2@c man begin SYNOPSIS
   3@command{qemu-img} [@var{standard} @var{options}] @var{command} [@var{command} @var{options}]
   4@c man end
   5@end example
   7@c man begin DESCRIPTION
   8qemu-img allows you to create, convert and modify images offline. It can handle
   9all image formats supported by QEMU.
  11@b{Warning:} Never use qemu-img to modify images in use by a running virtual
  12machine or any other process; this may destroy the image. Also, be aware that
  13querying an image that is being modified by another process may encounter
  14inconsistent state.
  15@c man end
  17@c man begin OPTIONS
  19Standard options:
  20@table @option
  21@item -h, --help
  22Display this help and exit
  23@item -V, --version
  24Display version information and exit
  25@item -T, --trace [[enable=]@var{pattern}][,events=@var{file}][,file=@var{file}]
  26@findex --trace
  27@include qemu-option-trace.texi
  28@end table
  30The following commands are supported:
  32@include qemu-img-cmds.texi
  34Command parameters:
  35@table @var
  37@item filename
  38is a disk image filename
  40@item fmt
  41is the disk image format. It is guessed automatically in most cases. See below
  42for a description of the supported disk formats.
  44@item size
  45is the disk image size in bytes. Optional suffixes @code{k} or @code{K}
  46(kilobyte, 1024) @code{M} (megabyte, 1024k) and @code{G} (gigabyte, 1024M)
  47and T (terabyte, 1024G) are supported.  @code{b} is ignored.
  49@item output_filename
  50is the destination disk image filename
  52@item output_fmt
  53is the destination format
  55@item options
  56is a comma separated list of format specific options in a
  57name=value format. Use @code{-o ?} for an overview of the options supported
  58by the used format or see the format descriptions below for details.
  60@item snapshot_param
  61is param used for internal snapshot, format is
  62'snapshot.id=[ID],snapshot.name=[NAME]' or '[ID_OR_NAME]'
  64@end table
  66@table @option
  68@item --object @var{objectdef}
  69is a QEMU user creatable object definition. See the @code{qemu(1)} manual
  70page for a description of the object properties. The most common object
  71type is a @code{secret}, which is used to supply passwords and/or encryption
  74@item --image-opts
  75Indicates that the source @var{filename} parameter is to be interpreted as a
  76full option string, not a plain filename. This parameter is mutually
  77exclusive with the @var{-f} parameter.
  79@item --target-image-opts
  80Indicates that the @var{output_filename} parameter(s) are to be interpreted as
  81a full option string, not a plain filename. This parameter is mutually
  82exclusive with the @var{-O} parameters. It is currently required to also use
  83the @var{-n} parameter to skip image creation. This restriction may be relaxed
  84in a future release.
  86@item --force-share (-U)
  87If specified, @code{qemu-img} will open the image in shared mode, allowing
  88other QEMU processes to open it in write mode. For example, this can be used to
  89get the image information (with 'info' subcommand) when the image is used by a
  90running guest.  Note that this could produce inconsistent results because of
  91concurrent metadata changes, etc. This option is only allowed when opening
  92images in read-only mode.
  94@item --backing-chain
  95will enumerate information about backing files in a disk image chain. Refer
  96below for further description.
  98@item -c
  99indicates that target image must be compressed (qcow format only)
 101@item -h
 102with or without a command shows help and lists the supported formats
 104@item -p
 105display progress bar (compare, convert and rebase commands only).
 106If the @var{-p} option is not used for a command that supports it, the
 107progress is reported when the process receives a @code{SIGUSR1} or
 108@code{SIGINFO} signal.
 110@item -q
 111Quiet mode - do not print any output (except errors). There's no progress bar
 112in case both @var{-q} and @var{-p} options are used.
 114@item -S @var{size}
 115indicates the consecutive number of bytes that must contain only zeros
 116for qemu-img to create a sparse image during conversion. This value is rounded
 117down to the nearest 512 bytes. You may use the common size suffixes like
 118@code{k} for kilobytes.
