Comment # 15 on bug 1207243 from Chris Puttick

Have now persuaded journalctl to be in a sensible configuration, but it doesn't
log anything particularly insightful the fail per se. Last line logged on
failed boots (both normal and  recovery) is

systemd-journald[610]: Time spent on flushing to /var/log/journal/

on the normal boot process with the older kernel, this line is followed with 

systemd-journald[605]: Received client request to flush runtime journal.
systemd-journald[605]: File
/var/log/journal/84403c8f9fc24658abd0248aa7f8cc73/system.journal corrupted or
uncleanly shut down, renaming and replacing.
kernel: thermal LNXTHERM:00: registered as thermal_zone10
kernel: ACPI: thermal: Thermal Zone [THM] (25 C)
kernel: intel-lpss 0000:00:15.0: enabling device (0000 -> 0002)
kernel: mc: Linux media interface: v0.10
kernel: idma64 idma64.0: Found Intel integrated DMA 64-bit
kernel: mei_me 0000:00:16.0: enabling device (0000 -> 0002)
kernel: ACPI: bus type thunderbolt registered

I took a pic of the lines showing on the failed boot in recovery mode and will
attach it.