1QEMU-IMG(1) QEMU QEMU-IMG(1)
2
3
4
6 qemu-img - QEMU disk image utility
7
9 qemu-img [standard options] command [command options]
10
12 qemu-img allows you to create, convert and modify images offline. It
13 can handle all image formats supported by QEMU.
14
15 Warning: Never use qemu-img to modify images in use by a running vir‐
16 tual machine or any other process; this may destroy the image. Also, be
17 aware that querying an image that is being modified by another process
18 may encounter inconsistent state.
19
21 Standard options:
22
23 -h, --help
24 Display this help and exit
25
26 -V, --version
27 Display version information and exit
28
29 -T, --trace [[enable=]PATTERN][,events=FILE][,file=FILE]
30 Specify tracing options.
31
32 [enable=]PATTERN
33 Immediately enable events matching PATTERN (either event name
34 or a globbing pattern). This option is only available if
35 QEMU has been compiled with the simple, log or ftrace tracing
36 backend. To specify multiple events or patterns, specify the
37 -trace option multiple times.
38
39 Use -trace help to print a list of names of trace points.
40
41 events=FILE
42 Immediately enable events listed in FILE. The file must con‐
43 tain one event name (as listed in the trace-events-all file)
44 per line; globbing patterns are accepted too. This option is
45 only available if QEMU has been compiled with the simple, log
46 or ftrace tracing backend.
47
48 file=FILE
49 Log output traces to FILE. This option is only available if
50 QEMU has been compiled with the simple tracing backend.
51
52 The following commands are supported:
53
54 amend [--object OBJECTDEF] [--image-opts] [-p] [-q] [-f FMT] [-t CACHE]
55 [--force] -o OPTIONS FILENAME
56
57 bench [-c COUNT] [-d DEPTH] [-f FMT] [--flush-interval=FLUSH_INTERVAL]
58 [-i AIO] [-n] [--no-drain] [-o OFFSET] [--pattern=PATTERN] [-q] [-s
59 BUFFER_SIZE] [-S STEP_SIZE] [-t CACHE] [-w] [-U] FILENAME
60
61 bitmap (--merge SOURCE | --add | --remove | --clear | --enable | --dis‐
62 able)... [-b SOURCE_FILE [-F SOURCE_FMT]] [-g GRANULARITY] [--object
63 OBJECTDEF] [--image-opts | -f FMT] FILENAME BITMAP
64
65 check [--object OBJECTDEF] [--image-opts] [-q] [-f FMT] [--output=OFMT]
66 [-r [leaks | all]] [-T SRC_CACHE] [-U] FILENAME
67
68 commit [--object OBJECTDEF] [--image-opts] [-q] [-f FMT] [-t CACHE] [-b
69 BASE] [-r RATE_LIMIT] [-d] [-p] FILENAME
70
71 compare [--object OBJECTDEF] [--image-opts] [-f FMT] [-F FMT] [-T
72 SRC_CACHE] [-p] [-q] [-s] [-U] FILENAME1 FILENAME2
73
74 convert [--object OBJECTDEF] [--image-opts] [--target-image-opts]
75 [--target-is-zero] [--bitmaps] [-U] [-C] [-c] [-p] [-q] [-n] [-f FMT]
76 [-t CACHE] [-T SRC_CACHE] [-O OUTPUT_FMT] [-B BACKING_FILE [-F BACK‐
77 ING_FMT]] [-o OPTIONS] [-l SNAPSHOT_PARAM] [-S SPARSE_SIZE] [-r
78 RATE_LIMIT] [-m NUM_COROUTINES] [-W] [--salvage] FILENAME [FILENAME2
79 [...]] OUTPUT_FILENAME
80
81 create [--object OBJECTDEF] [-q] [-f FMT] [-b BACKING_FILE] [-F BACK‐
82 ING_FMT] [-u] [-o OPTIONS] FILENAME [SIZE]
83
84 dd [--image-opts] [-U] [-f FMT] [-O OUTPUT_FMT] [bs=BLOCK_SIZE]
85 [count=BLOCKS] [skip=BLOCKS] if=INPUT of=OUTPUT
86
87 info [--object OBJECTDEF] [--image-opts] [-f FMT] [--output=OFMT]
88 [--backing-chain] [-U] FILENAME
89
90 map [--object OBJECTDEF] [--image-opts] [-f FMT] [--start-offset=OFF‐
91 SET] [--max-length=LEN] [--output=OFMT] [-U] FILENAME
92
93 measure [--output=OFMT] [-O OUTPUT_FMT] [-o OPTIONS] [--size N | [--ob‐
94 ject OBJECTDEF] [--image-opts] [-f FMT] [-l SNAPSHOT_PARAM] FILENAME]
95
96 snapshot [--object OBJECTDEF] [--image-opts] [-U] [-q] [-l | -a SNAP‐
97 SHOT | -c SNAPSHOT | -d SNAPSHOT] FILENAME
98
99 rebase [--object OBJECTDEF] [--image-opts] [-U] [-q] [-f FMT] [-t
100 CACHE] [-T SRC_CACHE] [-p] [-u] -b BACKING_FILE [-F BACKING_FMT] FILE‐
101 NAME
102
103 resize [--object OBJECTDEF] [--image-opts] [-f FMT] [--prealloca‐
104 tion=PREALLOC] [-q] [--shrink] FILENAME [+ | -]SIZE
105
106 Command parameters:
107
108 FILENAME is a disk image filename.
109
110 FMT is the disk image format. It is guessed automatically in most
111 cases. See below for a description of the supported disk formats.