 120@item -t @var{cache}
 121specifies the cache mode that should be used with the (destination) file. See
 122the documentation of the emulator's @code{-drive cache=...} option for allowed
 125@item -T @var{src_cache}
 126specifies the cache mode that should be used with the source file(s). See
 127the documentation of the emulator's @code{-drive cache=...} option for allowed
 130@end table
 132Parameters to snapshot subcommand:
 134@table @option
 136@item snapshot
 137is the name of the snapshot to create, apply or delete
 138@item -a
 139applies a snapshot (revert disk to saved state)
 140@item -c
 141creates a snapshot
 142@item -d
 143deletes a snapshot
 144@item -l
 145lists all snapshots in the given image
 146@end table
 148Parameters to compare subcommand:
 150@table @option
 152@item -f
 153First image format
 154@item -F
 155Second image format
 156@item -s
 157Strict mode - fail on different image size or sector allocation
 158@end table
 160Parameters to convert subcommand:
 162@table @option
 164@item -n
 165Skip the creation of the target volume
 166@item -m
 167Number of parallel coroutines for the convert process
 168@item -W
 169Allow out-of-order writes to the destination. This option improves performance,
 170but is only recommended for preallocated devices like host devices or other
 171raw block devices.
 172@item -C
 173Try to use copy offloading to move data from source image to target. This may
 174improve performance if the data is remote, such as with NFS or iSCSI backends,
 175but will not automatically sparsify zero sectors, and may result in a fully
 176allocated target image depending on the host support for getting allocation
 178@item --salvage
 179Try to ignore I/O errors when reading.  Unless in quiet mode (@code{-q}), errors
 180will still be printed.  Areas that cannot be read from the source will be
 181treated as containing only zeroes.
 182@end table
 184Parameters to dd subcommand:
 186@table @option
 188@item bs=@var{block_size}
 189defines the block size
 190@item count=@var{blocks}
 191sets the number of input blocks to copy
 192@item if=@var{input}
 193sets the input file
 194@item of=@var{output}
 195sets the output file
 196@item skip=@var{blocks}
 197sets the number of input blocks to skip
 198@end table
 200Command description:
 202@table @option
 204@item amend [--object @var{objectdef}] [--image-opts] [-p] [-q] [-f @var{fmt}] [-t @var{cache}] -o @var{options} @var{filename}
 206Amends the image format specific @var{options} for the image file
 207@var{filename}. Not all file formats support this operation.
 209@item bench [-c @var{count}] [-d @var{depth}] [-f @var{fmt}] [--flush-interval=@var{flush_interval}] [-n] [--no-drain] [-o @var{offset}] [--pattern=@var{pattern}] [-q] [-s @var{buffer_size}] [-S @var{step_size}] [-t @var{cache}] [-w] [-U] @var{filename}
 211Run a simple sequential I/O benchmark on the specified image. If @code{-w} is
 212specified, a write test is performed, otherwise a read test is performed.
 214A total number of @var{count} I/O requests is performed, each @var{buffer_size}
 215bytes in size, and with @var{depth} requests in parallel. The first request
 216starts at the position given by @var{offset}, each following request increases
 217the current position by @var{step_size}. If @var{step_size} is not given,
 218@var{buffer_size} is used for its value.
 220If @var{flush_interval} is specified for a write test, the request queue is
 221drained and a flush is issued before new writes are made whenever the number of
 222remaining requests is a multiple of @var{flush_interval}. If additionally
 223@code{--no-drain} is specified, a flush is issued without draining the request
 224queue first.
 226If @code{-n} is specified, the native AIO backend is used if possible. On
 227Linux, this option only works if @code{-t none} or @code{-t directsync} is
 228specified as well.
 230For write tests, by default a buffer filled with zeros is written. This can be
 231overridden with a pattern byte specified by @var{pattern}.
