This project provides a Go package to read Virtual Hard Disk (VHD) file, a CLI interface to upload local VHD to Azure storage and to inspect a local VHD.
An implementation of VHD VHD specification can be found in the vhdcore package.
Note: You must have Go installed on your machine, at version 1.11 or greater. https://golang.org/dl/
go get github.com/flatcar/azure-vhd-utils
- Fast uploads - This tool offers faster uploads by using multiple routines and balancing the load across them.
- Efficient uploads - This tool will only upload used (non-zero) portions of the disk.
- Parallelism - This tool can upload segements of the VHD concurrently (user configurable).
USAGE:
azure-vhd-utils upload [command options] [arguments...]
OPTIONS:
--localvhdpath Path to source VHD in the local machine.
--stgaccountname Azure storage account name.
--stgaccountkey Azure storage account key.
--containername Name of the container holding destination page blob. (Default: vhds)
--blobname Name of the destination page blob.
--parallelism Number of concurrent goroutines to be used for uploadThe upload command uploads local VHD to Azure storage as page blob. Once uploaded, you can use Microsoft Azure portal to register an image based on this page blob and use it to create Azure Virtual Machines.
When creating a VHD for Microsoft Azure, the size of the VHD must be a whole number in megabytes, otherwise you will see an error similar to the following when you attempt to create image from the uploaded VHD in Azure:
"The VHD http://.blob.core.windows.net/vhds/.vhd has an unsupported virtual size of bytes. The size must be a whole number (in MBs)."
You should ensure the VHD size is even MB before uploading
VBoxManage modifyhd --resize <size in MB>
Resize-VHD -Path -SizeBytes
http://azure.microsoft.com/blog/2014/05/22/running-freebsd-in-azure/
qemu-img resize <path-to-raw-file> size
http://azure.microsoft.com/en-us/documentation/articles/virtual-machines-linux-create-upload-vhd-generic/
Azure requires VHD to be in Fixed Disk format. The command converts Dynamic and Differencing Disk to Fixed Disk during upload process, the conversion will not consume any additional space in local machine.
In case of Fixed Disk, the command detects blocks containing zeros and those will not be uploaded. In case of expandable disks (dynamic and differencing) only the blocks those are marked as non-empty in the Block Allocation Table (BAT) will be uploaded.
The blocks containing data will be uploaded as chunks of 2 MB pages. Consecutive blocks will be merged to create 2 MB pages if the block size of disk is less than 2 MB. If the block size is greater than 2 MB, tool will split them as 2 MB pages.
With page blob, we can upload multiple pages in parallel to decrease upload time. The command accepts the number of concurrent goroutines to use for upload through parallelism parameter. If the parallelism parameter is not proivded then it default to 8 * number_of_cpus.
A subset of command are exposed under inspect command for inspecting various segments of VHD in the local machine.
USAGE:
azure-vhd-utils inspect footer [command options] [arguments...]
OPTIONS:
--path Path to VHD.USAGE:
azure-vhd-utils inspect header [command options] [arguments...]
OPTIONS:
--path Path to VHD.Only expandable disks (dynamic and differencing) VHDs has header.
USAGE:
azure-vhd-utils inspect bat [command options] [arguments...]
OPTIONS:
--path Path to VHD.
--start-range Start range.
--end-range End range.
--skip-empty Do not show BAT entries pointing to empty blocks.Only expandable disks (dynamic and differencing) VHDs has BAT.
USAGE:
azure-vhd-utils inspect block info [command options] [arguments...]
OPTIONS:
--path Path to VHD.This command shows the total number blocks, block size and size of block sector
USAGE:
azure-vhd-utils inspect block bitmap [command options] [arguments...]
OPTIONS:
--path Path to VHD.
--block-index Index of the block.
This project is published under MIT License.