1guestfs-faq(1) Virtualization Support guestfs-faq(1)
2
6 guestfs-faq - libguestfs Frequently Asked Questions (FAQ)
7
9 What is libguestfs?
10 libguestfs is a way to create, access and modify disk images. You can
11 look inside disk images, modify the files they contain, create them
12 from scratch, resize them, and much more. It’s especially useful from
13 scripts and programs and from the command line.
14 libguestfs is a C library (hence "lib-"), and a set of tools built on
16 this library, and bindings for many common programming languages.
17 For more information about what libguestfs can do read the introduction
19 on the home page (http://libguestfs.org).
20 What are the virt tools?
22 Virt tools (website: http://virt-tools.org) are a whole set of
23 virtualization management tools aimed at system administrators. Some
24 of them come from libguestfs, some from libvirt and many others from
25 other open source projects. So virt tools is a superset of libguestfs.
26 However libguestfs comes with many important tools. See
27 http://libguestfs.org for a full list.
28 Does libguestfs need { libvirt / KVM / Red Hat / Fedora }?
30 No!
31 libvirt is not a requirement for libguestfs.
33 libguestfs works with any disk image, including ones created in VMware,
35 KVM, qemu, VirtualBox, Xen, and many other hypervisors, and ones which
36 you have created from scratch.
37 Red Hat sponsors (ie. pays for) development of libguestfs and a huge
39 number of other open source projects. But you can run libguestfs and
40 the virt tools on many different Linux distros and Mac OS X. We try
41 our best to support all Linux distros as first-class citizens. Some
42 virt tools have been ported to Windows.
43 How does libguestfs compare to other tools?
45 vs. kpartx
46 Libguestfs takes a different approach from kpartx. kpartx needs
47 root, and mounts filesystems on the host kernel (which can be
48 insecure - see guestfs-security(1)). Libguestfs isolates your host
49 kernel from guests, is more flexible, scriptable, supports LVM,
50 doesn't require root, is isolated from other processes, and cleans
51 up after itself. Libguestfs is more than just file access because
52 you can use it to create images from scratch.
53 vs. vdfuse
55 vdfuse is like kpartx but for VirtualBox images. See the kpartx
56 comparison above. You can use libguestfs on the partition files
57 exposed by vdfuse, although it’s not necessary since libguestfs can
58 access VirtualBox images directly.
59 vs. qemu-nbd
61 NBD (Network Block Device) is a protocol for exporting block
62 devices over the network. qemu-nbd is an NBD server which can
63 handle any disk format supported by qemu (eg. raw, qcow2). You can
64 use libguestfs and qemu-nbd or nbdkit together to access block
65 devices over the network, for example: "guestfish -a nbd://remote"
66 vs. mounting filesystems in the host
68 Mounting guest filesystems in the host is insecure and should be
69 avoided completely for untrusted guests. Use libguestfs to provide
70 a layer of protection against filesystem exploits. See also
71 guestmount(1).
72 vs. parted
74 Libguestfs supports LVM. Libguestfs uses parted and provides most
75 parted features through the libguestfs API.
76
78 How do I know what version I'm using?
79 The simplest method is:
80 guestfish --version
82 Libguestfs development happens along an unstable branch and we
84 periodically create a stable branch which we backport stable patches
85 to. To find out more, read "LIBGUESTFS VERSION NUMBERS" in guestfs(3).
86 How can I get help?
88 What mailing lists or chat rooms are available?
89 If you are a Red Hat customer using Red Hat Enterprise Linux, please
90 contact Red Hat Support: http://redhat.com/support
91 There is a mailing list, mainly for development, but users are also
93 welcome to ask questions about libguestfs and the virt tools:
94 https://www.redhat.com/mailman/listinfo/libguestfs
95 You can also talk to us on IRC channel "#guestfs" on Libera Chat.
97 We're not always around, so please stay in the channel after asking
98 your question and someone will get back to you.
99 For other virt tools (not ones supplied with libguestfs) there is a
101 general virt tools mailing list:
102 https://www.redhat.com/mailman/listinfo/virt-tools-list
103 How do I report bugs?
105 Please use the following link to enter a bug in Bugzilla:
106 https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virtualization+Tools
108 Include as much detail as you can and a way to reproduce the problem.
110 Include the full output of libguestfs-test-tool(1).
112
114 See also "LIBGUESTFS GOTCHAS" in guestfs(3) for some "gotchas" with
115 using the libguestfs API.