 233@item check [--object @var{objectdef}] [--image-opts] [-q] [-f @var{fmt}] [--output=@var{ofmt}] [-r [leaks | all]] [-T @var{src_cache}] [-U] @var{filename}
 235Perform a consistency check on the disk image @var{filename}. The command can
 236output in the format @var{ofmt} which is either @code{human} or @code{json}.
 237The JSON output is an object of QAPI type @code{ImageCheck}.
 239If @code{-r} is specified, qemu-img tries to repair any inconsistencies found
 240during the check. @code{-r leaks} repairs only cluster leaks, whereas
 241@code{-r all} fixes all kinds of errors, with a higher risk of choosing the
 242wrong fix or hiding corruption that has already occurred.
 244Only the formats @code{qcow2}, @code{qed} and @code{vdi} support
 245consistency checks.
 247In case the image does not have any inconsistencies, check exits with @code{0}.
 248Other exit codes indicate the kind of inconsistency found or if another error
 249occurred. The following table summarizes all exit codes of the check subcommand:
 251@table @option
 253@item 0
 254Check completed, the image is (now) consistent
 255@item 1
 256Check not completed because of internal errors
 257@item 2
 258Check completed, image is corrupted
 259@item 3
 260Check completed, image has leaked clusters, but is not corrupted
 261@item 63
 262Checks are not supported by the image format
 264@end table
 266If @code{-r} is specified, exit codes representing the image state refer to the
 267state after (the attempt at) repairing it. That is, a successful @code{-r all}
 268will yield the exit code 0, independently of the image state before.
 270@item commit [--object @var{objectdef}] [--image-opts] [-q] [-f @var{fmt}] [-t @var{cache}] [-b @var{base}] [-d] [-p] @var{filename}
 272Commit the changes recorded in @var{filename} in its base image or backing file.
 273If the backing file is smaller than the snapshot, then the backing file will be
 274resized to be the same size as the snapshot.  If the snapshot is smaller than
 275the backing file, the backing file will not be truncated.  If you want the
 276backing file to match the size of the smaller snapshot, you can safely truncate
 277it yourself once the commit operation successfully completes.
 279The image @var{filename} is emptied after the operation has succeeded. If you do
 280not need @var{filename} afterwards and intend to drop it, you may skip emptying
 281@var{filename} by specifying the @code{-d} flag.
 283If the backing chain of the given image file @var{filename} has more than one
 284layer, the backing file into which the changes will be committed may be
 285specified as @var{base} (which has to be part of @var{filename}'s backing
 286chain). If @var{base} is not specified, the immediate backing file of the top
 287image (which is @var{filename}) will be used. Note that after a commit operation
 288all images between @var{base} and the top image will be invalid and may return
 289garbage data when read. For this reason, @code{-b} implies @code{-d} (so that
 290the top image stays valid).
 292@item compare [--object @var{objectdef}] [--image-opts] [-f @var{fmt}] [-F @var{fmt}] [-T @var{src_cache}] [-p] [-q] [-s] [-U] @var{filename1} @var{filename2}
 294Check if two images have the same content. You can compare images with
 295different format or settings.
 297The format is probed unless you specify it by @var{-f} (used for
 298@var{filename1}) and/or @var{-F} (used for @var{filename2}) option.
 300By default, images with different size are considered identical if the larger
 301image contains only unallocated and/or zeroed sectors in the area after the end
 302of the other image. In addition, if any sector is not allocated in one image
 303and contains only zero bytes in the second one, it is evaluated as equal. You
 304can use Strict mode by specifying the @var{-s} option. When compare runs in
 305Strict mode, it fails in case image size differs or a sector is allocated in
 306one image and is not allocated in the second one.
 308By default, compare prints out a result message. This message displays
 309information that both images are same or the position of the first different
 310byte. In addition, result message can report different image size in case
 311Strict mode is used.
 313Compare exits with @code{0} in case the images are equal and with @code{1}
 314in case the images differ. Other exit codes mean an error occurred during
 315execution and standard error output should contain an error message.
