Articles in this section

See more

Exporting a Machine Image from Zadara Compute

Support Team
  • Updated
Follow

Overview

There are many 3rd-party tools which can be used to create an image file from a block device. While you can always follow the general procedure for this—create a snapshot, create a volume from that snapshot, and attach that to a VM to utilize your own tooling—Zadara also provides a tool for this in our zCompute Toolbox Image.

Creating a Machine Image with zCompute Toolbox

Downloading zCompute Toolbox

In order to begin, you'll need to download the zCompute Toolbox if you have not previously done so. When doing so you may be able to choose a scope of Project, Account, or Public (depending on your permissions).

Screenshot_20230914_093351.png

Creating zCompute Toolbox VM

Create a new VM Instance by deploying the  zCompute Toolbox Image. Ensure that the VM has a minimum of double the image size available as free storage space. In other words, for a 10GB image, ensure that there is a minimum of 20GB of free storage space in the directory where the script is executed.

Creating an Image

  1. Connect to the zCompute Toolbox VM:

    • Begin by accessing the toolbox VM using the key pair specified during the VM creation process.

  2. Transfer the Image Conversion Script:

    • Transfer the script provided below, export_image_on_v2z.sh, to the VM. Ensure that the script is placed in a partition with sufficient space to accommodate the image conversion process. Verify that the script is set to be executable.

  3. Configure Cluster and Login Parameters as Environment Variables:

    • Configure essential cluster and login parameters as environment variables. The parameters are as follows:
      • SYMP_URL—the URL of the cluster's API endpoint
      • SYMP_COMMAND—the specified symp CLI command, use symp -k to bypass SSL certificate validation
      • Login Parameters:
        • SYMP_DOMAIN—the Zadara Compute Account name
        • SYMP_PROJECT—the Project name
        • SYMP_USERNAME—the Username to be used for authentication
        • SYMP_PASSWORD—the Password to be used for authentication
      • Alternatively, instead of using login parameters, you may provide a pre-existing API token, in which case all other login-related parameters are disregarded:
        • SYMP_TOKEN—the API token to be used for authentication

  4. Executing the Image Conversion Command:

    • Execute the following command to initiate the image conversion process, replacing <image uuid> with the relevant image UUID: 
      • ./export_image_on_v2z.sh <image uuid>
  1.  

After successful execution, the converted image will be accessible as volume_<image_snapshot_name>.qcow2 within the local directory and can then be exported as needed.

Image Creation Script

export_image_on_v2z.sh:

#!/bin/bash
# sudo yum install epel-release
# sudo yum install jq
# SYMP_URL=https://1.2.3.4 SYMP_USERNAME=exampleuser SYMP_DOMAIN=exampleaccount SYMP_PROJECT=exampleproject SYMP_PASSWORD=*** symp -k
set -ex

if [ "$(whoami)" != "root" ]
then
  echo "Not running as root - re-executing with sudo"
  sudo "$0" $1
  exit
fi

IMAGE_ID=$1
export SYMP_URL=https://1.2.3.4
export SYMP_USERNAME=exampleuser
export SYMP_DOMAIN=exampleaccount
export SYMP_PROJECT=exampleproject
export SYMP_PASSWORD=***

LOG="log_for_last_run"
echo `date` > $LOG
TLOG="tee -a $LOG"

SYMP="symp -k"
TOKEN=$(symp -k login $SYMP_USERNAME $SYMP_PASSWORD $SYMP_DOMAIN $SYMP_PROJECT -f value)
SYMP="symp -k --project-token=$TOKEN"

echo "Extracting own VM ID" | $TLOG
VM_INFO=$(curl http://169.254.169.254/openstack/latest/meta_data.json)
MYVM_ID=$(echo ${VM_INFO} | jq -r .uuid)
echo "Own VM ID is $VM_UUID" | $TLOG

## { "block_device_mapping": [ {"snapshot_id": uuid, asd: asd}, ...] }
echo "Getting IMAGE INFO" | $TLOG
IMAGE_INFO=$($SYMP machine-images get ${IMAGE_ID} -f json)
IMAGE_SNAPSHOTS=$(echo ${IMAGE_INFO} | jq -r .block_device_mapping[].snapshot_id)
for SN in ${IMAGE_SNAPSHOTS}
do
  # Saving current disk device names
  LVOLUMES=($(lsblk --output NAME -d -n -e 11 | sed -z 's/\n/\|/g'))
  echo "Currently attached devices $LVOLUMES" | $TLOG
  LVOLUMES=${LVOLUMES}dummyeoff

  echo "Cloning image snapshot: $SN" | $TLOG
  echo "running command: volume create --source-id $SN \"volume_for_image_download\"" | $TLOG
  VOLUME_INFO=$($SYMP volume create --source-id $SN -f json "volume_for_image_download_${IMAGE_ID}")
  VOLUME_ID=$(echo $VOLUME_INFO | jq -r .id)
  SNAPSHOT_INFO=$($SYMP snapshot get $SN -f json)
  SNAPSHOT_NAME=$(echo ${SNAPSHOT_INFO} | jq -r .name)
  echo "Attaching new volumes and waiting for them be seen by the OS" | $TLOG
  echo "running command: vm volumes attach $MYVM_ID $VOLUME_ID" | $TLOG
  $SYMP vm volumes attach $MYVM_ID $VOLUME_ID | $TLOG

  sleep 15 # allow kernel to recognize new voles.
  $SYMP vm get -c id -c name -c bootVolume -c volumes -c networks -c vpc_id -c instanceType $MYVM_ID 2>>$LOG || $TLOG
  echo "vm devices:" | $TLOG
  lsblk --output NAME -d -n -e 11 | $TLOG
  LOCALVOLS=(`lsblk --output NAME -d -n -e 11 | egrep -v "$LVOLUMES"`)
  for ADDED_VOL in ${LOCALVOLS[*]}
  do
	  echo "Detected added volume: $ADDED_VOL" | $TLOG
    qemu-img convert -f raw -O qcow2 /dev/$ADDED_VOL volume_$SNAPSHOT_NAME.qcow2
  done
done

 

The Toolbox VM has the aws cli binaries available which for example would allow you to upload the converted image from the Toolbox VM externally to your own object store (once configured to your target).
eg.

aws s3 cp volume_<image_snapshot_name>.qcow2 s3://examplebucket/examplefolder

 

 

Related articles