Demystifying the Install Process of Red Hat Enterprise Linux on IBM Z and LinuxONE Native LPAR

Red Hat OpenShift Platform is becoming more and more popular in the IBM Z and LinuxONE world, and many more features are being introduced over time. Red Hat OpenShift Platform 4.7 adds support for another great hypervisor option on IBM Z and LinuxONE: KVM (Kernel Virtual Machine).

Figure 1. Red Hat OpenShift on IBM LinuxONE with KVM

KVM has been around for quite some time for the IBM Z and LinuxONE platform, but the big news for Red Hat OpenShift Platform 4.7 is that Red Hat now provides official support when running on top of virtual machines provided by the kvm hypervisor. This is done by leveraging the Red Hat Enterprise Linux distribution, allowing the creation of virtual machines when RHEL is deployed natively on a Logical Partition (LPAR) on the IBM Z and LinuxONE hardware.

Installing Red Hat Enterprise Linux natively on a LPAR can have other purposes besides installing Red Hat OpenShift Platform on top of KVM guests, many Red Hat customers leveraged RHEL natively on the LPAR to run specific workloads such as databases or other applications. The ability to created KVM guests also provides an alternative to IBM z/VM for all other types of virtualized workloads as well.

Logical Partition? Virtual Machines? Hopefully this picture shows how the pieces of the stack fit together:

Figure 2: Red Hat Virtual Machines on IBM LinuxONE with KVM

How to get started?

A challenge many people face when trying to install RHEL natively on the LPAR, is finding an authoritative step-by-step guide to help with the process.

For the purpose of this article, it is assumed the following requirements have already been met:

• Logical Partition (LPAR) created (using PR/SM or DPM). That means logical resources as I/O devices (disk, DASD in this case), Network card (OSA), memory, and CPU (logical IFLs) allocation is done. Think of this process as carving out a computer within a computer, as the IBM Z and LinuxONE is a huge pool of resources, so the LPAR a logical division where a few of these resources are dedicated and/or shared.
• The network card to connects you to the rest of your network has an appropriated IP address will be allocated so that the future RHEL system installed on this LPAR can communicate correctly. Make sure to talk to the network team so they can provide what network you will connect in order to access other resources such as a ftp server to conduct a network install process of RHEL.
• The RHEL install process will use a network install and it will requires a RHEL DVD ISO image shared on the network using http or ftp from another server. To access the Red Hat Enterprise Linux images, either must already own a Red Hat subscription for RHEL or have requested an evaluation subscription.

NOTE: as mentioned before, we have different ways to create LPARs (using either HMC or DPM) Both are valid for installing Linux OSes natively on LPARs. This article will show you the method we used, once the LPAR was created and logical resources were attached to this LPAR.

Figure 3. Environment used for this article

Install Process

On your local desktop, using a web browser, log in on the IBM System Z Hardware Management Console (HMC) or the Support Element (SE) as a user with sufficient privileges to install a new operating system to an LPAR. The SYSPROG user is recommended.

Select the system (CEC from the Systems Management section) and the LPAR. In this case, RHL8INST LPAR will be used. Select it by clicking in the Selection box.

Expand Recovery and choose Load From Removable Media or Server

Choose FTP as your source and provide the Host name, User name, Password and path to the Red Hat Enterprise Linux repository. Click OK.

NOTE: This example uses FTP, but we recommend using SFTP or FTPS.

The next screen will show the path to the generic.ins found by the installer from the path to the RHEL repository provided earlier. The generic.ins file contains the basic parameters to start the IPL process of RHEL (this will be covered in a later section of this article). Click OK.

A new screen will prompt to continue to load the removable media process, select YES.

This next step you will be prompted with another task confirmation process, this will load the RHEL image to the LPAR. Select YES:

The next screen shows that the RHEL image is being loaded and the IPL process has started

Before proceeding to the next section of the install process let's pause for a moment and understand what the generic.ins provided at the beginning of the OS load into the LPAR does and what other kernel parameters will be used to customize this install process.

The generic.ins contains the minimal LPAR load parameters, this will provide from the mounted RHEL DVD ISO the kernel.img, and initrd.img, a genericdvd.prm and the initrd.addsize files:

* minimal lpar ins file
images/kernel.img 0x00000000
images/initrd.img 0x02000000
images/genericdvd.prm 0x00010480
images/initrd.addrsize 0x00010408

The kernel and initrd images are needed to load the Linux kernel and additional kernel modules.