112
113 SIZE is the disk image size in bytes. Optional suffixes k or K (kilo‐
114 byte, 1024) M (megabyte, 1024k) and G (gigabyte, 1024M) and T (ter‐
115 abyte, 1024G) are supported. b is ignored.
116
117 OUTPUT_FILENAME is the destination disk image filename.
118
119 OUTPUT_FMT is the destination format.
120
121 OPTIONS is a comma separated list of format specific options in a
122 name=value format. Use -o ? for an overview of the options supported by
123 the used format or see the format descriptions below for details.
124
125 SNAPSHOT_PARAM is param used for internal snapshot, format is 'snap‐
126 shot.id=[ID],snapshot.name=[NAME]' or '[ID_OR_NAME]'.
127
128 --object OBJECTDEF
129 is a QEMU user creatable object definition. See the qemu(1) man‐
130 ual page for a description of the object properties. The most
131 common object type is a secret, which is used to supply pass‐
132 words and/or encryption keys.
133
134 --image-opts
135 Indicates that the source FILENAME parameter is to be inter‐
136 preted as a full option string, not a plain filename. This pa‐
137 rameter is mutually exclusive with the -f parameter.
138
139 --target-image-opts
140 Indicates that the OUTPUT_FILENAME parameter(s) are to be inter‐
141 preted as a full option string, not a plain filename. This pa‐
142 rameter is mutually exclusive with the -O parameters. It is cur‐
143 rently required to also use the -n parameter to skip image cre‐
144 ation. This restriction may be relaxed in a future release.
145
146 --force-share (-U)
147 If specified, qemu-img will open the image in shared mode, al‐
148 lowing other QEMU processes to open it in write mode. For exam‐
149 ple, this can be used to get the image information (with 'info'
150 subcommand) when the image is used by a running guest. Note
151 that this could produce inconsistent results because of concur‐
152 rent metadata changes, etc. This option is only allowed when
153 opening images in read-only mode.
154
155 --backing-chain
156 Will enumerate information about backing files in a disk image
157 chain. Refer below for further description.
158
159 -c Indicates that target image must be compressed (qcow format
160 only).
161
162 -h With or without a command, shows help and lists the supported
163 formats.
164
165 -p Display progress bar (compare, convert and rebase commands
166 only). If the -p option is not used for a command that supports
167 it, the progress is reported when the process receives a SIGUSR1
168 or SIGINFO signal.
169
170 -q Quiet mode - do not print any output (except errors). There's no
171 progress bar in case both -q and -p options are used.
172
173 -S SIZE
174 Indicates the consecutive number of bytes that must contain only
175 zeros for qemu-img to create a sparse image during conversion.
176 This value is rounded down to the nearest 512 bytes. You may use
177 the common size suffixes like k for kilobytes.
178
179 -t CACHE
180 Specifies the cache mode that should be used with the (destina‐
181 tion) file. See the documentation of the emulator's -drive
182 cache=... option for allowed values.
183
184 -T SRC_CACHE
185 Specifies the cache mode that should be used with the source
186 file(s). See the documentation of the emulator's -drive
187 cache=... option for allowed values.
188
189 Parameters to compare subcommand:
190
191 -f First image format
192
193 -F Second image format
194
195 -s Strict mode - fail on different image size or sector allocation
196
197 Parameters to convert subcommand:
198
199 --bitmaps
200 Additionally copy all persistent bitmaps from the top layer of
201 the source
202
203 -n Skip the creation of the target volume
204
205 -m Number of parallel coroutines for the convert process
206
207 -W Allow out-of-order writes to the destination. This option im‐
208 proves performance, but is only recommended for preallocated de‐
209 vices like host devices or other raw block devices.
210
211 -C Try to use copy offloading to move data from source image to
212 target. This may improve performance if the data is remote, such
213 as with NFS or iSCSI backends, but will not automatically spar‐
214 sify zero sectors, and may result in a fully allocated target
215 image depending on the host support for getting allocation in‐
216 formation.
217
218 -r Rate limit for the convert process
219
220 --salvage
221 Try to ignore I/O errors when reading. Unless in quiet mode
222 (-q), errors will still be printed. Areas that cannot be read
223 from the source will be treated as containing only zeroes.