 316The following table sumarizes all exit codes of the compare subcommand:
 318@table @option
 320@item 0
 321Images are identical
 322@item 1
 323Images differ
 324@item 2
 325Error on opening an image
 326@item 3
 327Error on checking a sector allocation
 328@item 4
 329Error on reading data
 331@end table
 333@item convert [--object @var{objectdef}] [--image-opts] [--target-image-opts] [-U] [-C] [-c] [-p] [-q] [-n] [-f @var{fmt}] [-t @var{cache}] [-T @var{src_cache}] [-O @var{output_fmt}] [-B @var{backing_file}] [-o @var{options}] [-l @var{snapshot_param}] [-S @var{sparse_size}] [-m @var{num_coroutines}] [-W] @var{filename} [@var{filename2} [...]] @var{output_filename}
 335Convert the disk image @var{filename} or a snapshot @var{snapshot_param}
 336to disk image @var{output_filename} using format @var{output_fmt}. It can be optionally compressed (@code{-c}
 337option) or use any format specific options like encryption (@code{-o} option).
 339Only the formats @code{qcow} and @code{qcow2} support compression. The
 340compression is read-only. It means that if a compressed sector is
 341rewritten, then it is rewritten as uncompressed data.
 343Image conversion is also useful to get smaller image when using a
 344growable format such as @code{qcow}: the empty sectors are detected and
 345suppressed from the destination image.
 347@var{sparse_size} indicates the consecutive number of bytes (defaults to 4k)
 348that must contain only zeros for qemu-img to create a sparse image during
 349conversion. If @var{sparse_size} is 0, the source will not be scanned for
 350unallocated or zero sectors, and the destination image will always be
 351fully allocated.
 353You can use the @var{backing_file} option to force the output image to be
 354created as a copy on write image of the specified base image; the
 355@var{backing_file} should have the same content as the input's base image,
 356however the path, image format, etc may differ.
 358If a relative path name is given, the backing file is looked up relative to
 359the directory containing @var{output_filename}.
 361If the @code{-n} option is specified, the target volume creation will be
 362skipped. This is useful for formats such as @code{rbd} if the target
 363volume has already been created with site specific options that cannot
 364be supplied through qemu-img.
 366Out of order writes can be enabled with @code{-W} to improve performance.
 367This is only recommended for preallocated devices like host devices or other
 368raw block devices. Out of order write does not work in combination with
 369creating compressed images.
 371@var{num_coroutines} specifies how many coroutines work in parallel during
 372the convert process (defaults to 8).
 374@item create [--object @var{objectdef}] [-q] [-f @var{fmt}] [-b @var{backing_file}] [-F @var{backing_fmt}] [-u] [-o @var{options}] @var{filename} [@var{size}]
 376Create the new disk image @var{filename} of size @var{size} and format
 377@var{fmt}. Depending on the file format, you can add one or more @var{options}
 378that enable additional features of this format.
 380If the option @var{backing_file} is specified, then the image will record
 381only the differences from @var{backing_file}. No size needs to be specified in
 382this case. @var{backing_file} will never be modified unless you use the
 383@code{commit} monitor command (or qemu-img commit).
 385If a relative path name is given, the backing file is looked up relative to
 386the directory containing @var{filename}.
 388Note that a given backing file will be opened to check that it is valid. Use
 389the @code{-u} option to enable unsafe backing file mode, which means that the
 390image will be created even if the associated backing file cannot be opened. A
 391matching backing file must be created or additional options be used to make the
 392backing file specification valid when you want to use an image created this
 395The size can also be specified using the @var{size} option with @code{-o},
 396it doesn't need to be specified separately in this case.
 398@item dd [--image-opts] [-U] [-f @var{fmt}] [-O @var{output_fmt}] [bs=@var{block_size}] [count=@var{blocks}] [skip=@var{blocks}] if=@var{input} of=@var{output}
 400Dd copies from @var{input} file to @var{output} file converting it from
 401@var{fmt} format to @var{output_fmt} format.
