Hot answers tagged

17

If you are running server 2012 then you already have access to hyper-v. The free hyper-v should serve your needs, you can run the GUI from any client or the 2012 server. This would also get you shared nothing migrations if you can't afford a san. ...


11

You can't use switch port security on the Cisco since all the VMs will be sharing a physical switch port. And you can't use Linux iptables because the traffic is being bridged, not routed, through the hypervisor machine. But you can emulate switch port security on the hypervisor with Linux ebtables, which is a lesser-known layer 2/3 firewall on the Linux ...


9

You can use VMware ESXi for free... You don't necessarily need to use the web interface, and the product works very well with the thick client. Or better, buy a vSphere Essentials license ($500) and have the ability to manage up to 3 host servers. This will give you vCenter capabilities and the Web interface if you wish to utilize it. Either way, the cost ...


8

This is how I currently do installs of Windows with virtio drivers at install: Create the new VM with a genuine Windows installation ISO image. I cannot guarantee that anything will work if you use a pirated copy of Windows or if your ISO image download is incomplete or corrupt. Choose to customize the machine configuration before install. Select the ...


6

Use proxmox, an hypervisor based on KVM (other containers types also supported) all integrated via a web interface. We use this on dell rack servers and it is just perfect for small companies. You can also buy paid support. Note that Proxmox is based on RedHat long-term support kernels, which is synonym for stability and good driver support for servers. ...


6

Most current hypervisors that run on Intel hardware use CPUID leaves 0x40000000 et seq. to pass information about the hypervisor from host to guest. KVM, Xen, VMware and Hyper-V all use this method. This is in addition to the hypervisor feature flag set in CPUID leaf 0x1, which indicates that the machine is a virtual machine. It's therefore trivial for a ...


6

While it might not look like one, Hyper-V actually is a type-1 hypervisor, because the virtualization layer is placed below the Windows operating system and mediates all hardware access (some time ago, before hardware-assisted virtualization improved a lot, it was very common to experience really slow graphic performance on Hyper-V hosts, due to the O.S. ...


6

You have no way of predicting what the device names or IDs will be after a significant hardware change (virtual hardware in this case, from Xen to KVM). Linode is the only entity here who knows what's going to change and how you should proceed if you upgrade. With that said, don't do this. You're forcing yourself to take an outage. If it breaks, you'll be ...


5

For your specific use case (base image + qcow2 overlay), the RAW format should be preferred: It's faster: as it has no metadata associated, it is as fast as possible. On the other hand, Qcow2 has two layer of indirection that must be crossed before to hit the actual data As the overlay layer must be a Qcow2 file, you don't lose the ever-useful snapshot ...


5

It's difficult to troubleshoot remotely with limited information. But I'd like to try. First, just a guess. Usually the DEVICE name is eth0 or eth1, not eth01. This might explain the "invalid argument error". Make sure you are dealing with the right NIC by ifconfig -a or ip link in the VM. Another suspect is the static route. It should have 13 items in ...


4

Here is a serious problem: <domain type='qemu'> This should appear as: <domain type='kvm'> If this is set to qemu, then qemu will run the machine without any sort of hardware virtualization support. Without this option set, qemu will not be told to enable KVM hardware acceleration. This is fine for, e.g. emulating non-Intel processors, but ...


4

You can, but you probably shouldn't. The DL360 G7 doesn't present disks in a JBOD fashion. If you're using the onboard HP Smart Array controller, this won't work the way you expect. Depending on the OS you use, there are swap and other I/O activity considerations. Why go through all of this? If you want ZFS, just use it on the server in a baremetal OS ...


4

Virtio is a para-virtualized driver, which means the os and driver are aware that it's not a physical Device. The driver is really an API between the guest and the hypervisor so it's linkspeed is totally disconnect from any physical device or Ethernet standard. This is a good thing as this is faster than the hypervisor pretending to be a physical device and ...


4

You created your guest with virsh create (or its equivalent). This creates transient domains, which are deleted when they power off. To create persistent domains, use virsh define instead. These remain defined after they are powered off or destroyed, and can be started again at any time.


4

Online Method (using qemu, libvirt, and virtio-block) Perform the following from the KVM hypervisor. Increase the size of the disk image file itself (specify the amount to increase): qemu-img resize <my_vm>.img +10G Get the name of the virtio device, via the libvirt shell (drive-virtio-disk0 in this example): virsh qemu-monitor-command ...