224
225 --target-is-zero
226 Assume that reading the destination image will always return ze‐
227 ros. This parameter is mutually exclusive with a destination im‐
228 age that has a backing file. It is required to also use the -n
229 parameter to skip image creation.
230
231 Parameters to dd subcommand:
232
233 bs=BLOCK_SIZE
234 Defines the block size
235
236 count=BLOCKS
237 Sets the number of input blocks to copy
238
239 if=INPUT
240 Sets the input file
241
242 of=OUTPUT
243 Sets the output file
244
245 skip=BLOCKS
246 Sets the number of input blocks to skip
247
248 Parameters to snapshot subcommand:
249
250 snapshot
251 Is the name of the snapshot to create, apply or delete
252
253 -a Applies a snapshot (revert disk to saved state)
254
255 -c Creates a snapshot
256
257 -d Deletes a snapshot
258
259 -l Lists all snapshots in the given image
260
261 Command description:
262
263 amend [--object OBJECTDEF] [--image-opts] [-p] [-q] [-f FMT] [-t CACHE]
264 [--force] -o OPTIONS FILENAME
265 Amends the image format specific OPTIONS for the image file
266 FILENAME. Not all file formats support this operation.
267
268 The set of options that can be amended are dependent on the im‐
269 age format, but note that amending the backing chain relation‐
270 ship should instead be performed with qemu-img rebase.
271
272 --force allows some unsafe operations. Currently for -f luks, it
273 allows to erase the last encryption key, and to overwrite an ac‐
274 tive encryption key.
275
276 bench [-c COUNT] [-d DEPTH] [-f FMT] [--flush-interval=FLUSH_INTERVAL]
277 [-i AIO] [-n] [--no-drain] [-o OFFSET] [--pattern=PATTERN] [-q] [-s
278 BUFFER_SIZE] [-S STEP_SIZE] [-t CACHE] [-w] [-U] FILENAME
279 Run a simple sequential I/O benchmark on the specified image. If
280 -w is specified, a write test is performed, otherwise a read
281 test is performed.
282
283 A total number of COUNT I/O requests is performed, each BUF‐
284 FER_SIZE bytes in size, and with DEPTH requests in parallel. The
285 first request starts at the position given by OFFSET, each fol‐
286 lowing request increases the current position by STEP_SIZE. If
287 STEP_SIZE is not given, BUFFER_SIZE is used for its value.
288
289 If FLUSH_INTERVAL is specified for a write test, the request
290 queue is drained and a flush is issued before new writes are
291 made whenever the number of remaining requests is a multiple of
292 FLUSH_INTERVAL. If additionally --no-drain is specified, a flush
293 is issued without draining the request queue first.
294
295 if -i is specified, AIO option can be used to specify different
296 AIO backends: threads, native or io_uring.
297
298 If -n is specified, the native AIO backend is used if possible.
299 On Linux, this option only works if -t none or -t directsync is
300 specified as well.
301
302 For write tests, by default a buffer filled with zeros is writ‐
303 ten. This can be overridden with a pattern byte specified by
304 PATTERN.
305
306 bitmap (--merge SOURCE | --add | --remove | --clear | --enable | --dis‐
307 able)... [-b SOURCE_FILE [-F SOURCE_FMT]] [-g GRANULARITY] [--object
308 OBJECTDEF] [--image-opts | -f FMT] FILENAME BITMAP
309 Perform one or more modifications of the persistent bitmap BIT‐
310 MAP in the disk image FILENAME. The various modifications are:
311
312 --add to create BITMAP, enabled to record future edits.
313
314 --remove to remove BITMAP.
315
316 --clear to clear BITMAP.
317
318 --enable to change BITMAP to start recording future edits.
319
320 --disable to change BITMAP to stop recording future edits.
321
322 --merge to merge the contents of the SOURCE bitmap into BITMAP.
323
324 Additional options include -g which sets a non-default GRANULAR‐
325 ITY for --add, and -b and -F which select an alternative source
326 file for all SOURCE bitmaps used by --merge.
327
328 To see what bitmaps are present in an image, use qemu-img info.
329
330 check [--object OBJECTDEF] [--image-opts] [-q] [-f FMT] [--output=OFMT]
331 [-r [leaks | all]] [-T SRC_CACHE] [-U] FILENAME
332 Perform a consistency check on the disk image FILENAME. The com‐
333 mand can output in the format OFMT which is either human or
334 json. The JSON output is an object of QAPI type ImageCheck.
335
336 If -r is specified, qemu-img tries to repair any inconsistencies
337 found during the check. -r leaks repairs only cluster leaks,
338 whereas -r all fixes all kinds of errors, with a higher risk of
339 choosing the wrong fix or hiding corruption that has already oc‐
340 curred.
