Setup kvm » Historie » Revision 2
Revision 1 (Jeremias Keihsler, 26.08.2021 10:49) → Revision 2/5 (Jeremias Keihsler, 26.08.2021 10:50)
h1. KVM this is for a vanilla CentOS 8 7 minimal installation, largely based on @kvm_virtualization_in_rhel_7_made_easy.pdf@ good information is also found at http://virtuallyhyper.com/2013/06/migrate-from-libvirt-kvm-to-virtualbox/ h2. basic updates/installs <pre><code class="bash"> yum update yum install wget yum install vim reboot </code></pre> h2. check machine capability <pre><code class="bash"> grep -E 'svm|vmx' /proc/cpuinfo </code></pre> vmx ... Intel svm ... AMD h2. install KVM on CentOS minimal <pre><code class="bash"> yum install qemu-kvm libvirt libguestfs-tools virt-install systemctl enable libvirtd && systemctl start libvirtd </code></pre> verify the following kernel modules are loaded <pre><code class="bash"> lsmod | grep kvm </code></pre> <pre><code class="bash"> kvm kvm_intel </code></pre> <pre><code class="bash"> kvm kvm_amd </code></pre> h2. setup networking add to the network controller configuration file @/etc/sysconfig/network-scripts/ifcfg-em1@ <pre> ... BRIDGE=br0 </pre> add following new file @/etc/sysconfig/network-scripts/ifcfg-br0@ <pre> DEVICE="br0" # BOOTPROTO is up to you. If you prefer “static”, you will need to # specify the IP address, netmask, gateway and DNS information. BOOTPROTO="dhcp" IPV6INIT="yes" IPV6_AUTOCONF="yes" ONBOOT="yes" TYPE="Bridge" DELAY="0" </pre> enable network forwarding @/etc/sysctl.conf@ <pre> ... net.ipv4.ip_forward = 1 </pre> read the file and restart NetworkManager <pre><code class="bash"> sysctl -p /etc/sysctl.conf systemctl restart NetworkManager </code></pre> h2. can KVM and Virtualbox coexist http://www.dedoimedo.com/computers/kvm-virtualbox.html h2. convert Virtualbox to KVM h3. uninstall Virtualbox-guest-additions <pre><code class="bash"> opt/[VboxAddonsFolder]/uninstall.sh </code></pre> some people had to remove @/etc/X11/xorg.conf@ h3. convert image from Virtualbox to KWM <pre><code class="bash"> VBoxManage clonehd --format RAW Virt_Image.vdi Virt_Image.img </code></pre> RAW-Datei nach qcow konvertieren <pre><code class="bash"> qemu-img convert -f raw Virt_Image.img -O qcow2 Virt_Image.qcow </code></pre> h2. automatic start/shutdown of VMs with Host taken from https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization_Administration_Guide/sub-sect-Shutting_down_rebooting_and_force_shutdown_of_a_guest_virtual_machine-Manipulating_the_libvirt_guests_configuration_settings.html h3. enable libvirt-guests service <pre><code class="bash"> systemctl enable libvirt-guests systemctl start libvirt-guests </code></pre> all settings are to be done in @/etc/sysconfig/libvirt-guests@ h2. install <pre><code class="bash"> yum install virt-manager </code></pre> <pre><code class="bash"> usermod -a -G libvirt username </code></pre> h2. rename KVM-guest taken from http://www.taitclarridge.com/techlog/2011/01/rename-kvm-virtual-machine-with-virsh.html Power off the virtual machine and export the machine's XML configuration file: <pre><code class="bash"> virsh dumpxml name_of_vm > name_of_vm.xml </code></pre> Next, edit the XML file and change the name between the <name></name> tags (should be right near the top). As an added step you could also rename the disk file to reflect the change of the name and change the name of it in the <devices> section under <source file='/path/to/name_of_vm.img'>. Save the XML file and undefine the old VM name with: <pre><code class="bash"> virsh undefine name_of_vm </code></pre> Now just import the edited XML file to define the VM: <pre><code class="bash"> virsh define name_of_vm.xml </code></pre> And that should be it! You can now start up your vm either in the Virtual Machine Manager or with virsh using: <pre><code class="bash"> virsh start name_of_vm </code></pre> h2. set fixed IP-adr via DHCP (default-network) taken from https://wiki.libvirt.org/page/Networking <pre><code class="bash"> virsh edit <guest> </code></pre> where <guest> is the name or uuid of the guest. Add the following snippet of XML to the config file: <pre><code class="bash"> <interface type='network'> <source