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

I am triggering a crash in order to get a vmcore dump of an Azure ND H100v5 instance running RHEL 9.4 GA. When the crash is triggered and kdump runs, it prints the vmcore was saved successfully on the console. However when the system reboots, the vmcore is missing and the specified directory is empty.

What is the impact of this issue to you?

I cannot complete certification of the system due to the kdump test not working.

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

kexec-tools-2.0.27-8.el9.x86_64
kexec-tools-2.0.27-8.el9_4.3.x86_64

How reproducible is this bug?:

I can reproduce it every time.

Steps to reproduce

1. Get an Azure system with an official Gen 2 x64 RHEL 9.4 (LVM) image.
2. If needed use LVMs to expand disk space.
3. Configure kdump as follows and restart kdump: (/etc/kdump.conf)

auto_reset_crashkernel yes
core_collector makedumpfile -l --message-level 7 -d 31
failure_action shell
force_rebuild 1
path /var/crash/cc99c796-9c17-11ef-b9e7-000d3a98ecf2

1. Trigger a system crash with echo c > /proc/sysrq-trigger.
2. Observe console output to confirm vmcore is saved and let the system reboot.
3. Check the specified path to see if vmcore exists.

Expected results

vmcore is saved in the expected path.

Actual results

No vmcore is found on the system.