116 "Could not allocate dynamic translator buffer"
118 This obscure error is in fact an SELinux failure. You have to enable
119 the following SELinux boolean:
120 setsebool -P virt_use_execmem=on
122 For more information see
124 https://bugzilla.redhat.com/show_bug.cgi?id=806106.
125 "child process died unexpectedly"
127 [This error message was changed in libguestfs 1.21.18 to something more
128 explanatory.]
129 This error indicates that qemu failed or the host kernel could not
131 boot. To get further information about the failure, you have to run:
132 libguestfs-test-tool
134 If, after using this, you still don’t understand the failure, contact
136 us (see previous section).
137 libguestfs: error: cannot find any suitable libguestfs supermin, fixed or
139 old-style appliance on LIBGUESTFS_PATH
140 febootstrap-supermin-helper: ext2: parent directory not found
141 supermin-helper: ext2: parent directory not found
142 [This issue is fixed permanently in libguestfs ≥ 1.26.]
143 If you see any of these errors on Debian/Ubuntu, you need to run the
145 following command:
146 sudo update-guestfs-appliance
148 "Permission denied" when running libguestfs as root
150 You get a permission denied error when opening a disk image, even
151 though you are running libguestfs as root.
152 This is caused by libvirt, and so only happens when using the libvirt
154 backend. When run as root, libvirt decides to run the qemu appliance
155 as user "qemu.qemu". Unfortunately this usually means that qemu cannot
156 open disk images, especially if those disk images are owned by root, or
157 are present in directories which require root access.
158 There is a bug open against libvirt to fix this:
160 https://bugzilla.redhat.com/show_bug.cgi?id=1045069
161 You can work around this by one of the following methods:
163 • Switch to the direct backend:
165 export LIBGUESTFS_BACKEND=direct
167 • Don’t run libguestfs as root.
169 • Chmod the disk image and any parent directories so that the qemu
171 user can access them.
172 • (Nasty) Edit /etc/libvirt/qemu.conf and change the "user" setting.
174 execl: /init: Permission denied
176 Note: If this error happens when you are using a distro package of
177 libguestfs (eg. from Fedora, Debian, etc) then file a bug against the
178 distro. This is not an error which normal users should ever see if the
179 distro package has been prepared correctly.
180 This error happens during the supermin boot phase of starting the
182 appliance:
183 supermin: mounting new root on /root
185 supermin: chroot
186 execl: /init: Permission denied
187 supermin: debug: listing directory /
188 [...followed by a lot of debug output...]
189 This is a complicated bug related to supermin(1) appliances. The
191 appliance is constructed by copying files like /bin/bash and many
192 libraries from the host. The file "hostfiles" lists the files that
193 should be copied from the host into the appliance. If some files don't
194 exist on the host then they are missed out, but if these files are
195 needed in order to (eg) run /bin/bash then you'll see the above error.
196 Diagnosing the problem involves studying the libraries needed by
198 /bin/bash, ie:
199 ldd /bin/bash
201 comparing that with "hostfiles", with the files actually available in
203 the host filesystem, and with the debug output printed in the error
204 message. Once you've worked out which file is missing, install that
205 file using your package manager and try again.
206 You should also check that files like /init and /bin/bash (in the
208 appliance) are executable. The debug output shows file modes.
209
211 Where can I get the latest binaries for ...?
212 Fedora ≥ 11
213 Use:
214 yum install '*guestf*'
216 For the latest builds, see:
218 http://koji.fedoraproject.org/koji/packageinfo?packageID=8391
219 Red Hat Enterprise Linux
221 RHEL 6
222 RHEL 7
223 It is part of the default install. On RHEL 6 and 7 (only) you
224 have to install "libguestfs-winsupport" to get Windows guest
225 support.
226 Debian and Ubuntu
228 For libguestfs < 1.26, after installing libguestfs you need to do:
229 sudo update-guestfs-appliance
231 (This script has been removed on Debian/Ubuntu with libguestfs ≥
233 1.26 and instead the appliance is built on demand.)