 403The data is by default read and written using blocks of 512 bytes but can be
 404modified by specifying @var{block_size}. If count=@var{blocks} is specified
 405dd will stop reading input after reading @var{blocks} input blocks.
 407The size syntax is similar to dd(1)'s size syntax.
 409@item info [--object @var{objectdef}] [--image-opts] [-f @var{fmt}] [--output=@var{ofmt}] [--backing-chain] [-U] @var{filename}
 411Give information about the disk image @var{filename}. Use it in
 412particular to know the size reserved on disk which can be different
 413from the displayed size. If VM snapshots are stored in the disk image,
 414they are displayed too.
 416If a disk image has a backing file chain, information about each disk image in
 417the chain can be recursively enumerated by using the option @code{--backing-chain}.
 419For instance, if you have an image chain like:
 422base.qcow2 <- snap1.qcow2 <- snap2.qcow2
 423@end example
 425To enumerate information about each disk image in the above chain, starting from top to base, do:
 428qemu-img info --backing-chain snap2.qcow2
 429@end example
 431The command can output in the format @var{ofmt} which is either @code{human} or
 432@code{json}.  The JSON output is an object of QAPI type @code{ImageInfo}; with
 433@code{--backing-chain}, it is an array of @code{ImageInfo} objects.
 435@code{--output=human} reports the following information (for every image in the
 437@table @var
 438@item image
 439The image file name
 441@item file format
 442The image format
 444@item virtual size
 445The size of the guest disk
 447@item disk size
 448How much space the image file occupies on the host file system (may be shown as
 4490 if this information is unavailable, e.g. because there is no file system)
 451@item cluster_size
 452Cluster size of the image format, if applicable
 454@item encrypted
 455Whether the image is encrypted (only present if so)
 457@item cleanly shut down
 458This is shown as @code{no} if the image is dirty and will have to be
 459auto-repaired the next time it is opened in qemu.
 461@item backing file
 462The backing file name, if present
 464@item backing file format
 465The format of the backing file, if the image enforces it
 467@item Snapshot list
 468A list of all internal snapshots
 470@item Format specific information
 471Further information whose structure depends on the image format.  This section
 472is a textual representation of the respective @code{ImageInfoSpecific*} QAPI
 473object (e.g. @code{ImageInfoSpecificQCow2} for qcow2 images).
 474@end table
 476@item map [--object @var{objectdef}] [--image-opts] [-f @var{fmt}] [--output=@var{ofmt}] [-U] @var{filename}
 478Dump the metadata of image @var{filename} and its backing file chain.
 479In particular, this commands dumps the allocation state of every sector
 480of @var{filename}, together with the topmost file that allocates it in
 481the backing file chain.
 483Two option formats are possible.  The default format (@code{human})
 484only dumps known-nonzero areas of the file.  Known-zero parts of the
 485file are omitted altogether, and likewise for parts that are not allocated
 486throughout the chain.  @command{qemu-img} output will identify a file
 487from where the data can be read, and the offset in the file.  Each line
 488will include four fields, the first three of which are hexadecimal
 489numbers.  For example the first line of:
 491Offset          Length          Mapped to       File
 4920               0x20000         0x50000         /tmp/overlay.qcow2
 4930x100000        0x10000         0x95380000      /tmp/backing.qcow2
 494@end example
 496means that 0x20000 (131072) bytes starting at offset 0 in the image are
 497available in /tmp/overlay.qcow2 (opened in @code{raw} format) starting
 498at offset 0x50000 (327680).  Data that is compressed, encrypted, or
 499otherwise not available in raw format will cause an error if @code{human}
 500format is in use.  Note that file names can include newlines, thus it is
 501not safe to parse this output format in scripts.