341
342 Only the formats qcow2, qed and vdi support consistency checks.
343
344 In case the image does not have any inconsistencies, check exits
345 with 0. Other exit codes indicate the kind of inconsistency
346 found or if another error occurred. The following table summa‐
347 rizes all exit codes of the check subcommand:
348
349 0 Check completed, the image is (now) consistent
350
351 1 Check not completed because of internal errors
352
353 2 Check completed, image is corrupted
354
355 3 Check completed, image has leaked clusters, but is not
356 corrupted
357
358 63 Checks are not supported by the image format
359
360 If -r is specified, exit codes representing the image state re‐
361 fer to the state after (the attempt at) repairing it. That is, a
362 successful -r all will yield the exit code 0, independently of
363 the image state before.
364
365 commit [--object OBJECTDEF] [--image-opts] [-q] [-f FMT] [-t CACHE] [-b
366 BASE] [-r RATE_LIMIT] [-d] [-p] FILENAME
367 Commit the changes recorded in FILENAME in its base image or
368 backing file. If the backing file is smaller than the snapshot,
369 then the backing file will be resized to be the same size as the
370 snapshot. If the snapshot is smaller than the backing file, the
371 backing file will not be truncated. If you want the backing
372 file to match the size of the smaller snapshot, you can safely
373 truncate it yourself once the commit operation successfully com‐
374 pletes.
375
376 The image FILENAME is emptied after the operation has succeeded.
377 If you do not need FILENAME afterwards and intend to drop it,
378 you may skip emptying FILENAME by specifying the -d flag.
379
380 If the backing chain of the given image file FILENAME has more
381 than one layer, the backing file into which the changes will be
382 committed may be specified as BASE (which has to be part of
383 FILENAME's backing chain). If BASE is not specified, the immedi‐
384 ate backing file of the top image (which is FILENAME) will be
385 used. Note that after a commit operation all images between BASE
386 and the top image will be invalid and may return garbage data
387 when read. For this reason, -b implies -d (so that the top image
388 stays valid).
389
390 The rate limit for the commit process is specified by -r.
391
392 compare [--object OBJECTDEF] [--image-opts] [-f FMT] [-F FMT] [-T
393 SRC_CACHE] [-p] [-q] [-s] [-U] FILENAME1 FILENAME2
394 Check if two images have the same content. You can compare im‐
395 ages with different format or settings.
396
397 The format is probed unless you specify it by -f (used for FILE‐
398 NAME1) and/or -F (used for FILENAME2) option.
399
400 By default, images with different size are considered identical
401 if the larger image contains only unallocated and/or zeroed sec‐
402 tors in the area after the end of the other image. In addition,
403 if any sector is not allocated in one image and contains only
404 zero bytes in the second one, it is evaluated as equal. You can
405 use Strict mode by specifying the -s option. When compare runs
406 in Strict mode, it fails in case image size differs or a sector
407 is allocated in one image and is not allocated in the second
408 one.
409
410 By default, compare prints out a result message. This message
411 displays information that both images are same or the position
412 of the first different byte. In addition, result message can re‐
413 port different image size in case Strict mode is used.
414
415 Compare exits with 0 in case the images are equal and with 1 in
416 case the images differ. Other exit codes mean an error occurred
417 during execution and standard error output should contain an er‐
418 ror message. The following table sumarizes all exit codes of
419 the compare subcommand:
420
421 0 Images are identical (or requested help was printed)
422
423 1 Images differ
424
425 2 Error on opening an image
426
427 3 Error on checking a sector allocation
428
429 4 Error on reading data
430
431 convert [--object OBJECTDEF] [--image-opts] [--target-image-opts]
432 [--target-is-zero] [--bitmaps [--skip-broken-bitmaps]] [-U] [-C] [-c]
433 [-p] [-q] [-n] [-f FMT] [-t CACHE] [-T SRC_CACHE] [-O OUTPUT_FMT] [-B
434 BACKING_FILE [-F BACKING_FMT]] [-o OPTIONS] [-l SNAPSHOT_PARAM] [-S
435 SPARSE_SIZE] [-r RATE_LIMIT] [-m NUM_COROUTINES] [-W] FILENAME [FILE‐
436 NAME2 [...]] OUTPUT_FILENAME
437 Convert the disk image FILENAME or a snapshot SNAPSHOT_PARAM to
438 disk image OUTPUT_FILENAME using format OUTPUT_FMT. It can be
439 optionally compressed (-c option) or use any format specific op‐
440 tions like encryption (-o option).
441
442 Only the formats qcow and qcow2 support compression. The com‐
443 pression is read-only. It means that if a compressed sector is
444 rewritten, then it is rewritten as uncompressed data.
445
446 Image conversion is also useful to get smaller image when using
447 a growable format such as qcow: the empty sectors are detected
448 and suppressed from the destination image.
449
450 SPARSE_SIZE indicates the consecutive number of bytes (defaults
451 to 4k) that must contain only zeros for qemu-img to create a
452 sparse image during conversion. If SPARSE_SIZE is 0, the source
453 will not be scanned for unallocated or zero sectors, and the
454 destination image will always be fully allocated.
