FDI 4.1.0 kexec issues with RHEL7.9 kickstart kernel on HPE DL360 Gen10 Plus

Problem:
After clicking “Auto Provision” in the discovered host and kexec is launched, a few seconds into the new kernel loading, the server gets a cold boot and never reaches the kickstart installation part.

Expected outcome:
New kernel is loaded and kickstart installation starts.

Distribution and version:
RHEL7.9
FDI 4.1.0

Other relevant data:
This only happens on HPE DL360 Gen10 Plus and only with RHEL7.X installer kernel.
All works fine with this hardware if I’m deploying RHEL8.X.
All works fines with other hardware like DL 365 Gen10 Pus with RHEL7.X and 8.X.
Have tried with:
nomodeset
acpi=off
acpi_rsdp= <value taken from (grep -m1 ^ACPI /sys/firmware/efi/systab | cut -f2- -d=)>

Anyone come across similar behavior?

Tks.

Yes, unfortunately kexec is quite buggy. Typically graphics drivers are offenders. With UEFI it got worse, so we no longer recommend the kexec workflow for production use.

Hi Lzap, thanks for your feedback. What do you recommend then? I still only see kexec based discovery templates.

Not using kexec workflow, I know that discovery “pxeless” is based on that. Use bootdisk, the most reliable is full host bootdisk which is SYSLINUX based and require no network.