#1 Enable SSH on ESX(i) Host

To enable SSH on the ESX(i) host please follow this guide:

http://www.thomasmaurer.ch/2014/01/enable-ssh-on-vmware-esxi-5-5-via-vsphere-client/

#2 Create a Ubuntu VM

1. Download Ubuntu Server 12.04 and create a new VM
   Direct Downloads:
   • 32 bit: http://releases.ubuntu.com/12.04/ubuntu-12.04.5-server-i386.iso
   • 64 bit: http://releases.ubuntu.com/12.04/ubuntu-12.04.5-server-amd64.iso
2. Install Python 2.6.9 (latest) – the patch of VMWare ESX(i) I was using had Python 2.6.8:

   sudo apt-get install python-software-properties
   sudo add-apt-repository ppa:fkrull/deadsnakes
   sudo apt-get update
   sudo apt-get install python2.6
   wget https://bootstrap.pypa.io/ez_setup.py -O - | sudo python2.6

   If you experience DNS problems after setting up the VM, please see: http://miami-nice.co.uk/how-to-fix-ubuntu-dns-problems/

3. Install s3cmd: http://s3tools.org/download

   wget http://downloads.sourceforge.net/project/s3tools/s3cmd/1.6.1/s3cmd-1.6.1.tar.gz
   tar xzf s3cmd-1.6.1.tar.gz
   cd s3cmd-1.6.1/
   sudo python2.6 setup.py install

4. I couldn’t get the latest version of python-dateutils to work, so I fell back to dateutils-1.5 (https://pypi.python.org/pypi/python-dateutil/1.5):

   wget https://pypi.python.org/packages/b4/7c/df59c89a753eb33c7c44e1dd42de0e9bc2ccdd5a4d576e0bfad97cc280cb/python-dateutil-1.5.tar.gz
   tar zxf python-dateutil-1.5.tar.gz
   cd python-dateutil-1.5/
   sudo python2.6 setup.py install
   rm -rf /usr/local/lib/python2.6/dist-packages/python_dateutil-2.6.0-py2.6.egg

5. Copy s3cmd and associated python libraries to the ESX(i) host; ensure you have enabled SSH as instructed in #1. Remember to replace the IP address in all the below commands with your own:

   scp /usr/local/bin/s3cmd root@192.168.99.1:/bin
   scp /usr/lib/python2.6/htmlentitydefs.py root@192.168.99.1:/lib/python2.6
   scp /usr/lib/python2.6/hmac.py root@192.168.99.1:/lib/python2.6
   scp -r /usr/lib/python2.6/distutils/ root@192.168.99.1:/lib/python2.6
   scpr -r /usr/local/lib/python2.6/dist-packages/ root@192.168.99.1:/lib/python2.6

6. Statically compile the latest version of gpg – this ensures that we’re not relying on any shared libraries and can copy the binary over to the ESX(i) host:

   sudo apt-get install build-essential
   sudo apt-get install libgpg-error-dev libgcrypt-dev libassuan-dev libksba-dev libpth-dev zlib1g-dev
   wget https://www.gnupg.org/ftp/gcrypt/libgpg-error/libgpg-error-1.25.tar.bz2
   tar xjf libgpg-error-1.25.tar.bz2
   cd libgpg-error-1.25/
   ./configure && sudo make && sudo make install
   
   cd ..
   
   wget https://www.gnupg.org/ftp/gcrypt/gnupg/gnupg-2.0.30.tar.bz2
   tar -xjf gnupg-2.0.30.tar.bz2
   cd gnupg-2.0.30/
   ./configure LDFLAGS="-static"
   make LDFLAGS="-static" LINKFORSHARED=" "

   Copy the statically compiled binary to the ESX(i) host; ensure you have enabled SSH as instructed in #1. Remember to replace the IP address in all the below commands with your own:

   scp g10/gpg2 root@192.168.99.1:/bin

#3 Configure ESX(i) Host

By this point we should have all the necessary programs on the host to configure backups:

1. Add the following to the bottom of /etc/profile.local:

   export PYTHONPATH="/lib/python2.6/dist-packages/"

2. Logout and back in from your console session to refresh the environment variables. Then check the environment variable has been set:

   echo $PYTHONPATH

3. Verify that both gpg2 and s3cmd are working:

   s3cmd --version
   gpg2 --version

   If s3cmd returns a not found error, ensure the hash bang points to the correct python binary.

