Jump to contentJump to page navigation: previous page [access key p]/next page [access key n]
documentation.suse.com / SUSE Linux Enterprise Serverマニュアル / Virtualization Guide / Managing Virtual Machines with QEMU / Setting Up a KVM VM Host Server
Applies to SUSE Linux Enterprise Server 12 SP5

27 Setting Up a KVM VM Host Server

This section documents how to set up and use SUSE Linux Enterprise Server 12 SP5 as a QEMU-KVM based virtual machine host.

Tip
Tip: Resources

In general, the virtual guest system needs the same hardware resources as if it were installed on a physical machine. The more guests you plan to run on the host system, the more hardware resources—CPU, disk, memory, and network—you need to add to the VM Host Server.

27.1 CPU Support for Virtualization

To run KVM, your CPU must support virtualization, and virtualization needs to be enabled in BIOS. The file /proc/cpuinfo includes information about your CPU features.

To find out whether your system supports virtualization, see Section 7.3, “KVM Hardware Requirements”.

27.2 Required Software

The KVM host requires several packages to be installed. To install all necessary packages, do the following:

  1. Verify that the yast2-vm package is installed. This package is YaST's configuration tool that simplifies the installation of virtualization hypervisors.

  2. Run YaST ›  Virtualization › Install Hypervisor and Tools.

    Installing the KVM Hypervisor and Tools
    Figure 27.1: Installing the KVM Hypervisor and Tools
  3. Select KVM server and preferably also KVM tools, and confirm with Accept.

  4. During the installation process, you can optionally let YaST create a Network Bridge for you automatically. If you do not plan to dedicate an additional physical network card to your virtual guests, network bridge is a standard way to connect the guest machines to the network.

    Network Bridge
    Figure 27.2: Network Bridge
  5. After all the required packages are installed (and new network setup activated), try to load the KVM kernel module relevant for your CPU type—kvm-intel or kvm-amd:

    root # modprobe kvm-intel

    Check if the module is loaded into memory:

    tux > lsmod | grep kvm
    kvm_intel              64835  6
    kvm                   411041  1 kvm_intel

    Now the KVM host is ready to serve KVM VM Guests. For more information, see Chapter 29, Running Virtual Machines with qemu-system-ARCH.

27.3 KVM Host-Specific Features

You can improve the performance of KVM-based VM Guests by letting them fully use specific features of the VM Host Server's hardware (paravirtualization). This section introduces techniques to make the guests access the physical host's hardware directly—without the emulation layer—to make the most use of it.

Tip
Tip

Examples included in this section assume basic knowledge of the qemu-system-ARCH command line options. For more information, see Chapter 29, Running Virtual Machines with qemu-system-ARCH.

27.3.1 Using the Host Storage with virtio-scsi

virtio-scsi is an advanced storage stack for KVM. It replaces the former virtio-blk stack for SCSI devices pass-through. It has several advantages over virtio-blk:

Improved scalability

KVM guests have a limited number of PCI controllers, which results in a limited number of possibly attached devices. virtio-scsi solves this limitation by grouping multiple storage devices on a single controller. Each device on a virtio-scsi controller is represented as a logical unit, or LUN.

Standard command set

virtio-blk uses a small set of commands that need to be known to both the virtio-blk driver and the virtual machine monitor, and so introducing a new command requires updating both the driver and the monitor.

By comparison, virtio-scsi does not define commands, but rather a transport protocol for these commands following the industry-standard SCSI specification. This approach is shared with other technologies, such as Fibre Channel, ATAPI, and USB devices.

Device naming

virtio-blk devices are presented inside the guest as /dev/vdX, which is different from device names in physical systems and may cause migration problems.

virtio-scsi keeps the device names identical to those on physical systems, making the virtual machines easily relocatable.

SCSI device pass-through

For virtual disks backed by a whole LUN on the host, it is preferable for the guest to send SCSI commands directly to the LUN (pass-through). This is limited in virtio-blk, as guests need to use the virtio-blk protocol instead of SCSI command pass-through, and, moreover, it is not available for Windows guests. virtio-scsi natively removes these limitations.