network='default'/> <mac address='00:16:3e:1a:b3:4a'/> </interface> </code></pre> Applying modifications to the network Sometimes, one needs to edit the network definition and apply the changes on the fly. The most common scenario for this is adding new static MAC+IP mappings for the network's DHCP server. If you edit the network with "virsh net-edit", any changes you make won't take effect until the network is destroyed and re-started, which unfortunately will cause a all guests to lose network connectivity with the host until their network interfaces are explicitly re-attached. virsh net-update Fortunately, many changes to the network configuration (including the aforementioned addition of a static MAC+IP mapping for DHCP) can be done with "virsh net-update", which can be told to enact the changes immediately. For example, to add a DHCP static host entry to the network named "default" mapping MAC address 53:54:00:00:01 to IP address 192.168.122.45 and hostname "bob", you could use this command: <pre><code class="bash"> virsh net-update default add ip-dhcp-host \ "<host mac='52:54:00:00:00:01' \ name='bob' ip='192.168.122.45' />" \ --live --config </code></pre> h2. forwarding incoming connections taken from https://wiki.libvirt.org/page/Networking By default, guests that are connected via a virtual network with <forward mode='nat'/> can make any outgoing network connection they like. Incoming connections are allowed from the host, and from other guests connected to the same libvirt network, but all other incoming connections are blocked by iptables rules. If you would like to make a service that is on a guest behind a NATed virtual network publicly available, you can setup libvirt's "hook" script for qemu to install the necessary iptables rules to forward incoming connections to the host on any given port HP to port GP on the guest GNAME: 1) Determine a) the name of the guest "G" (as defined in the libvirt domain XML), b) the IP address of the guest "I", c) the port on the guest that will receive the connections "GP", and d) the port on the host that will be forwarded to the guest "HP". (To assure that the guest's IP address remains unchanged, you can either configure the guest OS with static ip information, or add a <host> element inside the <dhcp> element of the network that is used by your guest. See the libvirt network XML documentation address section for defails and an example.) 2) Stop the guest if it's running. 3) Create the file /etc/libvirt/hooks/qemu (or add the following to an already existing hook script), with contents similar to the following (replace GNAME, IP, GP, and HP appropriately for your setup): Use the basic script below or see an "advanced" version, which can handle several different machines and port mappings here (improvements are welcome) or here's a python script which does a similar thing and is easy to understand and configure (improvements are welcome): <pre> #!/bin/bash # used some from advanced script to have multiple ports: use an equal number of guest and host ports # Update the following variables to fit your setup Guest_name=GUEST_NAME Guest_ipaddr=GUEST_IP Host_ipaddr=HOST_IP Host_port=( 'HOST_PORT1' 'HOST_PORT2' ) Guest_port=( 'GUEST_PORT1' 'GUEST_PORT2' ) length=$(( ${#Host_port[@]} - 1 )) if [ "${1}" = "${Guest_name}" ]; then if [ "${2}" = "stopped" ] || [ "${2}" = "reconnect" ]; then for i in `seq 0 $length`; do iptables -t nat -D PREROUTING -d ${Host_ipaddr} -p tcp --dport ${Host_port[$i]} -j DNAT --to ${Guest_ipaddr}:${Guest_port[$i]} iptables -D FORWARD -d ${Guest_ipaddr}/32 -p tcp -m state --state NEW -m tcp --dport ${Guest_port[$i]} -j ACCEPT done fi if [ "${2}" = "start" ] || [ "${2}" = "reconnect" ]; then for i in `seq 0 $length`; do iptables -t nat -A PREROUTING -d ${Host_ipaddr} -p tcp --dport ${Host_port[$i]} -j DNAT --to ${Guest_ipaddr}:${Guest_port[$i]} iptables -I FORWARD -d ${Guest_ipaddr}/32 -p tcp -m state --state NEW -m tcp --dport ${Guest_port[$i]} -j ACCEPT done fi fi </pre> 4) chmod +x /etc/libvirt/hooks/qemu 5) Restart the libvirtd service. 