455
456 You can use the BACKING_FILE option to force the output image to
457 be created as a copy on write image of the specified base image;
458 the BACKING_FILE should have the same content as the input's
459 base image, however the path, image format (as given by BACK‐
460 ING_FMT), etc may differ.
461
462 If a relative path name is given, the backing file is looked up
463 relative to the directory containing OUTPUT_FILENAME.
464
465 If the -n option is specified, the target volume creation will
466 be skipped. This is useful for formats such as rbd if the target
467 volume has already been created with site specific options that
468 cannot be supplied through qemu-img.
469
470 Out of order writes can be enabled with -W to improve perfor‐
471 mance. This is only recommended for preallocated devices like
472 host devices or other raw block devices. Out of order write does
473 not work in combination with creating compressed images.
474
475 NUM_COROUTINES specifies how many coroutines work in parallel
476 during the convert process (defaults to 8).
477
478 Use of --bitmaps requests that any persistent bitmaps present in
479 the original are also copied to the destination. If any bitmap
480 is inconsistent in the source, the conversion will fail unless
481 --skip-broken-bitmaps is also specified to copy only the consis‐
482 tent bitmaps.
483
484 create [--object OBJECTDEF] [-q] [-f FMT] [-b BACKING_FILE] [-F BACK‐
485 ING_FMT] [-u] [-o OPTIONS] FILENAME [SIZE]
486 Create the new disk image FILENAME of size SIZE and format FMT.
487 Depending on the file format, you can add one or more OPTIONS
488 that enable additional features of this format.
489
490 If the option BACKING_FILE is specified, then the image will
491 record only the differences from BACKING_FILE. No size needs to
492 be specified in this case. BACKING_FILE will never be modified
493 unless you use the commit monitor command (or qemu-img commit).
494
495 If a relative path name is given, the backing file is looked up
496 relative to the directory containing FILENAME.
497
498 Note that a given backing file will be opened to check that it
499 is valid. Use the -u option to enable unsafe backing file mode,
500 which means that the image will be created even if the associ‐
501 ated backing file cannot be opened. A matching backing file must
502 be created or additional options be used to make the backing
503 file specification valid when you want to use an image created
504 this way.
505
506 The size can also be specified using the SIZE option with -o, it
507 doesn't need to be specified separately in this case.
508
509 dd [--image-opts] [-U] [-f FMT] [-O OUTPUT_FMT] [bs=BLOCK_SIZE]
510 [count=BLOCKS] [skip=BLOCKS] if=INPUT of=OUTPUT
511 dd copies from INPUT file to OUTPUT file converting it from FMT
512 format to OUTPUT_FMT format.
513
514 The data is by default read and written using blocks of 512
515 bytes but can be modified by specifying BLOCK_SIZE. If
516 count=BLOCKS is specified dd will stop reading input after read‐
517 ing BLOCKS input blocks.
518
519 The size syntax is similar to dd(1)'s size syntax.
520
521 info [--object OBJECTDEF] [--image-opts] [-f FMT] [--output=OFMT]
522 [--backing-chain] [-U] FILENAME
523 Give information about the disk image FILENAME. Use it in par‐
524 ticular to know the size reserved on disk which can be different
525 from the displayed size. If VM snapshots are stored in the disk
526 image, they are displayed too.
527
528 If a disk image has a backing file chain, information about each
529 disk image in the chain can be recursively enumerated by using
530 the option --backing-chain.
531
532 For instance, if you have an image chain like:
533
534 base.qcow2 <- snap1.qcow2 <- snap2.qcow2
535
536 To enumerate information about each disk image in the above
537 chain, starting from top to base, do:
538
539 qemu-img info --backing-chain snap2.qcow2
540
541 The command can output in the format OFMT which is either human
542 or json. The JSON output is an object of QAPI type ImageInfo;
543 with --backing-chain, it is an array of ImageInfo objects.
544
545 --output=human reports the following information (for every im‐
546 age in the chain):
547
548 image The image file name
549
550 file format
551 The image format
552
553 virtual size
554 The size of the guest disk
555
556 disk size
557 How much space the image file occupies on the host file
558 system (may be shown as 0 if this information is unavail‐
559 able, e.g. because there is no file system)
560
561 cluster_size
562 Cluster size of the image format, if applicable
563
564 encrypted
565 Whether the image is encrypted (only present if so)
566
567 cleanly shut down
568 This is shown as no if the image is dirty and will have
569 to be auto-repaired the next time it is opened in qemu.
570
571 backing file
572 The backing file name, if present
573
574 backing file format
575 The format of the backing file, if the image enforces it
576
577 Snapshot list
578 A list of all internal snapshots
579
580 Format specific information
581 Further information whose structure depends on the image
582 format. This section is a textual representation of the
583 respective ImageInfoSpecific* QAPI object (e.g. ImageIn‐
584 foSpecificQCow2 for qcow2 images).
