Comment 15 for bug 1946149

Hey Kleber,

Thanks for confirming.

I guess there might be something wrong with the boot process on r5.metal, specifically:
- there's no issue with kexec boot, just with normal boot (same code and from/to versions)
- there's no issue with normal boot on similar instance types (r5d.metal, r5.24xlarge)
- rarely, but sometimes, it works with normal boot.

I'm looking forward to hearing back from Ian on whether he could
get additional information from AWS, or how to reach out to them,
about getting the serial console log from the r5.metal instances
internally, via AWS support, so we know what the actual issue is.