234 On Ubuntu only:
236 sudo chmod 0644 /boot/vmlinuz*
238 You may need to add yourself to the "kvm" group:
240 sudo usermod -a -G kvm yourlogin
242 Debian Squeeze (6)
244 Hilko Bengen has built libguestfs in squeeze backports:
245 http://packages.debian.org/search?keywords=guestfs&searchon=names§ion=all&suite=squeeze-backports
246 Debian Wheezy and later (7+)
248 Hilko Bengen supports libguestfs on Debian. Official Debian
249 packages are available:
250 http://packages.debian.org/search?keywords=libguestfs
251 Ubuntu
253 We don’t have a full time Ubuntu maintainer, and the packages
254 supplied by Canonical (which are outside our control) are
255 sometimes broken.
256 Canonical decided to change the permissions on the kernel so
258 that it's not readable except by root. This is completely
259 stupid, but they won't change it
260 (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/759725).
261 So every user should do this:
262 sudo chmod 0644 /boot/vmlinuz*
264 Ubuntu 12.04
266 libguestfs in this version of Ubuntu works, but you need to
267 update febootstrap and seabios to the latest versions.
268 You need febootstrap ≥ 3.14-2 from:
270 http://packages.ubuntu.com/precise/febootstrap
271 After installing or updating febootstrap, rebuild the
273 appliance:
274 sudo update-guestfs-appliance
276 You need seabios ≥ 0.6.2-0ubuntu2.1 or ≥ 0.6.2-0ubuntu3
278 from: http://packages.ubuntu.com/precise-updates/seabios or
279 http://packages.ubuntu.com/quantal/seabios
280 Also you need to do (see above):
282 sudo chmod 0644 /boot/vmlinuz*
284 Gentoo
286 Libguestfs was added to Gentoo in 2012-07 by Andreis Vinogradovs
287 (libguestfs) and Maxim Koltsov (mainly hivex). Do:
288 emerge libguestfs
290 Mageia
292 Libguestfs was added to Mageia in 2013-08. Do:
293 urpmi libguestfs
295 SuSE
297 Libguestfs was added to SuSE in 2012 by Olaf Hering.
298 ArchLinux
300 Libguestfs was added to the AUR in 2010.
301 Other Linux distro
303 Compile from source (next section).
304 Other non-Linux distro
306 You'll have to compile from source, and port it.
307 How can I compile and install libguestfs from source?
309 You can compile libguestfs from git or a source tarball. Read the
310 README file before starting.
311 Git: https://github.com/libguestfs/libguestfs Source tarballs:
313 http://libguestfs.org/download
314 Don’t run "make install"! Use the "./run" script instead (see README).
316 How can I compile and install libguestfs if my distro doesn't have new
318 enough qemu/supermin/kernel?
319 Libguestfs needs supermin 5. If supermin 5 hasn't been ported to your
320 distro, then see the question below.
321 First compile qemu, supermin and/or the kernel from source. You do not
323 need to "make install" them.
324 In the libguestfs source directory, create two files. "localconfigure"
326 should contain:
327 source localenv
329 #export PATH=/tmp/qemu/x86_64-softmmu:$PATH
330 ./configure --prefix /usr "$@"
331 Make "localconfigure" executable.
333 "localenv" should contain:
335 #export SUPERMIN=/tmp/supermin/src/supermin
337 #export LIBGUESTFS_HV=/tmp/qemu/x86_64-softmmu/qemu-system-x86_64
338 #export SUPERMIN_KERNEL=/tmp/linux/arch/x86/boot/bzImage
339 #export SUPERMIN_KERNEL_VERSION=4.XX.0
340 #export SUPERMIN_MODULES=/tmp/lib/modules/4.XX.0
341 Uncomment and adjust these lines as required to use the alternate
343 programs you have compiled.
344 Use "./localconfigure" instead of "./configure", but otherwise you
346 compile libguestfs as usual.
347 Don’t run "make install"! Use the "./run" script instead (see README).
349 How can I compile and install libguestfs without supermin?
351 If supermin 5 supports your distro, but you don’t happen to have a new
352 enough supermin installed, then see the previous question.
353 If supermin 5 doesn't support your distro at all, you will need to use
355 the "fixed appliance method" where you use a pre-compiled binary
356 appliance. To build libguestfs without supermin, you need to pass
357 "--disable-appliance --disable-daemon" to either ./configure or
358 ./configure (depending whether you are building respectively from git
359 or from tarballs). Then, when using libguestfs, you must set the
360 "LIBGUESTFS_PATH" environment variable to the directory of a pre-
361 compiled appliance, as also described in "FIXED APPLIANCE" in
362 guestfs-internals(1).