585
586 map [--object OBJECTDEF] [--image-opts] [-f FMT] [--start-offset=OFF‐
587 SET] [--max-length=LEN] [--output=OFMT] [-U] FILENAME
588 Dump the metadata of image FILENAME and its backing file chain.
589 In particular, this commands dumps the allocation state of every
590 sector of FILENAME, together with the topmost file that allo‐
591 cates it in the backing file chain.
592
593 Two option formats are possible. The default format (human)
594 only dumps known-nonzero areas of the file. Known-zero parts of
595 the file are omitted altogether, and likewise for parts that are
596 not allocated throughout the chain. qemu-img output will iden‐
597 tify a file from where the data can be read, and the offset in
598 the file. Each line will include four fields, the first three
599 of which are hexadecimal numbers. For example the first line
600 of:
601
602 Offset Length Mapped to File
603 0 0x20000 0x50000 /tmp/overlay.qcow2
604 0x100000 0x10000 0x95380000 /tmp/backing.qcow2
605
606 means that 0x20000 (131072) bytes starting at offset 0 in the
607 image are available in /tmp/overlay.qcow2 (opened in raw format)
608 starting at offset 0x50000 (327680). Data that is compressed,
609 encrypted, or otherwise not available in raw format will cause
610 an error if human format is in use. Note that file names can
611 include newlines, thus it is not safe to parse this output for‐
612 mat in scripts.
613
614 The alternative format json will return an array of dictionaries
615 in JSON format. It will include similar information in the
616 start, length, offset fields; it will also include other more
617 specific information:
618
619 • boolean field data: true if the sectors contain actual data,
620 false if the sectors are either unallocated or stored as opti‐
621 mized all-zero clusters
622
623 • boolean field zero: true if the data is known to read as zero
624
625 • boolean field present: true if the data belongs to the backing
626 chain, false if rebasing the backing chain onto a deeper file
627 would pick up data from the deeper file;
628
629 • integer field depth: the depth within the backing chain at
630 which the data was resolved; for example, a depth of 2 refers
631 to the backing file of the backing file of FILENAME.
632
633 In JSON format, the offset field is optional; it is absent in
634 cases where human format would omit the entry or exit with an
635 error. If data is false and the offset field is present, the
636 corresponding sectors in the file are not yet in use, but they
637 are preallocated.
638
639 For more information, consult include/block/block.h in QEMU's
640 source code.
641
642 measure [--output=OFMT] [-O OUTPUT_FMT] [-o OPTIONS] [--size N | [--ob‐
643 ject OBJECTDEF] [--image-opts] [-f FMT] [-l SNAPSHOT_PARAM] FILENAME]
644 Calculate the file size required for a new image. This informa‐
645 tion can be used to size logical volumes or SAN LUNs appropri‐
646 ately for the image that will be placed in them. The values re‐
647 ported are guaranteed to be large enough to fit the image. The
648 command can output in the format OFMT which is either human or
649 json. The JSON output is an object of QAPI type BlockMeasure‐
650 Info.
651
652 If the size N is given then act as if creating a new empty image
653 file using qemu-img create. If FILENAME is given then act as if
654 converting an existing image file using qemu-img convert. The
655 format of the new file is given by OUTPUT_FMT while the format
656 of an existing file is given by FMT.
657
658 A snapshot in an existing image can be specified using SNAP‐
659 SHOT_PARAM.
660
661 The following fields are reported:
662
663 required size: 524288
664 fully allocated size: 1074069504
665 bitmaps size: 0
666
667 The required size is the file size of the new image. It may be
668 smaller than the virtual disk size if the image format supports
669 compact representation.
670
671 The fully allocated size is the file size of the new image once
672 data has been written to all sectors. This is the maximum size
673 that the image file can occupy with the exception of internal
674 snapshots, dirty bitmaps, vmstate data, and other advanced image
675 format features.
676
677 The bitmaps size is the additional size required in order to
678 copy bitmaps from a source image in addition to the guest-visi‐
679 ble data; the line is omitted if either source or destination
680 lacks bitmap support, or 0 if bitmaps are supported but there is
681 nothing to copy.
682
683 snapshot [--object OBJECTDEF] [--image-opts] [-U] [-q] [-l | -a SNAP‐
684 SHOT | -c SNAPSHOT | -d SNAPSHOT] FILENAME
685 List, apply, create or delete snapshots in image FILENAME.
686
687 rebase [--object OBJECTDEF] [--image-opts] [-U] [-q] [-f FMT] [-t
688 CACHE] [-T SRC_CACHE] [-p] [-u] -b BACKING_FILE [-F BACKING_FMT] FILE‐
689 NAME
690 Changes the backing file of an image. Only the formats qcow2 and
691 qed support changing the backing file.
692
693 The backing file is changed to BACKING_FILE and (if the image
694 format of FILENAME supports this) the backing file format is
695 changed to BACKING_FMT. If BACKING_FILE is specified as "" (the
696 empty string), then the image is rebased onto no backing file
697 (i.e. it will exist independently of any backing file).
698
699 If a relative path name is given, the backing file is looked up
700 relative to the directory containing FILENAME.