 503The alternative format @code{json} will return an array of dictionaries
 504in JSON format.  It will include similar information in
 505the @code{start}, @code{length}, @code{offset} fields;
 506it will also include other more specific information:
 507@itemize @minus
 509whether the sectors contain actual data or not (boolean field @code{data};
 510if false, the sectors are either unallocated or stored as optimized
 511all-zero clusters);
 514whether the data is known to read as zero (boolean field @code{zero});
 517in order to make the output shorter, the target file is expressed as
 518a @code{depth}; for example, a depth of 2 refers to the backing file
 519of the backing file of @var{filename}.
 520@end itemize
 522In JSON format, the @code{offset} field is optional; it is absent in
 523cases where @code{human} format would omit the entry or exit with an error.
 524If @code{data} is false and the @code{offset} field is present, the
 525corresponding sectors in the file are not yet in use, but they are
 528For more information, consult @file{include/block/block.h} in QEMU's
 529source code.
 531@item measure [--output=@var{ofmt}] [-O @var{output_fmt}] [-o @var{options}] [--size @var{N} | [--object @var{objectdef}] [--image-opts] [-f @var{fmt}] [-l @var{snapshot_param}] @var{filename}]
 533Calculate the file size required for a new image.  This information can be used
 534to size logical volumes or SAN LUNs appropriately for the image that will be
 535placed in them.  The values reported are guaranteed to be large enough to fit
 536the image.  The command can output in the format @var{ofmt} which is either
 537@code{human} or @code{json}.  The JSON output is an object of QAPI type
 540If the size @var{N} is given then act as if creating a new empty image file
 541using @command{qemu-img create}.  If @var{filename} is given then act as if
 542converting an existing image file using @command{qemu-img convert}.  The format
 543of the new file is given by @var{output_fmt} while the format of an existing
 544file is given by @var{fmt}.
 546A snapshot in an existing image can be specified using @var{snapshot_param}.
 548The following fields are reported:
 550required size: 524288
 551fully allocated size: 1074069504
 552@end example
 554The @code{required size} is the file size of the new image.  It may be smaller
 555than the virtual disk size if the image format supports compact representation.
 557The @code{fully allocated size} is the file size of the new image once data has
 558been written to all sectors.  This is the maximum size that the image file can
 559occupy with the exception of internal snapshots, dirty bitmaps, vmstate data,
 560and other advanced image format features.
 562@item snapshot [--object @var{objectdef}] [--image-opts] [-U] [-q] [-l | -a @var{snapshot} | -c @var{snapshot} | -d @var{snapshot}] @var{filename}
 564List, apply, create or delete snapshots in image @var{filename}.
 566@item rebase [--object @var{objectdef}] [--image-opts] [-U] [-q] [-f @var{fmt}] [-t @var{cache}] [-T @var{src_cache}] [-p] [-u] -b @var{backing_file} [-F @var{backing_fmt}] @var{filename}
 568Changes the backing file of an image. Only the formats @code{qcow2} and
 569@code{qed} support changing the backing file.
 571The backing file is changed to @var{backing_file} and (if the image format of
 572@var{filename} supports this) the backing file format is changed to
 573@var{backing_fmt}. If @var{backing_file} is specified as ``'' (the empty
 574string), then the image is rebased onto no backing file (i.e. it will exist
 575independently of any backing file).
 577If a relative path name is given, the backing file is looked up relative to
 578the directory containing @var{filename}.
 580@var{cache} specifies the cache mode to be used for @var{filename}, whereas
 581@var{src_cache} specifies the cache mode for reading backing files.
 583There are two different modes in which @code{rebase} can operate:
 584@table @option
 585@item Safe mode
 586This is the default mode and performs a real rebase operation. The new backing
 587file may differ from the old one and qemu-img rebase will take care of keeping
 588the guest-visible content of @var{filename} unchanged.
 590In order to achieve this, any clusters that differ between @var{backing_file}
 591and the old backing file of @var{filename} are merged into @var{filename}
 592before actually changing the backing file.
 594Note that the safe mode is an expensive operation, comparable to converting
 595an image. It only works if the old backing file still exists.
