Comment # 10 on bug 1158241 from Ulrich Windl

I don't like systemd playing operating system: The handling of I/O errors and
slow I/O should be handled by the kernel, not by systemd. Unfortunately the
journal does not log the reason of the core dump: SIGABRT?
But I see that the journal is not written with ascending time:

Dec 02 12:49:57 pc kernel: SFW2-INext-DROP-DEFLT IN=eth0 OUT=
MAC=00:22:11:dd:00:11:bc:ea:fa:a4:79:00:08:00 SRC=77.247.110.72 1.9.5.2 LEN=40
TOS=0x00 PREC=0x00 TTL=246 ID=34189 PROT
O=TCP SPT=59403 DPT=5072 WINDOW=1024 RES=0x00 SYN URGP=0 
Dec 02 12:49:57 pc systemd-coredump[21380]: Process 1726 (systemd-journal) of
user 0 dumped core.
...
Dec 02 12:49:57 pc systemd-journald[21398]: Journal started
Dec 02 12:49:57 pc systemd-journald[21398]: System journal
(/var/log/journal/89c660865c00403a9bacef32b6828556) is 832.0M, max 819.2M, 0B
free.
...
Dec 02 12:48:02 pc systemd[1]: systemd-journald.service: Watchdog timeout
(limit 3min)!
Dec 02 12:48:02 pc systemd[1]: systemd-journald.service: Killing process 1726
(systemd-journal) with signal SIGABRT.

>From the user's perspective journald was killed while actually writing log
messages, so the killing made no sense.