363 For pre-compiled appliances, see also:
365 http://libguestfs.org/download/binaries/appliance/.
366 Patches to port supermin to more Linux distros are welcome.
368 How can I add support for sVirt?
370 Note for Fedora/RHEL users: This configuration is the default starting
371 with Fedora 18 and RHEL 7. If you find any problems, please let us
372 know or file a bug.
373 SVirt provides a hardened appliance using SELinux, making it very hard
375 for a rogue disk image to "escape" from the confinement of libguestfs
376 and damage the host (it's fair to say that even in standard libguestfs
377 this would be hard, but sVirt provides an extra layer of protection for
378 the host and more importantly protects virtual machines on the same
379 host from each other).
380 Currently to enable sVirt you will need libvirt ≥ 0.10.2 (1.0 or later
382 preferred), libguestfs ≥ 1.20, and the SELinux policies from recent
383 Fedora. If you are not running Fedora 18+, you will need to make
384 changes to your SELinux policy - contact us on the mailing list.
385 Once you have the requirements, do:
387 ./configure --with-default-backend=libvirt # libguestfs >= 1.22
389 ./configure --with-default-attach-method=libvirt # libguestfs <= 1.20
390 make
391 Set SELinux to Enforcing mode, and sVirt should be used automatically.
393 All, or almost all, features of libguestfs should work under sVirt.
395 There is one known shortcoming: virt-rescue(1) will not use libvirt
396 (hence sVirt), but falls back to direct launch of qemu. So you won't
397 currently get the benefit of sVirt protection when using virt-rescue.
398 You can check if sVirt is being used by enabling libvirtd logging (see
400 /etc/libvirt/libvirtd.log), killing and restarting libvirtd, and
401 checking the log files for "Setting SELinux context on ..." messages.
402 In theory sVirt should support AppArmor, but we have not tried it. It
404 will almost certainly require patching libvirt and writing an AppArmor
405 policy.
406 Libguestfs has a really long list of dependencies!
408 The base library doesn't depend on very much, but there are three
409 causes of the long list of other dependencies:
410 1. Libguestfs has to be able to read and edit many different disk
412 formats. For example, XFS support requires XFS tools.
413 2. There are language bindings for many different languages, all
415 requiring their own development tools. All language bindings
416 (except C) are optional.
417 3. There are some optional library features which can be disabled.
419 Since libguestfs ≥ 1.26 it is possible to split up the appliance
421 dependencies (item 1 in the list above) and thus have (eg)
422 "libguestfs-xfs" as a separate subpackage for processing XFS disk
423 images. We encourage downstream packagers to start splitting the base
424 libguestfs package into smaller subpackages.
425 Errors during launch on Fedora ≥ 18, RHEL ≥ 7
427 In Fedora ≥ 18 and RHEL ≥ 7, libguestfs uses libvirt to manage the
428 appliance. Previously (and upstream) libguestfs runs qemu directly:
429 ┌──────────────────────────────────┐
431 │ libguestfs │
432 ├────────────────┬─────────────────┤
433 │ direct backend │ libvirt backend │
434 └────────────────┴─────────────────┘
435 ↓ ↓
436 ┌───────┐ ┌──────────┐
437 │ qemu │ │ libvirtd │
438 └───────┘ └──────────┘
439 ↓
440 ┌───────┐
441 │ qemu │
442 └───────┘
443 upstream Fedora 18+
445 non-Fedora RHEL 7+
446 non-RHEL
447 The libvirt backend is more sophisticated, supporting SELinux/sVirt
449 (see above) and more. It is, however, more complex and so less robust.
450 If you have permissions problems using the libvirt backend, you can
452 switch to the direct backend by setting this environment variable:
453 export LIBGUESTFS_BACKEND=direct
455 before running any libguestfs program or virt tool.
457 How can I switch to a fixed / prebuilt appliance?
459 This may improve the stability and performance of libguestfs on Fedora
460 and RHEL.