Contents of the genericdvd.prm:

ro ramdisk_size=40000 cio_ignore=all,!condev rd.cmdline=ask

This created a ramdisk, and basically tells the install process to ask for additional kernel parameters needed for the installation process.

Now let's continue the installation process now we understand what files are loaded during the minimum LPAR load from the DVD ISO.

Select the Home tab and select the LPAR you were using for this installation. From the Daily sub-menu select Operating System Messages.

From the the Operating System Messages tab, additional kernel parameters will be inserted using the Command box to complete the set of kernel parameters specific to the RHEL systems that will be loaded into this LPAR. Note that there is a Command text box where the kernel parameters will have to be inserted, one line at a time with a 80 characters per line limit. After each line, press the Send button.

Below is a sample of kernel parameters used for this install process. Insert these parameters one line at a time:

ro ramdisk_size=40000 cio_ignore=all,!condev
rd.znet=qeth,<0.0.0130,0.0.0131,0.0.0132>,layer2=1,portno=0
ip=<IP_Address>::<Network_Gateway>:<Network_Netmask>:<rhl8inst>:<enc130>:none
nameserver=<IP_DNS_Server>
inst.repo=ftp://<userid>:<mypassword>@<ftp_server>/<RHEL-repository>
rd.dasd=<0.0.0f00>
vnc vncpassword=<VNC_Password>

A few explanations about the parameters:

rd.znet=qeth,<0.0.0130,0.0.0131,0.0.0132> these are the logical addresses for the OSA device attached to the LPAR

rd.dasd=<0.0.0f00> this is the logical address of the allocated dasd disk to this LPAR

inst.repo=ftp://<userid>:<mypassword>@<ftp_server>/<RHEL-repository> this configuration tells where to find the stages of the RHEL install process. Usually this contains the network path to find the install tree for RHEL the install process will use the network install method.

After inserting all necessary kernel parameters, to load this configuration send the last line with a dot ("."), this will tell the install process to continue to load the kernel and find the next stage of the install process.

The next step of the install process is to connect to the newly system with the IP address or hostname of the instance using the ssh command from a command prompt/terminal window.

From the Desktop, open a terminal and type the following command:

# ssh install@<IP_Adress_New_Linux>

Once you are connected new instructions will show you the next step which is to access the next stage of the installer using a VNC client:

Starting installer, one moment...
anaconda 33.16.3.26-1.el8 for Red Hat Enterprise Linux 8.3 Started
  * installation log files are stored in /tmp during the installation
  * shell is available in second TMUX pane (ctrl+b, then press 2)
  * when reporting a bug add logs from /tmp as a separate text/plain attachments
15:28:18 Starting VNC...
15:28:20 The VNC server is now running
15:28:20 

You Chose to execute vnc with a password.

15:28:20 Please manually connect yout vnc client to <IP_Address_Linux>:1 to begin the install.
15:28:20 Attempting to start vncconfig

Leave this terminal open on your desktop and now open the VNC Client:

When prompted verify the password for the VNC session (previously defined by the kernel parameters):

Once you are connected to the VNC session, the Welcome to the Red Hat Enterprise Linux install screen will be visible. The install process is now the same as any other RHEL installation on any other platform.

NOTE: we noticed when using DASD disks, the installer will show the DASD disk as 0B Free. This is not a problem. This happens because the disks are not yet formatted. Select the disks you want to use and select Done.

This will automatically start the format process.

This concludes the objective of this article which is to guide others on how to install Red Hat Enterprise Linux natively in a logical partition from the IBM Z/LinuxONE system.

Authors:

Bob Blessing Hartley  is the Chief Architect for Industry and Regulatory Compliance. His key responsibilities are leading client engagement development of solutions for compliance for IBM Z and IBM LinuxONE.

Filipe Miranda is the Worldwide OpenShift Architect for the IBM zAcceleration team part of the IBM Hybrid Cloud for IBM Z/LinuxONE.

Contributors:

Anderson Augusto da Silveira is a consultant for the IBM Technology Lab Services for Linux and IBM Cloud. Focused in LinuxONE, he delivers and advises in complex enterprise projects around the world.

The views expressed in this article are based on our experiences alone, it does not represent IBM.