pve-qemu/vma_spec.txt

133 lines
3.6 KiB
Plaintext

= Virtual Machine Archive format (VMA) =
This format contains a header which includes the VM configuration as
binary blobs, and a list of devices (dev_id, name).
The actual VM image data is stored inside extents. An extent contains
up to 64 clusters, and start with a 512 byte header containing
additional information for those clusters.
We use a cluster size of 65536, and use 8 bytes for each
cluster in the header to store the following information:
* 1 byte dev_id (to identity the drive)
* 1 byte not used (reserved)
* 2 bytes zero indicator (mark zero regions (16x4096))
* 4 bytes cluster number
We only store non-zero blocks (such block is 4096 bytes).
Each archive is marked with a uuid. The archive header and all
extent headers includes that uuid and a MD5 checksum (over header
data).
All numbers in VMA archive are stored in Big Endian byte order.
== VMA Header ==
Byte 0 - 3: magic
VMA magic string ("VMA\x00")
4 - 7: version
Version number (valid value is 1)
8 - 23: uuid
Unique ID, Same uuid is used to mark extents.
24 - 31: ctime
Backup time stamp (seconds since epoch)
32 - 47: md5sum
Header checksum (from byte 0 to header_size). This field
is filled with zero to generate the checksum.
48 - 51: blob_buffer_offset
Start of blob buffer (multiple of 512)
52 - 55: blob_buffer_size
Size of blob buffer (multiple of 512)
56 - 59: header_size
Overall size of this header (multiple of 512)
60 - 2043: reserved
2044 - 3067: uint32_t config_names[256]
Offsets into blob_buffer table
3068 - 4091: uint32_t config_data[256]
Offsets into blob_buffer table
4092 - 4095: reserved
4096 - 12287: VmaDeviceInfoHeader dev_info[256]
The offset in this table is used as 'dev_id' inside
the data streams.
12288 - header_size: Blob buffer
=== Devive Info Header (VmaDeviceInfoHeader) ===
This is use to store details about the contained disk images.
Byte 0 - 3: devive name (offsets into blob_buffer table)
4 - 7: reserved
8 - 15: device size in bytes
16 - 31: reserved
Note: Devive name 'vmstate' is reserved to store VM RAM state.
=== Blob buffer ===
The blob buffer is used to store both configuration file names and
configuration data.
This region contain a list of binary data blobs. Each blob starts with
a 2 byte size field, followed by the actual data.
== Image Data Streams ==
The VMA header is followed by the image data stream. Image data is grouped
with extents, which contains up to 59 clusters from different images.
=== VMA Extent Header ===
Byte 0 - 3: magic
VMA extent magic string ("VMAE")
4 - 5: reserved
6 - 7: block_count
Overall number of contained 4K block
8 - 23: uuid
Unique ID, Same uuid as used in the VMA header.
24 - 39: md5sum
Header checksum (from byte 0 to header_size). This field
is filled with zero to generate the checksum.
40 - 511: blockinfo[59]
Each 'blockinfo' (8 bytes) give further details about contained clusters:
Byte 0 - 1: mask
Bitmap used to indicate non-zero 4K blocks inside the
cluster.
2: reserved
3: dev_id
Device ID (offset into dev_info table)
4 - 7: cluster_num
The extend header if followed by the actual cluster data, where we only
store non-zero 4K blocks.