461 Any time after installing libguestfs, run the following commands as
463 root:
464 mkdir -p /usr/local/lib/guestfs/appliance
466 libguestfs-make-fixed-appliance /usr/local/lib/guestfs/appliance
467 ls -l /usr/local/lib/guestfs/appliance
468 Now set the following environment variable before using libguestfs or
470 any virt tool:
471 export LIBGUESTFS_PATH=/usr/local/lib/guestfs/appliance
473 Of course you can change the path to any directory you want. You can
475 share the appliance across machines that have the same architecture
476 (eg. all x86-64), but note that libvirt will prevent you from sharing
477 the appliance across NFS because of permissions problems (so either
478 switch to the direct backend or don't use NFS).
479 How can I speed up libguestfs builds?
481 By far the most important thing you can do is to install and properly
482 configure Squid. Note that the default configuration that ships with
483 Squid is rubbish, so configuring it is not optional.
484 A very good place to start with Squid configuration is here:
486 https://fedoraproject.org/wiki/Extras/MockTricks#Using_Squid_to_Speed_Up_Mock_package_downloads
487 Make sure Squid is running, and that the environment variables
489 $http_proxy and $ftp_proxy are pointing to it.
490 With Squid running and correctly configured, appliance builds should be
492 reduced to a few minutes.
493 How can I speed up libguestfs builds (Debian)?
495 Hilko Bengen suggests using "approx" which is a Debian archive proxy
497 (http://packages.debian.org/approx). This tool is documented on Debian
498 in the approx(8) manual page.
499
501 Note: Most of the information in this section has moved:
502 guestfs-performance(1).
503 Upload or write seem very slow.
505 If the underlying disk is not fully allocated (eg. sparse raw or qcow2)
506 then writes can be slow because the host operating system has to do
507 costly disk allocations while you are writing. The solution is to use a
508 fully allocated format instead, ie. non-sparse raw, or qcow2 with the
509 "preallocation=metadata" option.
510 Libguestfs uses too much disk space!
512 libguestfs caches a large-ish appliance in:
513 /var/tmp/.guestfs-<UID>
515 If the environment variable "TMPDIR" is defined, then
517 $TMPDIR/.guestfs-<UID> is used instead.
518 It is safe to delete this directory when you are not using libguestfs.
520 virt-sparsify seems to make the image grow to the full size of the virtual
522 disk
523 If the input to virt-sparsify(1) is raw, then the output will be raw
524 sparse. Make sure you are measuring the output with a tool which
525 understands sparseness such as "du -sh". It can make a huge
526 difference:
527 $ ls -lh test1.img
529 -rw-rw-r--. 1 rjones rjones 100M Aug 8 08:08 test1.img
530 $ du -sh test1.img
531 3.6M test1.img
532 (Compare the apparent size 100M vs the actual size 3.6M)
534 If all this confuses you, use a non-sparse output format by specifying
536 the --convert option, eg:
537 virt-sparsify --convert qcow2 disk.raw disk.qcow2
539 Why doesn't virt-resize work on the disk image in-place?
541 Resizing a disk image is very tricky -- especially making sure that you
542 don't lose data or break the bootloader. The current method
543 effectively creates a new disk image and copies the data plus
544 bootloader from the old one. If something goes wrong, you can always
545 go back to the original.
546 If we were to make virt-resize work in-place then there would have to
548 be limitations: for example, you wouldn't be allowed to move existing
549 partitions (because moving data across the same disk is most likely to
550 corrupt data in the event of a power failure or crash), and LVM would
551 be very difficult to support (because of the almost arbitrary mapping
552 between LV content and underlying disk blocks).
553 Another method we have considered is to place a snapshot over the
555 original disk image, so that the original data is untouched and only
556 differences are recorded in the snapshot. You can do this today using
557 "qemu-img create" + "virt-resize", but qemu currently isn't smart
558 enough to recognize when the same block is written back to the snapshot
559 as already exists in the backing disk, so you will find that this
560 doesn't save you any space or time.
561 In summary, this is a hard problem, and what we have now mostly works
563 so we are reluctant to change it.
564 Why doesn't virt-sparsify work on the disk image in-place?
566 In libguestfs ≥ 1.26, virt-sparsify can now work on disk images in
567 place. Use:
568 virt-sparsify --in-place disk.img
570 But first you should read "IN-PLACE SPARSIFICATION" in
572 virt-sparsify(1).
