Can not boot RHEL 7.3 by using iPXE

[djamelo]

there is no official ipxe git version of 0c8d16 Where does your ipxe come from? is it based on the latest git master?

HP is somewhat problematic, and so is several RHEL and CentOS releases
But if I understood you correctly you have tried other EFI boots on the same hardware that works?
Have you tried the problematic version in a VM with EFI boot?

[djamelo]

I am new to this forum and iPXE. I have been struggling to get my CentOS 7.4 1708 to PXEboot successfully ultimately to proceed to KickStart - using a VMware ESXi VM (BIOS and not EFI). Like this thread, both the CentOS's pxeboot initrd and vmlinuz files are being loaded via HTTP but vmlinuz fails to complete its boot due to not being able to mount the root filesystem.

I note that "mathp" says he/she had success with iPXE and "stock" Centos 7.4.1708 installer kernel. I would appreciate a snippet of the iPXE script that did this so that I can confirm my environment.

Thanks

[djamelo]

Quick Update, I had great success with iPXE and "stock" CentOS 7.4.1708 installer kernel - no modification needed during my tests.

[MultimediaMan]

This is a known issue.

You can "solve" the problem (most of the time) by doing the following:

DD a 16MB or 32MB file:

Code:

dd if=/dev/urandom of=32MB.bin bs=32MB count=1

Then place the file in the Webserver or TFTP directory...

Place the following line in your boot script:

Code:

initrd http://path.to.file/32MB.bin

Some things can get tricky... the order in which they are downloaded can affect things; in general I will download the initrd(s) and any *.bin file before the kernel:

Code:

initrd http://path.to.file/initrd
initrd http://path.to.file/32MB.bin
chain http://path.to.file/vmlinuz initrd=initrd

Yes, you are just filling some parts of the memory with random characters, but if it works...

This scheme works most of the time...YMMV.

Cheers,

M^3

[djamelo]

there is no official ipxe git version of 0c8d16 Where does your ipxe come from? is it based on the latest git master?

HP is somewhat problematic, and so is several RHEL and CentOS releases
But if I understood you correctly you have tried other EFI boots on the same hardware that works?
Have you tried the problematic version in a VM with EFI boot?