On Wed, Apr 15, 2020 at 3:08 AM Łukasz Posadowski
<mail(a)lukaszposadowski.pl> wrote:
Today I had an issue during regular system update. Apparently time was
set on Unix 0 (with +1 hour for a timezone), which confused at least
dnf, upgrade failed and system didn't respond to anything, including
ctrl+alt+del. Kernel was not installed correctly and panicked during
boot. I was able to reinstall kernel being chrooted from rescue mode of
VPS provider.
I have no idea what happened. Did You experienced anything similar
before? Or lately with Fedora 32?
Some logs:
https://pastebin.com/YtPEV7d5
Maybe this:
https://bugzilla.redhat.com/show_bug.cgi?id=1823463
I can reproduce the complaint from 'hwclock -s' but it seems at least
on my hardware, the kernel is setting the system time from the
hardware clock during boot.
$ dmesg | grep -i rtc
[ 0.120076] PM: RTC time: 18:41:34, date: 2020-04-15
[ 0.770228] rtc_cmos 00:03: RTC can wake from S4
[ 0.770713] rtc_cmos 00:03: registered as rtc0
[ 0.770714] rtc_cmos 00:03: alarms up to one month, y3k, 242 bytes
nvram, hpet irqs
[ 0.855104] rtc_cmos 00:03: setting system clock to
2020-04-15T18:41:35 UTC (1586976095)
That is before systemd starts up from initramfs.
[ 1.244044] flap.local systemd[1]: systemd v245.4-1.fc32 running in
system mode. (+PAM +AUDIT +SELINUX +IMA -APPARMOR +SMACK +SYSVINIT
+UTMP +LIBCRYPTSETUP +GCRYPT +GNUT>
--
Chris Murphy