Poweroff failing, core dump failing as well

Andrew Bloomgarden

unread,

Apr 10, 2018, 2:51:32 PM4/10/18

to CoreOS User

Hi everyone, I'm seeing a strange issue on a CoreOS VM I'm running on a Xen-based VM platform I don't control. When the VM platform sends a shutdown request to the VM, I see the following in journalctl:

Apr 10 18:26:36 hostname systemd-coredump[4321]: Resource limits disable core dumping for process 4319 (poweroff).

Apr 10 18:26:36 hostname systemd-coredump[4321]: Process 4319 (poweroff) of user 0 dumped core.

and the VM fails to shut down.

Running poweroff manually works completely fine, which leads me to a couple of questions I hope for some advice on:

* Is it possible to simulate the kind of shutdown request that the VM receives from a shell inside the VM? That should make this easier to reproduce…

* Any ideas as to what could be causing this? I've convinced systemd to give me a core dump, and all I'm seeing is SIGABRT in __printf_fphex in libc.

Thank you,

Andrew

Andrew Bloomgarden

unread,

Apr 10, 2018, 2:53:30 PM4/10/18

to CoreOS User

Should have mentioned: this is on 1688.5.3.

Andrew Bloomgarden

unread,

Apr 10, 2018, 2:57:12 PM4/10/18

to CoreOS User

And, I did get a more useful stack trace:

#0  0x00007f590a198008 in raise () from /media/root/lib64/libc.so.6
#1  0x00007f590a1994f8 in abort () from /media/root/lib64/libc.so.6
#2  0x00007f590a81290e in log_assert_failed_realm () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#3  0x00007f590a82e513 in getpeergroups () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#4  0x00007f590a882e08 in ?? () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#5  0x00007f590a883122 in bus_socket_start_auth () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#6  0x00007f590a883fad in bus_socket_connect () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#7  0x00007f590a88c3ff in ?? () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#8  0x00007f590a88c911 in sd_bus_start () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#9  0x00007f590a78a4a0 in bus_connect_system_systemd () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#10 0x00007f590a78cacb in bus_connect_transport_systemd () from /media/root/usr/lib/systemd/libsystemd-shared-237.so
#11 0x0000562d57fdece8 in ?? ()
#12 0x0000562d57fe9a2a in ?? ()
#13 0x0000562d57ffcefe in ?? ()
#14 0x0000562d57ffd8fe in ?? ()
#15 0x0000562d57ffde1e in ?? ()
#16 0x00007f590a1847f6 in __libc_start_main () from /media/root/lib64/libc.so.6
#17 0x0000562d57fddf09 in ?? ()

On Tuesday, April 10, 2018 at 2:51:32 PM UTC-4, Andrew Bloomgarden wrote:

Andrew Bloomgarden

unread,

Apr 10, 2018, 3:07:59 PM4/10/18

to CoreOS User

OK, it appears that this is fixed by https://github.com/systemd/systemd/commit/75f40779607ea79f20441c7fb46744d04ee2c7ae in systemd 238. Is there a way to backport systemd fixes to stable-channel CoreOS? Happy to send a PR if someone points me in the right direction.

Thank you,

Andrew

Benjamin Gilbert

unread,

Apr 10, 2018, 3:28:44 PM4/10/18

to Andrew Bloomgarden, CoreOS User

Hi Andrew,

Thanks for tracking this down! Yes, there is. You can PR the v237-coreos branch of https://github.com/coreos/systemd/, and after that merges, PR the build-1688 branch of https://github.com/coreos/coreos-overlay to update the CROS_WORKON_COMMIT in the systemd ebuild. I'd be happy to review the PRs.

Best,

--Benjamin Gilbert

--
You received this message because you are subscribed to the Google Groups "CoreOS User" group.
To unsubscribe from this group and stop receiving emails from it, send an email to coreos-user+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply all

Reply to author

Forward