Linked clone with qemu-img

2018-04-09 2 min read Vurtualization

As you would have seen in Virtualbox or vmware, there is option to create a linked clone. I wanted to use the same feature as “Snapshot” feature anyway does not look/work so great with virt-manager. So, I created a script to create a linked clone VM and here it is :

 

#!/bin/bash - 
#===============================================================================
#
#          FILE: qcow2-linked-clone.sh
# 
#         USAGE: ./qcow2-linked-clone.sh 
# 
#   DESCRIPTION: 
# 
#       OPTIONS: ---
#  REQUIREMENTS: ---
#          BUGS: ---
#         NOTES: ---
#        AUTHOR: Amit Agarwal (aka)
#  ORGANIZATION: Mobileum
#       CREATED: 01/05/2018 09:54
# Last modified: Wed Feb 28, 2018  04:39PM
#      REVISION:  ---
#===============================================================================

set -o nounset                              # Treat unset variables as an error
UP="amitag:amitag" #Here you need to put your username and group.

if [[ $# == 0 ]]
then
    read -p "Enter the source path :: " spath
    read -p "Enter the source disk :: " sdisk
    read -p "Enter the destin path :: " dpath
    read -p "Enter the destin disk :: " ddisk
    read -p "Enter new VMName :: " vmname
else
    spath=$(dirname $1)
    dpath=$spath
    sdisk=$(basename $1)
    ddisk=$2.qcow2
    vmname=$2
fi


sudo chown $UP "$spath/$sdisk"
qemu-img create -f qcow2 -b "$spath/$sdisk" "$dpath/$ddisk"

virt-install --disk $dpath/$ddisk --ram 512 \
    --virt-type kvm --vcpus 1 --name "$vmname" --import

The script will create a linked qcow2 and then create a VM with that image. Running it is simple, either provide command line options or just run and it will ask you for details.

Continue reading

vagrant box to libvirtd (QEMU) VM

2018-03-26 1 min read Fedora Vurtualization

Like ova images, you can use box images as well with Qemu. After all, both have the disk images, so here is the script to do that. Just put the script somewhere in your path and run with ova or box image name :

 

#!/bin/bash - 
#===============================================================================
#
#          FILE: ova2vm.sh
# 
#         USAGE: ./ova2vm.sh 
# 
#   DESCRIPTION: 
# 
#       OPTIONS: ---
#  REQUIREMENTS: ---
#          BUGS: ---
#         NOTES: ---
#        AUTHOR: Amit Agarwal (aka),
#  ORGANIZATION: Mobileum
#       CREATED: 12/28/2017 13:59
# Last modified: Sun Mar 11, 2018  12:01PM
#      REVISION:  ---
#===============================================================================

set -o nounset                              # Treat unset variables as an error
dest='/mnt/Backup/VM'
ORIG=${PWD}

if [[ $# == 0 ]]
then
    echo "You need to provide ova/vmdk filename"
    exit
fi
if [[ $1 == *ova || $1 == *box ]]
then
    tmp=$(mktemp -d /tmp/amitXXXXXXX)
    cd  $tmp
    tar xvf $ORIG/$1
    file=$(echo $PWD/*vmdk)
else
    file=$1
    echo "Not a OVA file"
fi
dfile="$dest/$(basename $file)"

read -p "Enter the name for VM :: " vmname
qemu-img convert $file $dfile -p -c -O qcow2
virt-install --disk $dfile --ram 512 \
    --virt-type kvm --vcpus 1 --name "$vmname" --import

mv command with progress

2018-03-19 1 min read Bash

When moving large files/directories, I would like to see the progress.

Idea for this is to use rsync with progress and remove source files. But that option does not remove the empty directories left behind so find command to delete that.

So, here is function for that:

mv-progress () 
{ 
    rsync -ah --progress --remove-source-files "$1" "$2";
    find "$1" -empty -delete
}

Highest disk usage of directory in subdirectories

2018-03-12 1 min read Bash

I find myself doing this lot of times so thought will share this with you all. Basically, once I want to clear out the directory, I first want to find out the sub-directory using the maximum disk space so I wrote a function for that and here it is:

 

disk_usage_dirs () 
{ 
    find . -maxdepth 1 -type d -not -name '.' | while read line; do
        du -s "$line";
    done | sort -n | tail -${1:-5}
}

Some other posts you might find useful on this :

Continue reading

Send history of current host to some other host over ssh

2018-02-26 1 min read Fedora Learning

Sometimes I want to save the history of current host on another host. This is to ensure that I can use copy/paste on other host to run the commands. To this, I found a simple solution –

history| ssh <user>@<host> 'cat - > /tmp/history'

And on the new host, you can find the history in file “/tmp/history”, cool :). Now I can quick edit this file to create this as shell script as well if required. How cool is that.

Continue reading
Older posts Newer posts