Bug 1371299

There is not a lot of difference between automatically installing RHVH and RHEL, thus the RHEL automated installation guide can be taken for reference:

https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Installation_Guide/chap-kickstart-installations.html

However, as RHVH does not use a package management system, the %packages section is not needed.
Instead there are a few other constraints:
a) Use he kernel, initrd and stage2 squashfs from a RHVH installation DVD
b) liveimg --url=$URL_TO_SQUASHFS directive must be used instead of %packages
c) %post section which calls "nodectl init" is required.
d) autopart must be used (as long as the manual partitioning is not documented)

These additional constraints should be added to the RHVH installation docs, because that is where RHVH installation differs from RHEL.

The RHEV 2.x guides have been removed from the Portal, so for reference, the sections from the bug description can now be found at: 

https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Virtualization/3.1/html/Hypervisor_Deployment_Guide/sect-Deployment_Guide-Preparing_Red_Hat_Enterprise_Virtualization_Hypervisor_installation_media-Deploying_RHEV_Hypervisors_with_PXE_and_tftp.html

https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Virtualization/3.1/html/Hypervisor_Deployment_Guide/sect-Deployment_Guide-Installing_Red_Hat_Enterprise_Virtualization_Hypervisors-RHEV_Hypervisor_Kernel_Parameters_and_Automated_Installation.html

For PXE, the RHEL docs have the following sections:

Preparing for a Network Installation:
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Installation_Guide/chap-installation-server-setup.html

Booting from the Network Using PXE:
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Installation_Guide/chap-booting-installer-x86.html#sect-booting-from-pxe-x86

I think that we can just have a section linking to these, the same as we will for kickstart, since I can't imagine there's a lot of difference between RHEL and RHVH. Fabian, can you confirm? Any requirements/limitations/differences that should be noted?

Yes. Linking the two and the ks section is good.

But we have a few differences.

The required kickstart must contain:

autopart --type=thinp
liveimg --url=<url to squashfs>

%post
nodectl init
%end

If autopart is not used, then there are a few constraints:
/ needs to be a thin LV
/var needs to be a separate volume
/boot needs to be on a separate partition

I'm a little concerned that the warning about the squashfs/kernel is not more explicit.

Anaconda must discover the correct product as part of /.buildstamp, which is part of the squashfs. If the squashfs from the RHV-H ISO is not explicitly referenced, anaconda will not know this is a RHV-H host, and the wrong partitioning will be selected.

Hi Ryan, 
I tried to put those constraints in the order they would appear in the Kickstart file, according to the RHEL docs. However, since the RHVH squashfs is probably the most important part, I can put it first in that list if that would help?

Hi Tahlia,

I would suggest leaving it in the same position as the "url" parameter normally is. It's true that liveimg (along with autopart) is the most important parameter in the main section of the kickstart, but replacing 'url' most closely matches the RHEL documentation.

That said, I'm not a documentation expert at all, so I'll trust your opinion.

Thanks Ryan, I'll leave it as is.

Hi Tahlia -

Just to clarify, the squashfs (the `liveimg` directive) should be in the same place as url.

However, the requirement on the kernel/initramfs (stage2) from the ISO still needs to be added.

Now published at
https://access.redhat.com/documentation/en/red-hat-virtualization/4.0/single/installation-guide/#Automating_RHVH_Deployment
and 
https://access.redhat.com/documentation/en/red-hat-virtualization/4.1-beta/single/installation-guide/#Automating_RHVH_Deployment