27.3.1.1 virtio-scsi Usage

KVM supports the SCSI pass-through feature with the virtio-scsi-pci device:

qemu-system-x86_64 [...] \
-device virtio-scsi-pci,id=scsi

27.3.2 Accelerated Networking with vhost-net

The vhost-net module is used to accelerate KVM's paravirtualized network drivers. It provides better latency and greater network throughput. Use the vhost-net driver by starting the guest with the following example command line:

qemu-system-x86_64 [...] \
-netdev tap,id=guest0,vhost=on,script=no \
-net nic,model=virtio,netdev=guest0,macaddr=00:16:35:AF:94:4B

Note that guest0 is an identification string of the vhost-driven device.

27.3.3 Scaling Network Performance with Multiqueue virtio-net

As the number of virtual CPUs increases in VM Guests, QEMU offers a way of improving the network performance using multiqueue. Multiqueue virtio-net scales the network performance by allowing VM Guest virtual CPUs to transfer packets in parallel. Multiqueue support is required on both the VM Host Server and VM Guest sides.

Tip
Tip: Performance Benefit

The multiqueue virtio-net solution is most beneficial in the following cases:

  • Network traffic packets are large.

  • VM Guest has many connections active at the same time, mainly between the guest systems, or between the guest and the host, or between the guest and an external system.

  • The number of active queues is equal to the number of virtual CPUs in the VM Guest.

Note
Note

While multiqueue virtio-net increases the total network throughput, it increases CPU consumption as it uses of the virtual CPU's power.

Procedure 27.1: How to Enable Multiqueue virtio-net

The following procedure lists important steps to enable the multiqueue feature with qemu-system-ARCH. It assumes that a tap network device with multiqueue capability (supported since kernel version 3.8) is set up on the VM Host Server.

  1. In qemu-system-ARCH, enable multiqueue for the tap device:

    -netdev tap,vhost=on,queues=2*N

    where N stands for the number of queue pairs.

  2. In qemu-system-ARCH, enable multiqueue and specify MSI-X (Message Signaled Interrupt) vectors for the virtio-net-pci device:

    -device virtio-net-pci,mq=on,vectors=2*N+2

    where the formula for the number of MSI-X vectors results from: N vectors for TX (transmit) queues, N for RX (receive) queues, one for configuration purposes, and one for possible VQ (vector quantization) control.

  3. In VM Guest, enable multiqueue on the relevant network interface (eth0 in this example):

    ethtool -L eth0 combined 2*N

The resulting qemu-system-ARCH command line will look similar to the following example:

qemu-system-x86_64 [...] -netdev tap,id=guest0,queues=8,vhost=on \
-device virtio-net-pci,netdev=guest0,mq=on,vectors=10

Note that the id of the network device (guest0 ) needs to be identical for both options.

Inside the running VM Guest, specify the following command as root:

ethtool -L eth0 combined 8

Now the guest system networking uses the multiqueue support from the qemu-system-ARCH hypervisor.

27.3.4 VFIO: Secure Direct Access to Devices

Directly assigning a PCI device to a VM Guest (PCI pass-through) avoids performance issues caused by avoiding any emulation in performance-critical paths. VFIO replaces the traditional KVM PCI Pass-Through device assignment. A prerequisite for this feature is a VM Host Server configuration as described in Important: Requirements for VFIO and SR-IOV.

To be able to assign a PCI device via VFIO to a VM Guest, you need to find out which IOMMU Group it belongs to. The IOMMU (input/output memory management unit that connects a direct memory access-capable I/O bus to the main memory) API supports the notion of groups. A group is a set of devices that can be isolated from all other devices in the system. Groups are therefore the unit of ownership used by VFIO.

Procedure 27.2: Assigning a PCI Device to a VM Guest via VFIO
  1. Identify the host PCI device to assign to the guest.

    tux > sudo lspci -nn
    [...]
    00:10.0 Ethernet controller [0200]: Intel Corporation 82576 \
    Virtual Function [8086:10ca] (rev 01)
    [...]

    Note down the device ID (00:10.0 in this case) and the vendor ID (8086:10ca).

  2. Find the IOMMU group of this device:

    tux > sudo readlink /sys/bus/pci/devices/0000\:00\:10.0/iommu_group
    ../../../kernel/iommu_groups/20

    The IOMMU group for this device is 20. Now you can check the devices belonging to the same IOMMU group:

    ls -l /sys/bus/pci/devices/0000:01:10.0/iommu_group/devices/0000:01:10.0
    [...] 0000:00:1e.0 -> ../../../../devices/pci0000:00/0000:00:1e.0
    [...] 0000:01:10.0 -> ../../../../devices/pci0000:00/0000:00:1e.0/0000:01:10.0
    [...] 0000:01:10.1 -> ../../../../devices/pci0000:00/0000:00:1e.0/0000:01:10.1
  3. Unbind the device from the device driver:

    sudo echo "0000:01:10.0" > /sys/bus/pci/devices/0000\:01\:10.0/driver/unbind
  4. Bind the device to the vfio-pci driver using the vendor ID from step 1:

    sudo echo "8086 153a" > /sys/bus/pci/drivers/vfio-pci/new_id

    A new device /dev/vfio/IOMMU_GROUP will be created as a result, /dev/vfio/20 in this case.