701
702 CACHE specifies the cache mode to be used for FILENAME, whereas
703 SRC_CACHE specifies the cache mode for reading backing files.
704
705 There are two different modes in which rebase can operate:
706
707 Safe mode
708 This is the default mode and performs a real rebase oper‐
709 ation. The new backing file may differ from the old one
710 and qemu-img rebase will take care of keeping the
711 guest-visible content of FILENAME unchanged.
712
713 In order to achieve this, any clusters that differ be‐
714 tween BACKING_FILE and the old backing file of FILENAME
715 are merged into FILENAME before actually changing the
716 backing file.
717
718 Note that the safe mode is an expensive operation, compa‐
719 rable to converting an image. It only works if the old
720 backing file still exists.
721
722 Unsafe mode
723 qemu-img uses the unsafe mode if -u is specified. In this
724 mode, only the backing file name and format of FILENAME
725 is changed without any checks on the file contents. The
726 user must take care of specifying the correct new backing
727 file, or the guest-visible content of the image will be
728 corrupted.
729
730 This mode is useful for renaming or moving the backing
731 file to somewhere else. It can be used without an acces‐
732 sible old backing file, i.e. you can use it to fix an im‐
733 age whose backing file has already been moved/renamed.
734
735 You can use rebase to perform a "diff" operation on two disk im‐
736 ages. This can be useful when you have copied or cloned a
737 guest, and you want to get back to a thin image on top of a tem‐
738 plate or base image.
739
740 Say that base.img has been cloned as modified.img by copying it,
741 and that the modified.img guest has run so there are now some
742 changes compared to base.img. To construct a thin image called
743 diff.qcow2 that contains just the differences, do:
744
745 qemu-img create -f qcow2 -b modified.img diff.qcow2
746 qemu-img rebase -b base.img diff.qcow2
747
748 At this point, modified.img can be discarded, since base.img +
749 diff.qcow2 contains the same information.
750
751 resize [--object OBJECTDEF] [--image-opts] [-f FMT] [--prealloca‐
752 tion=PREALLOC] [-q] [--shrink] FILENAME [+ | -]SIZE
753 Change the disk image as if it had been created with SIZE.
754
755 Before using this command to shrink a disk image, you MUST use
756 file system and partitioning tools inside the VM to reduce allo‐
757 cated file systems and partition sizes accordingly. Failure to
758 do so will result in data loss!
759
760 When shrinking images, the --shrink option must be given. This
761 informs qemu-img that the user acknowledges all loss of data be‐
762 yond the truncated image's end.
763
764 After using this command to grow a disk image, you must use file
765 system and partitioning tools inside the VM to actually begin
766 using the new space on the device.
767
768 When growing an image, the --preallocation option may be used to
769 specify how the additional image area should be allocated on the
770 host. See the format description in the Notes section which
771 values are allowed. Using this option may result in slightly
772 more data being allocated than necessary.
773
775 Supported image file formats:
776
777 raw
778 Raw disk image format (default). This format has the advantage of
779 being simple and easily exportable to all other emulators. If your
780 file system supports holes (for example in ext2 or ext3 on Linux or
781 NTFS on Windows), then only the written sectors will reserve space.
782 Use qemu-img info to know the real size used by the image or ls -ls
783 on Unix/Linux.
784
785 Supported options:
786
787 preallocation
788 Preallocation mode (allowed values: off, falloc, full). fal‐
789 loc mode preallocates space for image by calling posix_fallo‐
790 cate(). full mode preallocates space for image by writing
791 data to underlying storage. This data may or may not be
792 zero, depending on the storage location.
793
794 qcow2
795 QEMU image format, the most versatile format. Use it to have smaller
796 images (useful if your filesystem does not supports holes, for exam‐
797 ple on Windows), optional AES encryption, zlib based compression and
798 support of multiple VM snapshots.
799
800 Supported options:
801
802 compat Determines the qcow2 version to use. compat=0.10 uses the
803 traditional image format that can be read by any QEMU since
804 0.10. compat=1.1 enables image format extensions that only
805 QEMU 1.1 and newer understand (this is the default). Amongst
806 others, this includes zero clusters, which allow efficient
807 copy-on-read for sparse images.
808
809 backing_file
810 File name of a base image (see create subcommand)
811
812 backing_fmt
813 Image format of the base image
814
815 encryption
816 If this option is set to on, the image is encrypted with
817 128-bit AES-CBC.
818
819 The use of encryption in qcow and qcow2 images is considered
820 to be flawed by modern cryptography standards, suffering from
821 a number of design problems:
822
823 • The AES-CBC cipher is used with predictable initialization
824 vectors based on the sector number. This makes it vulnera‐
825 ble to chosen plaintext attacks which can reveal the exis‐
826 tence of encrypted data.
827
828 • The user passphrase is directly used as the encryption key.
829 A poorly chosen or short passphrase will compromise the se‐
830 curity of the encryption.