 597@item Unsafe mode
 598qemu-img uses the unsafe mode if @code{-u} is specified. In this mode, only the
 599backing file name and format of @var{filename} is changed without any checks
 600on the file contents. The user must take care of specifying the correct new
 601backing file, or the guest-visible content of the image will be corrupted.
 603This mode is useful for renaming or moving the backing file to somewhere else.
 604It can be used without an accessible old backing file, i.e. you can use it to
 605fix an image whose backing file has already been moved/renamed.
 606@end table
 608You can use @code{rebase} to perform a ``diff'' operation on two
 609disk images.  This can be useful when you have copied or cloned
 610a guest, and you want to get back to a thin image on top of a
 611template or base image.
 613Say that @code{base.img} has been cloned as @code{modified.img} by
 614copying it, and that the @code{modified.img} guest has run so there
 615are now some changes compared to @code{base.img}.  To construct a thin
 616image called @code{diff.qcow2} that contains just the differences, do:
 619qemu-img create -f qcow2 -b modified.img diff.qcow2
 620qemu-img rebase -b base.img diff.qcow2
 621@end example
 623At this point, @code{modified.img} can be discarded, since
 624@code{base.img + diff.qcow2} contains the same information.
 626@item resize [--object @var{objectdef}] [--image-opts] [-f @var{fmt}] [--preallocation=@var{prealloc}] [-q] [--shrink] @var{filename} [+ | -]@var{size}
 628Change the disk image as if it had been created with @var{size}.
 630Before using this command to shrink a disk image, you MUST use file system and
 631partitioning tools inside the VM to reduce allocated file systems and partition
 632sizes accordingly.  Failure to do so will result in data loss!
 634When shrinking images, the @code{--shrink} option must be given. This informs
 635qemu-img that the user acknowledges all loss of data beyond the truncated
 636image's end.
 638After using this command to grow a disk image, you must use file system and
 639partitioning tools inside the VM to actually begin using the new space on the
 642When growing an image, the @code{--preallocation} option may be used to specify
 643how the additional image area should be allocated on the host.  See the format
 644description in the @code{NOTES} section which values are allowed.  Using this
 645option may result in slightly more data being allocated than necessary.
 647@end table
 648@c man end
 651@c man begin NOTES
 652Supported image file formats:
 654@table @option
 655@item raw
 657Raw disk image format (default). This format has the advantage of
 658being simple and easily exportable to all other emulators. If your
 659file system supports @emph{holes} (for example in ext2 or ext3 on
 660Linux or NTFS on Windows), then only the written sectors will reserve
 661space. Use @code{qemu-img info} to know the real size used by the
 662image or @code{ls -ls} on Unix/Linux.
 664Supported options:
 665@table @code
 666@item preallocation
 667Preallocation mode (allowed values: @code{off}, @code{falloc}, @code{full}).
 668@code{falloc} mode preallocates space for image by calling posix_fallocate().
 669@code{full} mode preallocates space for image by writing zeros to underlying
 671@end table
 673@item qcow2
 674QEMU image format, the most versatile format. Use it to have smaller
 675images (useful if your filesystem does not supports holes, for example
 676on Windows), optional AES encryption, zlib based compression and
 677support of multiple VM snapshots.
 679Supported options:
 680@table @code
 681@item compat
 682Determines the qcow2 version to use. @code{compat=0.10} uses the
 683traditional image format that can be read by any QEMU since 0.10.
 684@code{compat=1.1} enables image format extensions that only QEMU 1.1 and
 685newer understand (this is the default). Amongst others, this includes zero
 686clusters, which allow efficient copy-on-read for sparse images.
 688@item backing_file
 689File name of a base image (see @option{create} subcommand)
 690@item backing_fmt
 691Image format of the base image
 692@item encryption
 693If this option is set to @code{on}, the image is encrypted with 128-bit AES-CBC.
 695The use of encryption in qcow and qcow2 images is considered to be flawed by
 696modern cryptography standards, suffering from a number of design problems:
 698@itemize @minus
 700The AES-CBC cipher is used with predictable initialization vectors based
 701on the sector number. This makes it vulnerable to chosen plaintext attacks
 702which can reveal the existence of encrypted data.
