What were you trying to do that didn't work?

I entered the kdump emergency shell to debug an issue but the system got rebooted automatically after a few minutes.

What is the impact of this issue to you?

kdump emergency shell (failure_action shell) is unusable.

Please provide the package NVR for which the bug is seen:

kdump-utils-1.0.54-1.el10.x86_64

rhel-guest-image-10.1-20250603.1.x86_64.qcow2

How reproducible is this bug?:

always

Steps to reproduce

1.  Use virt-install --import to install RHEL10 guest machine e.g. rhel-guest-image-10.1-20250603.1.x86_64.qcow2
2. kdumpctl restart
3.  printf "core_collector nop\nfailure_action shell\n" > /etc/kdump.conf
4.  kdumpctl restart
5.  trigger sysrq

Expected results

I can stay in the emergency shell as long as I want.

Actual results

The system gets rebooted automatically.

Additional notes

1. The system won't be rebooted if I change watchdog action from action=reset to action=none by "virsh edit GUEST_NAME",

<watchdog model='itco' action='none'>
  <alias name='watchdog0'/>
</watchdog>

2. rhel-guest-image-10.0-20250111.0.x86_64.qcow2 also have this issue but not RHEL9 guest image.