qemu-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: qemu-system-x86_64 run VM (nested) doesn't reboot occasionally


From: Anoop Karollil
Subject: Re: qemu-system-x86_64 run VM (nested) doesn't reboot occasionally
Date: Tue, 2 Jun 2020 12:56:29 -0700

Forgot to mention the version of qemu:

$ qemu-system-x86_64 --version
QEMU emulator version 2.11.1(Debian 1:2.11+dfsg-1ubuntu7.26)
Copyright (c) 2003-2017 Fabrice Bellard and the QEMU Project developers

On Tue, 2 Jun 2020 at 12:50, Anoop Karollil <anoop.karollil@gmail.com> wrote:
>
> Hello,
>
> TL;DR:
>
> QEMU VM (run within another KVM) doesn't reboot occasionally when a
> kernel panic is induced, with either the kernel configured to reboot
> on panics, or when a watchdog is configured. The VM kernel log (set to
> be printed to console) doesn't show anything after the kernel panic,
> and qemu-system-x86_64 is still running, but the VM doesn't reboot.
> Any suggestions to debug this would be appreciated!
>
> Details:
>
> A test runs a Linux OS (Yocto based, with 4.14.157 kernel) in a QEMU
> VM, with the test itself running in a KVM. The OS uses systemd, which
> is configured to use any watchdog available to reboot the VM if the
> watchdog isn't tickled for more than 10 seconds (using
> RuntimeWatchdogSec=10s). The VM has been setup to use the i6300esb
> watchdog. This is the command used to start the VM:
>
> qemu-system-x86_64 -global isa-fdc.driveA= -boot c -bios ovmf/ovmf.fd
> -nographic -cpu core2duo -m 1000 -net nic,model=virtio -net
> user,hostfwd=tcp::22221-:22 -boot c -drive
> file=image.hddirect.qcow2,if=virtio -watchdog i6300esb -enable-kvm
> -global virtio-pci.disable-modern=on
>
> The test induces a kernel panic with `echo c > /proc/sysrq-trigger`.
> This results in the VM being rebooted by the watchdog most of the
> time, but occasionally the reboot doesn't happen though the watchdog
> does seem to trigger every time (I tested this by setting the watchdog
> action to debug). Also when I removed the watchdog from the VM (by
> removing -watchdog parameter) and enabled configuration to reboot on
> kernel panic, then too the VM doesn't reboot occasionally.
>
> Thanks,
> Anoop



reply via email to

[Prev in Thread] Current Thread [Next in Thread]