  5. Change the ownership of the newly created device:

    chown qemu.qemu /dev/vfio/DEVICE
  6. Now run the VM Guest with the PCI device assigned.

    qemu-system-ARCH [...] -device
         vfio-pci,host=00:10.0,id=ID
Important
Important: No Hotplugging

As of SUSE Linux Enterprise Server 12 SP5 hotplugging of PCI devices passed to a VM Guest via VFIO is not supported.

You can find more detailed information on the VFIO driver in the /usr/src/linux/Documentation/vfio.txt file (package kernel-source needs to be installed).

27.3.5 VirtFS: Sharing Directories between Host and Guests

VM Guests usually run in a separate computing space—they are provided their own memory range, dedicated CPUs, and file system space. The ability to share parts of the VM Host Server's file system makes the virtualization environment more flexible by simplifying mutual data exchange. Network file systems, such as CIFS and NFS, have been the traditional way of sharing directories. But as they are not specifically designed for virtualization purposes, they suffer from major performance and feature issues.

KVM introduces a new optimized method called VirtFS (sometimes called file system pass-through). VirtFS uses a paravirtual file system driver, which avoids converting the guest application file system operations into block device operations, and then again into host file system operations.

You typically use VirtFS for the following situations:

  • To access a shared directory from several guests, or to provide guest-to-guest file system access.

  • To replace the virtual disk as the root file system to which the guest's RAM disk connects during the guest boot process.

  • To provide storage services to different customers from a single host file system in a cloud environment.

27.3.5.1 Implementation

In QEMU, the implementation of VirtFS is simplified by defining two types of devices:

  • virtio-9p-pci device which transports protocol messages and data between the host and the guest.

  • fsdev device which defines the export file system properties, such as file system type and security model.

Example 27.1: Exporting Host's File System with VirtFS
qemu-system-x86_64 [...] \
-fsdev local,id=exp11,path=/tmp/2,security_model=mapped3 \
-device virtio-9p-pci,fsdev=exp14,mount_tag=v_tmp5

1

Identification of the file system to be exported.

2

File system path on the host to be exported.

3

Security model to be used—mapped keeps the guest file system modes and permissions isolated from the host, while none invokes a pass-through security model in which permission changes on the guest's files are reflected on the host as well.

4

The exported file system ID defined before with -fsdev id= .

5

Mount tag used later on the guest to mount the exported file system.

Such an exported file system can be mounted on the guest as follows:

sudo mount -t 9p -o trans=virtio v_tmp /mnt

where v_tmp is the mount tag defined earlier with -device mount_tag= and /mnt is the mount point where you want to mount the exported file system.

27.3.6 KSM: Sharing Memory Pages between Guests

Kernel Same Page Merging (KSM) is a Linux kernel feature that merges identical memory pages from multiple running processes into one memory region. Because KVM guests run as processes under Linux, KSM provides the memory overcommit feature to hypervisors for more efficient use of memory. Therefore, if you need to run multiple virtual machines on a host with limited memory, KSM may be helpful to you.

KSM stores its status information in the files under the /sys/kernel/mm/ksm directory:

tux > ls -1 /sys/kernel/mm/ksm
full_scans
merge_across_nodes
pages_shared
pages_sharing
pages_to_scan
pages_unshared
pages_volatile
run
sleep_millisecs

For more information on the meaning of the /sys/kernel/mm/ksm/* files, see /usr/src/linux/Documentation/vm/ksm.txt (package kernel-source).

To use KSM, do the following.

  1. Although SLES includes KSM support in the kernel, it is disabled by default. To enable it, run the following command:

    root # echo 1 > /sys/kernel/mm/ksm/run
  2. Now run several VM Guests under KVM and inspect the content of files pages_sharing and pages_shared, for example:

    while [ 1 ]; do cat /sys/kernel/mm/ksm/pages_shared; sleep 1; done
    13522
    13523
    13519
    13518
    13520
    13520
    13528