Comment 0 for bug 1886115

Revision history for this message
guenthert (guenthert) wrote :

  After applying updates to Ubuntu 18.04 my desktop (apple mini with i5-2415M CPU) failed to complete the boot process. A few seconds into the boot, the last message displayed is "/var mounted". The system then appears to hang indefinitely.

  Luckily, the 'rescue' boot image allows the boot process to proceed sufficiently far to allow a root shell to be spawned. Unfortunately no log files were written during the unsuccessful attempts to boot. Spawning a 2nd root shell (# nohup getty tty5) on a 2nd virtual terminal (tty5) I was able to observe the message 'systemd freezing execution' after I closed the first root shell and resumed the boot process. Further a core file was created (belonging to /sbin/init) in the root fs
--8<--
(gdb) bt
#0 0x00007f16807ba187 in kill () at ../sysdeps/unix/syscall-template.S:78
#1 0x0000563b957223b7 in ?? ()
#2 <signal handler called>
#3 __GI___libc_free (mem=0x4a60d140dfd9a5) at malloc.c:3103
#4 0x0000563b9577c22e in ?? ()
#5 0x0000563b957672d6 in ?? ()
#6 0x0000563b9576ba22 in ?? ()
#7 0x0000563b9574f51a in ?? ()
#8 0x00007f16803a509a in ?? () from /lib/systemd/libsystemd-shared-237.so
#9 0x00007f16803a53ea in sd_event_dispatch () from /lib/systemd/libsystemd-shared-237.so
#10 0x00007f16803a5579 in sd_event_run () from /lib/systemd/libsystemd-shared-237.so
#11 0x0000563b9572a49d in ?? ()
#12 0x0000563b9571560c in ?? ()
#13 0x00007f168079cb97 in __libc_start_main (main=0x563b957139c0, argc=3, argv=0x7ffe78153758,
    init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>,
    stack_end=0x7ffe78153748) at ../csu/libc-start.c:310
#14 0x0000563b957164fa in ?? ()
(gdb)
-->8--
 and the kernel message buffer lists
--8<--
traps: systemd[1] general protection fault ip:7f17ebf6e98d sp:7ffd774d6020 error:0 in libc-2.27.so[7f17ebed7000+1e7000]
-->8--
.

  To me that looked a bit like Bug 669702 of Gentoo (https://bugs.gentoo.org/669702) and indeed one of the (few) updates applied just prior the reboot was the update of libseccomp.

  I was able to circumvent the problem by disabling (commenting out) the syscall filtering requested by systemd (on my system, only /etc/systemd/system/dbus-org.freedesktop.resolve1.service needed to be modified).