On Thu, Jun 13, 2019 at 16:01:18 +0200, Lentes, Bernd wrote:
----- On Jun 13, 2019, at 1:08 PM, Bernd Lentes bernd.lentes(a)helmholtz-muenchen.de
wrote:
I found further information in /var/log/messages for both occurrences:
2019-06-01T03:05:31.620725+02:00 ha-idg-2 systemd-coredump[14253]: Core Dumping has been
disabled for process 30590 (qemu-system-x86).
2019-06-01T03:05:31.712673+02:00 ha-idg-2 systemd-coredump[14253]: Process 30590
(qemu-system-x86) of user 488 dumped core.
2019-06-01T03:05:32.173272+02:00 ha-idg-2 kernel: [294682.387828] br0: port 4(vnet2)
entered disabled state
2019-06-01T03:05:32.177111+02:00 ha-idg-2 kernel: [294682.388384] device vnet2 left
promiscuous mode
2019-06-01T03:05:32.177122+02:00 ha-idg-2 kernel: [294682.388391] br0: port 4(vnet2)
entered disabled state
2019-06-01T03:05:32.208916+02:00 ha-idg-2 wickedd[2954]: error retrieving tap attribute
from sysfs
2019-06-01T03:05:41.395685+02:00 ha-idg-2 systemd-machined[2824]: Machine qemu-31-severin
terminated.
2019-06-08T05:59:17.502899+02:00 ha-idg-1 systemd-coredump[31089]: Core Dumping has been
disabled for process 19489 (qemu-system-x86).
2019-06-08T05:59:17.523050+02:00 ha-idg-1 systemd-coredump[31089]: Process 19489
(qemu-system-x86) of user 489 dumped core.
2019-06-08T05:59:17.650334+02:00 ha-idg-1 kernel: [999258.577132] br0: port 9(vnet7)
entered disabled state
2019-06-08T05:59:17.650354+02:00 ha-idg-1 kernel: [999258.578103] device vnet7 left
promiscuous mode
2019-06-08T05:59:17.650355+02:00 ha-idg-1 kernel: [999258.578108] br0: port 9(vnet7)
entered disabled state
2019-06-08T05:59:25.983702+02:00 ha-idg-1 systemd-machined[1383]: Machine
qemu-205-severin terminated.
Core Dumping is disabled, but nevertheless a core dump has been created ?
Where could i find it ?
Would it be useful to provide it ?
So this really hints to qemu crashing. It certainly will be beneficial
to collect the backtrace, but you really should report this (including
the error message from the vm log file) to the qemu team.
They might have even fixed it by now, so a plain update might help.