831
832 • In the event of the passphrase being compromised there is
833 no way to change the passphrase to protect data in any qcow
834 images. The files must be cloned, using a different encryp‐
835 tion passphrase in the new file. The original file must
836 then be securely erased using a program like shred, though
837 even this is ineffective with many modern storage technolo‐
838 gies.
839
840 • Initialization vectors used to encrypt sectors are based on
841 the guest virtual sector number, instead of the host physi‐
842 cal sector. When a disk image has multiple internal snap‐
843 shots this means that data in multiple physical sectors is
844 encrypted with the same initialization vector. With the CBC
845 mode, this opens the possibility of watermarking attacks if
846 the attack can collect multiple sectors encrypted with the
847 same IV and some predictable data. Having multiple qcow2
848 images with the same passphrase also exposes this weakness
849 since the passphrase is directly used as the key.
850
851 Use of qcow / qcow2 encryption is thus strongly discouraged.
852 Users are recommended to use an alternative encryption tech‐
853 nology such as the Linux dm-crypt / LUKS system.
854
855 cluster_size
856 Changes the qcow2 cluster size (must be between 512 and 2M).
857 Smaller cluster sizes can improve the image file size whereas
858 larger cluster sizes generally provide better performance.
859
860 preallocation
861 Preallocation mode (allowed values: off, metadata, falloc,
862 full). An image with preallocated metadata is initially
863 larger but can improve performance when the image needs to
864 grow. falloc and full preallocations are like the same op‐
865 tions of raw format, but sets up metadata also.
866
867 lazy_refcounts
868 If this option is set to on, reference count updates are
869 postponed with the goal of avoiding metadata I/O and improv‐
870 ing performance. This is particularly interesting with
871 cache=writethrough which doesn't batch metadata updates. The
872 tradeoff is that after a host crash, the reference count ta‐
873 bles must be rebuilt, i.e. on the next open an (automatic)
874 qemu-img check -r all is required, which may take some time.
875
876 This option can only be enabled if compat=1.1 is specified.
877
878 nocow If this option is set to on, it will turn off COW of the
879 file. It's only valid on btrfs, no effect on other file sys‐
880 tems.
881
882 Btrfs has low performance when hosting a VM image file, even
883 more when the guest on the VM also using btrfs as file sys‐
884 tem. Turning off COW is a way to mitigate this bad perfor‐
885 mance. Generally there are two ways to turn off COW on btrfs:
886
887 • Disable it by mounting with nodatacow, then all newly cre‐
888 ated files will be NOCOW
889
890 • For an empty file, add the NOCOW file attribute. That's
891 what this option does.
892
893 Note: this option is only valid to new or empty files. If
894 there is an existing file which is COW and has data blocks
895 already, it couldn't be changed to NOCOW by setting nocow=on.
896 One can issue lsattr filename to check if the NOCOW flag is
897 set or not (Capital 'C' is NOCOW flag).
898
899 data_file
900 Filename where all guest data will be stored. If this option
901 is used, the qcow2 file will only contain the image's meta‐
902 data.
903
904 Note: Data loss will occur if the given filename already ex‐
905 ists when using this option with qemu-img create since
906 qemu-img will create the data file anew, overwriting the
907 file's original contents. To simply update the reference to
908 point to the given pre-existing file, use qemu-img amend.
909
910 data_file_raw
911 If this option is set to on, QEMU will always keep the exter‐
912 nal data file consistent as a standalone read-only raw image.
913
914 It does this by forwarding all write accesses to the qcow2
915 file through to the raw data file, including their offsets.
916 Therefore, data that is visible on the qcow2 node (i.e., to
917 the guest) at some offset is visible at the same offset in
918 the raw data file. This results in a read-only raw image.
919 Writes that bypass the qcow2 metadata may corrupt the qcow2
920 metadata because the out-of-band writes may result in the
921 metadata falling out of sync with the raw image.
922
923 If this option is off, QEMU will use the data file to store
924 data in an arbitrary manner. The file’s content will not make
925 sense without the accompanying qcow2 metadata. Where data is
926 written will have no relation to its offset as seen by the
927 guest, and some writes (specifically zero writes) may not be
928 forwarded to the data file at all, but will only be handled
929 by modifying qcow2 metadata.
930
931 This option can only be enabled if data_file is set.
932
933 Other
934 QEMU also supports various other image file formats for compatibil‐
935 ity with older QEMU versions or other hypervisors, including VMDK,
936 VDI, VHD (vpc), VHDX, qcow1 and QED. For a full list of supported
937 formats see qemu-img --help. For a more detailed description of
938 these formats, see the QEMU block drivers reference documentation.
939
940 The main purpose of the block drivers for these formats is image
941 conversion. For running VMs, it is recommended to convert the disk
942 images to either raw or qcow2 in order to achieve good performance.
943
945 Fabrice Bellard
946
948 2022, The QEMU Project Developers
949
950
951
952
9536.2.0 Jun 11, 2022 QEMU-IMG(1)