KVM in a PowerVM LPAR

KVM in a PowerVM LPAR utilizes the industry standard Linux KVM virtualization stack and can easily integrate within an existing Linux virtualization ecosystem. KVM in an LPAR is enabled by:

• IBM Power architecture and Power10: This implementation has advanced virtualization capabilities to run multiple operating system (OS) instances that share the same hardware resources while providing isolation. The Radix MMU architecture provides the capability to independently manage page tables for the LPAR and the KVM guest instances on the LPAR.

• PowerVM : The industry-leading virtualization stack provides new functions to create and manage KVM guests. These changes extend the Power platform architecture to include new hypervisor interfaces.

• Linux kernel that includes the KVM kernel module (KVM): Provides core virtualization infrastructure to run multiple virtual machines in a Linux host LPAR. Upstream kernels and enabled downstream distributions such as Fedora and Ubuntu use the newly introduced Power architecture extensions to create and manage KVM guests in the PowerLinux LPAR.

• QEMU: User space component that implements virtual machines on the host that use KVM functions.

• LibVirt: Provides a toolkit for virtual machine management.

The following features are not supported on KVM logical partitions:

• Shared processors
• vPMEM LUNs
• Platform keystore
• Live partition migration (LPM)
• Dynamic platform optimization
• Add or Remove memory, processor, and I/O DLPAR

The following KVM features for guests are not supported:

• PCI pass-through of LPAR-attached PCI devices to KVM guests. This feature will be supported in future releases.

In KVM guests, the disks available are based on virtIO (virtual I/O) by default. The two emulated storage controllers that are commonly used are virtio-blk and virtio-scsi.

1. Virtio-blk: The virtio-blk device provides the guest with a virtual block device. This controller is older than virtio-scsi. When higher performance is required, virtio-scsi is preferred.
2. Virtio-SCSI: This device presents a SCSI host bus adapter to the guest virtual machine. In contrast to virtio-blk, SCSI offers more features and flexibility. It is more scalable and can support hundreds of devices. Virtio-scsi uses standard SCSI command sets that simplify feature addition.

User Networking: This is the default networking configuration (virtio-net) for a KVM guest where the guest OS has access to the network services by using network address translation (NAT). The following characteristics or use cases are for user networking:

1. Uses default network of Libvirt (no additional network configuration is required).
2. The simplest way to configure and access the internet or resources that are available on the local network.
3. By default, the guest OS receives an IP address in the 192.168.122.0/24 address space.
4. The guest is not accessible from outside of the host in the network directly. You can make the guest accessible by configuring port forwarding through iptables.
5. Might not support some networking features such as ICMP (pings might not function properly).
6. Other networking options can provide better performance.
7. Can be configured with the following extensible markup language (XML) snippet in the guest XML:

       <interface type='network'>
         <source network='default'/>
         <model type='virtio'/>
       </interface>
   

You can configure MacVtap in three different modes:

• Virtual Ethernet Port Aggregator (VEPA)
• Bridge (most commonly used)
• Private mode

To configure MacVtap, complete the following steps:

1. Define a network by using the following network XML code:

   <network>
     <name>macvtapnet</name>
     <forward dev='enP32769p1s0' mode='bridge'>
       <interface dev='enP32769p1s0'/>
     </forward>
   </network>
   

2. Put the following XML code in the XML of the guest:

       <interface type='network'>
         <source network='macvtapnet'/>
         <model type='virtio'/>
       </interface>
   

You can enable KVM by using the eBMC ASMI GUI, complete the following steps:

1. Connect to the eBMC ASMI GUI at https://<eBMC IP>.
2. Navigate to Server power operations.
3. From the Default partition environment menu, select Linux KVM.
4. Click Save.

You can enable KVM by using the HMC GUI, complete the following steps:

1. Connect to the HMC GUI.
2. Select the managed system.
3. Select the LPAR.
4. Click Advanced Settings.
5. Select KVM Capable.