 704The user passphrase is directly used as the encryption key. A poorly
 705chosen or short passphrase will compromise the security of the encryption.
 707In the event of the passphrase being compromised there is no way to
 708change the passphrase to protect data in any qcow images. The files must
 709be cloned, using a different encryption passphrase in the new file. The
 710original file must then be securely erased using a program like shred,
 711though even this is ineffective with many modern storage technologies.
 713Initialization vectors used to encrypt sectors are based on the
 714guest virtual sector number, instead of the host physical sector. When
 715a disk image has multiple internal snapshots this means that data in
 716multiple physical sectors is encrypted with the same initialization
 717vector. With the CBC mode, this opens the possibility of watermarking
 718attacks if the attack can collect multiple sectors encrypted with the
 719same IV and some predictable data. Having multiple qcow2 images with
 720the same passphrase also exposes this weakness since the passphrase
 721is directly used as the key.
 722@end itemize
 724Use of qcow / qcow2 encryption is thus strongly discouraged. Users are
 725recommended to use an alternative encryption technology such as the
 726Linux dm-crypt / LUKS system.
 728@item cluster_size
 729Changes the qcow2 cluster size (must be between 512 and 2M). Smaller cluster
 730sizes can improve the image file size whereas larger cluster sizes generally
 731provide better performance.
 733@item preallocation
 734Preallocation mode (allowed values: @code{off}, @code{metadata}, @code{falloc},
 735@code{full}). An image with preallocated metadata is initially larger but can
 736improve performance when the image needs to grow. @code{falloc} and @code{full}
 737preallocations are like the same options of @code{raw} format, but sets up
 738metadata also.
 740@item lazy_refcounts
 741If this option is set to @code{on}, reference count updates are postponed with
 742the goal of avoiding metadata I/O and improving performance. This is
 743particularly interesting with @option{cache=writethrough} which doesn't batch
 744metadata updates. The tradeoff is that after a host crash, the reference count
 745tables must be rebuilt, i.e. on the next open an (automatic) @code{qemu-img
 746check -r all} is required, which may take some time.
 748This option can only be enabled if @code{compat=1.1} is specified.
 750@item nocow
 751If this option is set to @code{on}, it will turn off COW of the file. It's only
 752valid on btrfs, no effect on other file systems.
 754Btrfs has low performance when hosting a VM image file, even more when the guest
 755on the VM also using btrfs as file system. Turning off COW is a way to mitigate
 756this bad performance. Generally there are two ways to turn off COW on btrfs:
 757a) Disable it by mounting with nodatacow, then all newly created files will be
 758NOCOW. b) For an empty file, add the NOCOW file attribute. That's what this option
 761Note: this option is only valid to new or empty files. If there is an existing
 762file which is COW and has data blocks already, it couldn't be changed to NOCOW
 763by setting @code{nocow=on}. One can issue @code{lsattr filename} to check if
 764the NOCOW flag is set or not (Capital 'C' is NOCOW flag).
 766@end table
 768@item Other
 769QEMU also supports various other image file formats for compatibility with
 770older QEMU versions or other hypervisors, including VMDK, VDI, VHD (vpc), VHDX,
 771qcow1 and QED. For a full list of supported formats see @code{qemu-img --help}.
 772For a more detailed description of these formats, see the QEMU Emulation User
 775The main purpose of the block drivers for these formats is image conversion.
 776For running VMs, it is recommended to convert the disk images to either raw or
 777qcow2 in order to achieve good performance.
 778@end table
 781@c man end
 783@setfilename qemu-img
 784@settitle QEMU disk image utility
 786@c man begin SEEALSO
 787The HTML documentation of QEMU for more precise information and Linux
 788user mode emulator invocation.
 789@c man end
 791@c man begin AUTHOR
 792Fabrice Bellard
 793@c man end
 795@end ignore