4. Configure s3cmd with your Amazon S3 account:

   s3cmd --configure

5. Download GhettoVCB backup script and upload it to somewhere that your ESX(i) host can access. Unfortunately at the time of publishing this article the script had a few bugs. You can see a diff of the fixes here: https://github.com/bytestream/ghettoVCB/commit/e7c5db852a12318c8b50366e968ee2038467115b

#4 Using s3cmd and GhettoVCB

Backup a given VM:

ghettoVCB.sh -m "VM_NAME"

Upload backup to Amazon S3:

s3cmd put /vmfs/volumes/datastore1/backups/vm_name.gz s3://mybucket.com-backups

#5 Script to backup multiple VMs via cron

1. Create a ghettoVCB.conf file, replacing /ABSOLUTE/PATH/backups with a directory where backups should be placed:

   VM_BACKUP_VOLUME=/ABSOLUTE/PATH/backups
   DISK_BACKUP_FORMAT=thin
   VM_BACKUP_ROTATION_COUNT=3
   POWER_VM_DOWN_BEFORE_BACKUP=0
   ENABLE_HARD_POWER_OFF=0
   ITER_TO_WAIT_SHUTDOWN=3
   POWER_DOWN_TIMEOUT=5
   ENABLE_COMPRESSION=1
   VM_SNAPSHOT_MEMORY=0
   VM_SNAPSHOT_QUIESCE=0
   ALLOW_VMS_WITH_SNAPSHOTS_TO_BE_BACKEDUP=0
   ENABLE_NON_PERSISTENT_NFS=0
   UNMOUNT_NFS=0
   
   SNAPSHOT_TIMEOUT=15
   EMAIL_LOG=0
   
   WORKDIR_DEBUG=0
   VM_SHUTDOWN_ORDER=
   VM_STARTUP_ORDER=
   

2. Create VMs_2_BACKUP file with the name of each VM to backup on separate lines
3. Create backup.sh, replacing BUCKET_NAME_HERE, /ABSOLUTE/PATH/TO with your relevant information:

   #!/bin/sh
   #
   # Auto Backup VMs.
   #
   
   # s3 protocol uri, including bucket and optional path.
   S3_URI="s3://BUCKET_NAME_HERE/vms"
   
   # Absolute path to GhettoVCB global config file.
   GLOBAL_CONF=/ABSOLUTE/PATH/TO/ghettoVCB.conf
   
   # File containing the names of VMs to backup.
   # Each VM name should be on a separate line.
   VMs=/ABSOLUTE/PATH/TO/VMs_2_BACKUP
   
   # Add GLOBAL_CONF variables to current context.
   source "${GLOBAL_CONF}"
   
   # Ensure default value if one is not selected or variable is null
   if [[ -z ${VM_BACKUP_DIR_NAMING_CONVENTION} ]] ; then
       VM_BACKUP_DIR_NAMING_CONVENTION="$(date +%F_%k-%M-%S)"
   fi
   
   # Back up each VM in the VMs file.
   while IFS= read -r VM_NAME <&3; do
   
       # Backup
       ghettoVCB.sh -g ${GLOBAL_CONF} -m "${VM_NAME}"
   
       # Lets assume that this is the compressed backup file...
       BACKUP_DIR="${VM_BACKUP_VOLUME}/${VM_NAME}"
       BACKUP_FILE=$(ls "$BACKUP_DIR" | head -1)
   
       # Send to S3
       s3cmd put "${BACKUP_DIR}/${BACKUP_FILE}" "${S3_URI}/${FILENAME}"
   
       # Delete backup file
       rm -rf "${BACKUP_DIR}"
   
   done 3< "${VMs}"
   

4. Run the script automatically by creating a cron job in /var/spool/cron/crontabs/root

References:

• How to install Amazon s3cmd utility on ESX(i) 4.0 & Backup VMs to the Cloud -https://communities.vmware.com/docs/DOC-11745
• ghettoVCB.sh - Free alternative for backing up VM's for ESX(i) - https://communities.vmware.com/docs/DOC-8760