trace:
PID: 3995 (drkonqi-coredum)
UID: 1000 (arch)
GID: 1000 (arch)
Signal: 6 (ABRT)
Timestamp: Tue 2026-06-23 16:14:13 CEST (12s ago)
Command Line: /usr/lib/drkonqi-coredump-launcher
Executable: /usr/lib/drkonqi-coredump-launcher
Control Group: /user.slice/user-1000.slice/user@1000.service/app.slice/drkonqi-coredump-launcher@114-53252-3932_33125-0.service
Unit: user@1000.service
User Unit: drkonqi-coredump-launcher@114-53252-3932_33125-0.service
Slice: user-1000.slice
Owner UID: 1000 (arch)
Boot ID: c1edc04520ca46bab49c3a9289847143
Machine ID: 93c9f07cba3d469e8257f108d514bbee
Hostname: archlinux
Storage: /var/lib/systemd/coredump/core.drkonqi-coredum.1000.c1edc04520ca46bab49c3a9289847143.3995.1782224053000000.zst (present)
Size on Disk: 683K
Message: Process 3995 (drkonqi-coredum) of user 1000 dumped core.
Stack trace of thread 3995:
#0 0x00007f060e09a29c n/a (libc.so.6 + 0x9a29c)
#1 0x00007f060e03e7d0 raise (libc.so.6 + 0x3e7d0)
#2 0x00007f060e025681 abort (libc.so.6 + 0x25681)
#3 0x00007f060e8968e4 n/a (libQt6Core.so.6 + 0x968e4)
#4 0x00007f060e89771d _ZNK14QMessageLogger5fatalEPKcz (libQt6Core.so.6 + 0x9771d)
#5 0x00007f060f0e72da n/a (libQt6Gui.so.6 + 0xe72da)
#6 0x00007f060f1ae228 _ZN22QGuiApplicationPrivate21createEventDispatcherEv (libQt6Gui.so.6 + 0x1ae228)
#7 0x00007f060e989175 _ZN23QCoreApplicationPrivate4initEv (libQt6Core.so.6 + 0x189175)
#8 0x00007f060f1ae32b _ZN22QGuiApplicationPrivate4initEv (libQt6Gui.so.6 + 0x1ae32b)
#9 0x00007f060f1af69e _ZN15QGuiApplicationC2ERiPPci (libQt6Gui.so.6 + 0x1af69e)
#10 0x000056241d978d85 n/a (drkonqi-coredump-launcher + 0x6d85)
#11 0x00007f060e027741 n/a (libc.so.6 + 0x27741)
#12 0x00007f060e027879 __libc_start_main (libc.so.6 + 0x27879)
#13 0x000056241d97a385 n/a (drkonqi-coredump-launcher + 0x8385)
ELF object binary architecture: AMD x86-64
-- Notice: 1 systemd-coredump@.service unit is running, output may be incomplete.