6) Start the guest. (NB: This method is a hack, and has one annoying flaw in versions of libvirt prior to 0.9.13 - if libvirtd is restarted while the guest is running, all of the standard iptables rules to support virtual networks that were added by libvirtd will be reloaded, thus changing the order of the above FORWARD rule relative to a reject rule for the network, hence rendering this setup non-working until the guest is stopped and restarted. Thanks to the new "reconnect" hook in libvirt-0.9.13 and newer (which is used by the above script if available), this flaw is not present in newer versions of libvirt (however, this hook script should still be considered a hack). h2. wrapper script for virsh <pre> #! /bin/sh # kvm_control Startup script for KVM Virtual Machines # # description: Manages KVM VMs # processname: kvm_control.sh # # pidfile: /var/run/kvm_control/kvm_control.pid # ### BEGIN INIT INFO # ### END INIT INFO # # Version 20171103 by Jeremias Keihsler added ionice prio 'idle' # Version 20161228 by Jeremias Keihsler based on: # virsh-specific parts are taken from: # https://github.com/kumina/shutdown-kvm-guests/blob/master/shutdown-kvm-guests.sh # Version 20110509 by Jeremias Keihsler (vboxcontrol) based on: # Version 20090301 by Kevin Swanson <kswan.info> based on: # Version 2008051100 by Jochem Kossen <jochem.kossen@gmail.com> # http://farfewertoes.com # # Released in the public domain # # This file came with a README file containing the instructions on how # to use this script. # # this is no more to be used as an init.d-script (vboxcontrol was an init.d-script) # ################################################################################ # INITIAL CONFIGURATION export PATH="${PATH:+$PATH:}/bin:/usr/bin:/usr/sbin:/sbin" VIRSH=/usr/bin/virsh TIMEOUT=300 declare -i VM_isrunning ################################################################################ # FUNCTIONS log_failure_msg() { echo $1 } log_action_msg() { echo $1 } # list running domains list_running_domains() { $VIRSH list | grep running | awk '{ print $2}' } # Check for running machines every few seconds; return when all machines are # down wait_for_closing_machines() { RUNNING_MACHINES=`list_running_domains | wc -l` if [ $RUNNING_MACHINES != 0 ]; then log_action_msg "machines running: "$RUNNING_MACHINES sleep 2 wait_for_closing_machines fi } ################################################################################ # RUN case "$1" in start) if [ -f /etc/kvm_box/machines_enabled_start ]; then cat /etc/kvm_box/machines_enabled_start | while read VM; do log_action_msg "Starting VM: $VM ..." $VIRSH start $VM sleep 20 RETVAL=$? done touch /tmp/kvm_control fi ;; stop) # NOTE: this stops first the listed VMs in the given order # and later all running VM's. # After the defined timeout all remaining VMs are killed # Create some sort of semaphore. touch /tmp/shutdown-kvm-guests echo "Try to cleanly shut down all listed KVM domains..." # Try to shutdown each listed domain, one by one. if [ -f /etc/kvm_box/machines_enabled_stop ]; then cat /etc/kvm_box/machines_enabled_stop | while read VM; do log_action_msg "Shutting down VM: $VM ..." $VIRSH shutdown $VM --mode acpi sleep 10 RETVAL=$? done fi sleep 10 echo "give still running machines some more time..." # wait 20s per still running machine list_running_domains | while read VM; do log_action_msg "waiting 20s ... for: $VM ..." sleep 20 done echo "Try to cleanly shut down all running KVM domains..." # Try to shutdown each remaining domain, one by one. list_running_domains | while read VM; do log_action_msg "Shutting down VM: $VM ..." $VIRSH shutdown $VM --mode acpi sleep 10 done # Wait until all domains are shut down or timeout has reached. END_TIME=$(date -d "$TIMEOUT seconds" +%s) while [ $(date +%s) -lt $END_TIME ]; do # Break while loop when no domains are left. test -z "$(list_running_domains)" && break # Wait a litte, we don't want to DoS libvirt. sleep 2 done # Clean up left over domains, one by one. list_running_domains | while read DOMAIN; do # Try to shutdown given domain. $VIRSH destroy $DOMAIN # Give libvirt some time for killing off the domain. sleep 10 done wait_for_closing_machines rm -f /tmp/shutdown-kvm-guests rm -f /tmp/kvm_control ;; export) JKE_DATE=$(date +%F) if [ -f /etc/kvm_box/machines_enabled_export ]; then cat /etc/kvm_box/machines_enabled_export | while read VM; do rm -f /tmp/kvm_control_VM_isrunning VM_isrunning=0 list_running_domains | while read RVM; do #echo "VM list -$VM- : -$RVM-" if [[ "$VM" == "$RVM" ]]; then #echo "VM found running..." touch /tmp/kvm_control_VM_isrunning VM_isrunning=1 #echo "$VM_isrunning" break fi #echo "$VM_isrunning" done # took me a while to figure out that the above 'while'-loop # runs in a separate process ... let's use the 'file' as a # kind of interprocess-communication :-) JKE 20161229 if [ -f /tmp/kvm_control_VM_isrunning ]; then VM_isrunning=1 fi rm -f /tmp/kvm_control_VM_isrunning #echo "VM status $VM_isrunning" if [ "$VM_isrunning" -ne 0 ]; then log_failure_msg "Exporting VM: $VM is not possible, it's running ..." else log_action_msg "Exporting VM: $VM ..." VM_BAK_DIR="$VM"_"$JKE_DATE" mkdir "$VM_BAK_DIR" $VIRSH dumpxml $VM > ./$VM_BAK_DIR/$VM.xml $VIRSH -q domblklist $VM | awk '{ print$2}' | while read VMHDD; do echo "$VM hdd=$VMHDD" if [ -f "$VMHDD" ]; then ionice -c 3 rsync --progress $VMHDD ./$VM_BAK_DIR/`basename $VMHDD` else log_failure_msg "Exporting VM: $VM image-file $VMHDD not found ..." fi done fi done else log_action_msg "export-list not found" fi ;; start-vm) log_action_msg "Starting VM: $2 ..." $VIRSH start $2 RETVAL=$? ;; stop-vm) log_action_msg "Stopping VM: $2 ..." $VIRSH shutdown $2 --mode acpi RETVAL=$? ;; poweroff-vm) log_action_msg "Powering off VM: $2 ..." $VIRSH destroy $2 RETVAL=$? ;; export-vm) # NOTE: this exports the given VM log_action_msg "Exporting VM: $2 ..." rm -f /tmp/kvm_control_VM_isrunning VM_isrunning=0 JKE_DATE=$(date +%F) list_running_domains | while read RVM; do #echo "VM list -$VM- : -$RVM-" if [[ "$2" == "$RVM" ]]; then #echo "VM found running..." touch /tmp/kvm_control_VM_isrunning VM_isrunning=1 #echo "$VM_isrunning" break fi #echo "$VM_isrunning" done # took me a while to figure out that the above 'while'-loop # runs in a separate process ... let's use the 'file' as a # kind of interprocess-communication :-) JKE 20161229 if [ -f /tmp/kvm_control_VM_isrunning ]; then VM_isrunning=1 fi rm -f /tmp/kvm_control_VM_isrunning #echo "VM status $VM_isrunning" if [ "$VM_isrunning" -ne 0 ]; then log_failure_msg "Exporting VM: $VM is not possible, it's running ..." else log_action_msg "Exporting VM: $VM ..." VM_BAK_DIR="$2"_"$JKE_DATE" mkdir "$VM_BAK_DIR" $VIRSH dumpxml $2 > ./$VM_BAK_DIR/$2.xml $VIRSH -q domblklist $2 | awk '{ print$2}' | while read VMHDD; do echo "$2 hdd=$VMHDD" if [ -f "$VMHDD" ]; then ionice -c 3 rsync --progress $VMHDD ./$VM_BAK_DIR/`basename $VMHDD` else log_failure_msg "Exporting VM: $2 image-file $VMHDD not found ..." fi done fi ;; status) echo "The following virtual machines are currently running:" list_running_domains | while read VM; do echo -n " $VM" echo " ... is running" done ;; *) echo "Usage: $0 {start|stop|status|export|start-vm <VM name>|stop-vm <VM name>|poweroff-vm <VM name>}|export-vm <VMname>" echo " start start all VMs listed in '/etc/kvm_box/machines_enabled_start'" echo " stop 1st step: acpi-shutdown all VMs listed in '/etc/kvm_box/machines_enabled_stop'" echo " 2nd step: wait 20s for each still running machine to give a chance to shut-down on their own" echo " 3rd step: acpi-shutdown all running VMs" echo " 4th step: wait for all machines shutdown or $TIMEOUT s" echo " 5th step: destroy all sitting VMs" echo " status list all running VMs" echo " export export all VMs listed in '/etc/kvm_box/machines_enabled_export' to the current directory" echo " start-vm <VM name> start the given VM" echo " stop-vm <VM name> acpi-shutdown the given VM" echo " poweroff-vm <VM name> poweroff the given VM" echo " export-vm <VM name> export the given VM to the current directory" exit 3 esac exit 0 </pre> h2. restore 'exported' kvm-machines <pre><code class="shell"> tar xvf mach-name_202x-01-01.tar.gz </code></pre> * copy the image-files to @/var/lib/libvirt/images/@ set ownership <pre><code class="shell"> chown qemu:qemu /var/lib/libvirt/images/* </code></pre> define the machine by <pre><code class="shell"> virsh define mach-name.xml </code></pre>