Chapitre 3. KVM Support Status

Table des matières

3.1. Supported Features and Tools
3.2. Unsupported Features and Tools

The following list contains features and tools as supported by SUSE—this does not necessarily reflect the support status of the software itself. For a list of qemu-kvm command switches supported by SUSE, refer to Section A.3, « QEMU Command Line Options ».

3.1. Supported Features and Tools

vm-install

Define and install VM Guests via vm-install including specifying the number of virtual processors, RAM, disk type and location, video type, keyboard mapping, NIC type, binding, MAC address, and boot method.

Restrictions: Currently only the raw, qcow2 and qed disk formats are supported. NIC creation is restricted to using Realtek, e1000 or virtio NICs.

Virtual Machine Manager

Manage guests via Virtual Machine Manager using the following functions: autostart, start, stop, restart, pause, unpause, save, restore, clone, migrate, special key sequence insertion, guest console viewers, performance monitoring, and CPU pinning. Furthermore, static modifications of CPU, RAM, boot method, disk, NIC, mouse, display, video and host PCI and USB assignments are supported.

Restrictions: The following features are currently not supported: sound devices, qxl, vmvga (vmware), Xen video, emulated SCSI and SATA disks, vmdk storage format, Spice graphics, eepro100, ne2k_pci, pcnet, Smartcard and USB redirection.

virsh

Manage guests via the command line.

Restrictions: Requires XML descriptions as created by vm-install or virt-manager. Altering these descriptions via virsh edit is not supported. The supported virsh functionality is restricted to life cycle functions.

qemu-kvm

Manage guests via the command line. Although managing via Virtual Machine Manager should be the preferred option, qemu-kvm may be used for greater flexibility. See Section A.3.1, « Supported qemu-kvm Command Line Options » for a list of supported options.

Restrictions: See Section A.3.2, « Unsupported qemu-kvm Command Line Options » for a list of not supported options.

kvm_stat

Debugging and monitoring tool.

USB Host Device Pass-through

A physical USB device may be passed from the VM Host Server to the VM Guest.

PCI Pass-through

PCI Pass-through improves performance of PCI devices. It requires an AMD CPU with IOMMU (I/O Memory Mapping Unit) or an Intel CPU with VT-d (Virtualization Technology for Directed I/O). VT-d requires the Kernel parameter "intel_iommu=on". Many PCIe cards from major vendors should be supportable. Refer to system level certifications for specific details, or contact the vendor for support statements.

Device hotplugging

Dynamically adding or removing emulated or pass-through physical devices in the VM Guest is supported.

Sharing Folders between VM Host Server and VM Guest

Sharing folders between host and guest is supported via VirtFs.

Kernel Samepage Merging (KSM)

KSM allows for automatic sharing of identical memory pages between guests to save host memory. KVM is optimized to use KSM if enabled on the VM Host Server.

Transparent Huge Pages (THP)

THP allows CPUs to address memory using pages larger than the default 4 KB. This helps reducing memory consumption and CPU cache usage. KVM is optimized to use THP (via madvise and opportunistic methods) if enabled on the VM Host Server.

Non-Uniform Memory Access (NUMA)

NUMA machines are supported. Using numactl to pin qemu-kvm processes to specific nodes is recommended.

3.2. Unsupported Features and Tools

CPU hotplugging

Dynamically changing the number of virtual CPUs assigned to the guest is currently not supported.

AHCI guest storage

The AHCI interface for SATA storage provides fast I/O performance. It is currently not supported.

Online Disk resizing

Online Disk resizing is currently not supported.

KVM Kernel Module Parameters

Specifying parameters for the KVM Kernel modules is currently not supported unless done under the direction of SUSE support personnel.

Guest Agent

The guest agent (qemu-ga) allows programs on the VM Host Server to directly communicate with a guest via an emulated or para-virtualized serial console. This feature is currently not supported.

Using QEMU without KVM

qemu-kvm can be invoked with the -no-kvm parameter. In this case guest CPU instructions are emulated instead of being executed directly by the processor. This mode is not supported, but may be useful for problem resolution.