3

A more "official"[1] way to do this is to create a hook script as described at the libvirt website: http://wiki.libvirt.org/page/Networking#Forwarding_Incoming_Connections ... basically this script will be invoked when a KVM guest is booted-up. The script itself will add the appropriate iptable rules (similar to Isaac Sutherland's answer above) with the ...


3

This feature was requested a long time ago. Now libvirt supports it by providing two new commands: domifaddr and net-dhcp-leases Usage: domifaddr <domain> [interface] [--full] [--source lease|agent] Example outputs: virsh # domifaddr f20 --source agent Name MAC address Protocol Address ...


3

To change the state of the network link for a KVM guest it would be much better to use the appropriate tool, i.e. the hypervisor virsh domif-setlink domain interface-device state Modify link state of the domain’s virtual interface. Possible values for state are "up" and "down". If --config is specified, only the persistent configuration of the ...


3

According to the community wiki (https://help.ubuntu.com/community/32bit_and_64bit), Ubuntu itself is limited to about 1TB of memory. I would assume this means that it has trouble assigning more memory than that to a guest OS.


3

You're seeing this error because Ubuntu put their own customizations into its version of libvirt and qemu/KVM. RHEL, of course, did the same thing. But there's always a machine type that will work, namely, pc. This is always aliased to the latest i440fx machine type available on the hypervisor at the time the VM is defined. Somewhere in the XML you'll find ...


3

-device e1000,netdev=hostnet0 This. Switch to virtio_net before you try anything else. step two: try to play with the NIC offload settings using ethtool -K turn off TSO, LSO, LRO, see if you get different results. step three: try a different distro. Ubuntu is notorious for KVM related bugs, so switch to CentOS or even Fedora as a test.


3

Check that your VMs have ip addresses on 10.x.x.x/24 (netmask 255.255.255.0) Set 10.x.x.11 (br0 ip address) as the default gateway of your VMs Enable ip forwarding on the physical host Enable SNAT with: iptables -t nat -A POSTROUTING -s 10.x.x.x/24 -o eth1 -j SNAT --to y.y.y.102


3

You may be overthinking this. The manual assignment of cores here could actually result in lower performance. In the VMware world, we don't do this unless there are very specific requirements, but for the workload and applications you've described, it's not necessary. Let KVM schedule things and be done. If in doubt, get more cores and sockets. But CPU ...


3

Not at present. It looks like CPU hotplug used to be possible in QEMU, but the patch is old and crusty, never worked with windows guests, and not shipped natively anymore. http://www.linux-kvm.org/page/CPUHotPlug


3

What you need to do is to resize the disk using the virt-resize tool. Install libguestfs-tools and perform the following steps: 1- Create a sparse disk 2- Inspect the template disk filesystem using "virt-filesystems" command in order to select which partition you want to resize 3- Perform the resizing process Here is an example: # qemu-img create -f ...


3

I can't say how well Windows works with KVM, but I think it's worth testing. There are PV drivers available for Windows. It wouldn't be so bad to setup a VM and run some benchmarks. KVM is nice because it's built into the kernel, unlike Xen. Performance is about the same. I find it easy to use via libvirt and the "Virtual Machine Manager" GUI. It can ...


3

As others already suggested, if you already have a Windows Server license, you thus have an enterprise-grade virtualization platform in your possession. If it is a Standard edition, then you can use it this way: (1) install one instance of Windows Server on a physical server with up to two physical sockets (number of cores doesn't matter, matter processor ...


3

Is it safe to boot Linux from a USB drive? Yes, why not? I do it regularly. But having said that, it's not entirely trivial to do it well. If you just straight install Linux to a USB disk it's usually really slow and doesn't really work that great. However, if you use a tool like unetbootin it'll create a more optimal installation with which you can ...


3

Define "safe". As in: Does the system boot from USB? Yes Data integrity? No Fault tolerance? No I would not trust a file server to be booted from a single USB pen, unless it's a server just for testing/lab etc. We use USB drives (expensive ones) in our vSphere Cluster, and I've encountered failures many times. You've indicated that you have replaced the ...



Only top voted, non community-wiki answers of a minimum length are eligible