573
575 Remote libvirt guests cannot be opened.
576 Opening remote libvirt guests is not supported at this time. For
577 example this won't work:
578 guestfish -c qemu://remote/system -d Guest
580 To open remote disks you have to export them somehow, then connect to
582 the export. For example if you decided to use NBD:
583 remote$ qemu-nbd -t -p 10809 guest.img
585 local$ guestfish -a nbd://remote:10809 -i
586 Other possibilities include ssh (if qemu is recent enough), NFS or
588 iSCSI. See "REMOTE STORAGE" in guestfs(3).
589 How can I open this strange disk source?
591 You have a disk image located inside another system that requires
592 access via a library / HTTP / REST / proprietary API, or is compressed
593 or archived in some way. (One example would be remote access to
594 OpenStack glance images without actually downloading them.)
595 We have a sister project called nbdkit
597 (https://github.com/libguestfs/nbdkit). This project lets you turn any
598 disk source into an NBD server. Libguestfs can access NBD servers
599 directly, eg:
600 guestfish -a nbd://remote
602 nbdkit is liberally licensed, so you can link it to or include it in
604 proprietary libraries and code. It also has a simple, stable plugin
605 API so you can easily write plugins against the API which will continue
606 to work in future.
607 Error opening VMDK disks: "uses a vmdk feature which is not supported by
609 this qemu version: VMDK version 3"
610 Qemu (and hence libguestfs) only supports certain VMDK disk images.
611 Others won't work, giving this or similar errors.
612 Ideally someone would fix qemu to support the latest VMDK features, but
614 in the meantime you have three options:
615 1. If the guest is hosted on a live, reachable ESX server, then locate
617 and download the disk image called somename-flat.vmdk. Despite the
618 name, this is a raw disk image, and can be opened by anything.
619 If you have a recent enough version of qemu and libguestfs, then
621 you may be able to access this disk image remotely using either
622 HTTPS or ssh. See "REMOTE STORAGE" in guestfs(3).
623 2. Use VMware’s proprietary vdiskmanager tool to convert the image to
625 raw format.
626 3. Use nbdkit with the proprietary VDDK plugin to live export the disk
628 image as an NBD source. This should allow you to read and write
629 the VMDK file.
630 UFS disks (as used by BSD) cannot be opened.
632 The UFS filesystem format has many variants, and these are not self-
633 identifying. The Linux kernel has to be told which variant of UFS it
634 has to use, which libguestfs cannot know.
635 You have to pass the right "ufstype" mount option when mounting these
637 filesystems.
638 See https://www.kernel.org/doc/Documentation/filesystems/ufs.txt
640 Windows ReFS
642 Windows ReFS is Microsoft’s ZFS/Btrfs copy. This filesystem has not
643 yet been reverse engineered and implemented in the Linux kernel, and
644 therefore libguestfs doesn't support it. At the moment it seems to be
645 very rare "in the wild".
646 Non-ASCII characters don’t appear on VFAT filesystems.
648 Typical symptoms of this problem:
649 • You get an error when you create a file where the filename contains
651 non-ASCII characters, particularly non 8-bit characters from Asian
652 languages (Chinese, Japanese, etc). The filesystem is VFAT.
653 • When you list a directory from a VFAT filesystem, filenames appear
655 as question marks.
656 This is a design flaw of the GNU/Linux system.
658 VFAT stores long filenames as UTF-16 characters. When opening or
660 returning filenames, the Linux kernel has to translate these to some
661 form of 8 bit string. UTF-8 would be the obvious choice, except for
662 Linux users who persist in using non-UTF-8 locales (the user’s locale
663 is not known to the kernel because it’s a function of libc).
664 Therefore you have to tell the kernel what translation you want done
666 when you mount the filesystem. The two methods are the "iocharset"
667 parameter (which is not relevant to libguestfs) and the "utf8" flag.
668 So to use a VFAT filesystem you must add the "utf8" flag when mounting.
670 From guestfish, use:
671 ><fs> mount-options utf8 /dev/sda1 /
673 or on the guestfish command line:
675 guestfish [...] -m /dev/sda1:/:utf8
677 or from the API:
679 guestfs_mount_options (g, "utf8", "/dev/sda1", "/");
681 The kernel will then translate filenames to and from UTF-8 strings.
683 We considered adding this mount option transparently, but unfortunately
685 there are several problems with doing that:
686 • On some Linux systems, the "utf8" mount option doesn't work. We
688 don't precisely understand what systems or why, but this was
689 reliably reported by one user.