You can enable or disable KVM by using the HMC command line interface (CLI), complete the following steps:

1. Login in through the CLI by running the following command: ssh hscroot@<IP>
2. Run any of the following required commands:
   • To enable KVM, run the following command: chsyscfg -r lpar -m <Managed System> -i "name=<lparname>,kvm_capable=1"
   • To disable KVM, run the following command: chsyscfg -r lpar -m <Managed System> -i "name=<lparname>,kvm_capable=0"
   • To check LPAR mode, run the following command: lssyscfg -r lpar -m <Managed System> | grep <lparname>

     Sample output

     #lssyscfg -r lpar -m ltcden6 | grep ltcden6-lp1

     name=ltcden6-lp1,lpar_id=1,lpar_env=aixlinux,state=Running,resource_config=1,os=linux,os_version=Unknown,
     logical_serial_num=134C9481,default_profile=default_profile,curr_profile=default_profile,work_group_id=none,
     shared_proc_pool_util_auth=0,allow_perf_collection=0,power_ctrl_lpar_ids=none,boot_mode=norm,lpar_keylock=norm,
     auto_start=0,redundant_err_path_reporting=0,rmc_state=inactive,rmc_ipaddr=,time_ref=0,lpar_avail_priority=127,
     desired_lp ar_proc_compat_mode=default,curr_lpar_proc_compat_mode=POWER10,sync_curr_profile=1,affinity_group_id=none,
     vtpm_enabled=0,migr_storage_vios_data_status=Re-collect Failed,migr_storage_vios_data_timestamp=unavailable,
     powervm_mgmt_capable=0,pend_secure_boot=0,curr_secure_boot=0,keystore_kbytes=0,keystore_signed_updates=0,
     keystore_signed_updates_without_verification=0,linux_dynamic_key_secure_boot=0,virtual_serial_num=none,
     kvm_capable=1,description=

You can use one of the following methods to install and setup a KVM guest:

1. Create a KVM guest by using an existing cloud (qcow2) image.
   1. Download a cloud qcow2 image from Fedora or other KVM enabled Linux distribution repositories.

      For example, https://pubmirror2.math.uh.edu/fedora-buffet/fedora-secondary/releases/40/Server/ppc64le/images/

   2. Prepare the cloud image and set the root password by running the following command: $ virt-customize --add Fedora-Server-KVM-40.ppc64le.qcow2 --root-password password:passw0rd
   3. Create the guest with required resources. The following command example creates a guest that is named Fedora-40 with 4 GB RAM and 4 virtual CPUs: $ virt-install --name Fedora-40 --ram 4096 --disk path=/home/test/Fedora-Server-KVM-40.ppc64le.qcow2 --vcpus 4 --os-variant generic --network bridge=virbr0 --graphics none --console pty,target_type=serial –import
2. Create a KVM Guest by using an installation image (ISO).
   1. Download a DVD-based ISO image for IBM Power ppc64le at https://fedoraproject.org/server/download.
   2. Create a disk image.

      The following example uses qcow2 as a format (-f) with 40 GB size: $ qemu-img create -f qcow2 Fedora-40.qcow2 40G

   3. Install the Linux operating system.

      The following example uses ISO (Fedora-Server-dvd-ppc64le-40.iso) and disk (Fedora-40.qcow2) image that were created in previous steps: $ virt-install --name Fedora_40 --memory 4096 --vcpus 4 --os-variant fedora40 --network bridge=virbr0 --disk path=/home/test/Fedora-40.qcow2,size=8 --graphics none --cdrom /home/test/Fedora-Server-dvd-ppc64le-40.iso

   4. Follow the installation screens to complete the Linux OS installation.

To list the running guests, run the following command:

$ virsh list --all
 Id   Name          State
------------------------------
 1    Fedora_40     running

To connect to the guest console, run the following command:

$ virsh console <guest_id>  
$ virsh console 1
Connected to domain 'Fedora_40'
Escape character is ^] (Ctrl + ])

localhost login: 

To start a guest, run the following command:

$ virsh start <domain-name> --console

To stop a guest, run the following command:

$ virsh shutdown <domain-name>

To delete a guest, run the following command:

$ virsh undefine <domain name>

• Fedora Discussion
• Ubuntu 24.04 LTS Release Notes