690 • It would prevent you from using the "iocharset" parameter because
692 it is incompatible with "utf8". It is probably not a good idea to
693 use this parameter, but we don't want to prevent it.
694 Non-ASCII characters appear as underscore (_) on ISO9660 filesystems.
696 The filesystem was not prepared correctly with mkisofs or genisoimage.
697 Make sure the filesystem was created using Joliet and/or Rock Ridge
698 extensions. libguestfs does not require any special mount options to
699 handle the filesystem.
700 Cannot open Windows guests which use NTFS.
702 You see errors like:
703 mount: unknown filesystem type 'ntfs'
705 On Red Hat Enterprise Linux or CentOS < 7.2, you have to install the
707 libguestfs-winsupport package. In RHEL ≥ 7.2, "libguestfs-winsupport"
708 is part of the base RHEL distribution, but see the next question.
709 "mount: unsupported filesystem type" with NTFS in RHEL ≥ 7.2
711 In RHEL 7.2 we were able to add "libguestfs-winsupport" to the base
712 RHEL distribution, but we had to disable the ability to use it for
713 opening and editing filesystems. It is only supported when used with
714 virt-v2v(1). If you try to use guestfish(1) or guestmount(1) or some
715 other programs on an NTFS filesystem, you will see the error:
716 mount: unsupported filesystem type
718 This is not a supported configuration, and it will not be made to work
720 in RHEL. Don't bother to open a bug about it, as it will be
721 immediately "CLOSED -> WONTFIX".
722 You may compile your own libguestfs removing this restriction, but that
724 won't be endorsed or supported by Red Hat.
725 Cannot open or inspect RHEL 7 guests.
727 Cannot open Linux guests which use XFS.
728 RHEL 7 guests, and any other guests that use XFS, can be opened by
729 libguestfs, but you have to install the "libguestfs-xfs" package.
730
732 The API has hundreds of methods, where do I start?
733 We recommend you start by reading the API overview: "API OVERVIEW" in
734 guestfs(3).
735 Although the API overview covers the C API, it is still worth reading
737 even if you are going to use another programming language, because the
738 API is the same, just with simple logical changes to the names of the
739 calls:
740 C guestfs_ln_sf (g, target, linkname);
742 Python g.ln_sf (target, linkname);
743 OCaml g#ln_sf target linkname;
744 Perl $g->ln_sf (target, linkname);
745 Shell (guestfish) ln-sf target linkname
746 PHP guestfs_ln_sf ($g, $target, $linkname);
747 Once you're familiar with the API overview, you should look at this
749 list of starting points for other language bindings: "USING LIBGUESTFS
750 WITH OTHER PROGRAMMING LANGUAGES" in guestfs(3).
751 Can I use libguestfs in my proprietary / closed source / commercial
753 program?
754 In general, yes. However this is not legal advice - read the license
755 that comes with libguestfs, and if you have specific questions contact
756 a lawyer.
757 In the source tree the license is in the file "COPYING.LIB" (LGPLv2+
759 for the library and bindings) and "COPYING" (GPLv2+ for the standalone
760 programs).
761
763 Help, it’s not working!
764 If no libguestfs program seems to work at all, run the program below
765 and paste the complete, unedited output into an email to "libguestfs" @
766 "redhat.com":
767 libguestfs-test-tool
769 If a particular operation fails, supply all the information in this
771 checklist, in an email to "libguestfs" @ "redhat.com":
772 1. What are you trying to do?
774 2. What exact command(s) did you run?
776 3. What was the precise error or output of these commands?
778 4. Enable debugging, run the commands again, and capture the complete
780 output. Do not edit the output.
781 export LIBGUESTFS_DEBUG=1
783 export LIBGUESTFS_TRACE=1
784 5. Include the version of libguestfs, the operating system version,
786 and how you installed libguestfs (eg. from source, "yum install",
787 etc.)
788 How do I debug when using any libguestfs program or tool (eg. virt-
790 customize or virt-df)?
791 There are two "LIBGUESTFS_*" environment variables you can set in order
792 to get more information from libguestfs.
793 "LIBGUESTFS_TRACE"
795 Set this to 1 and libguestfs will print out each command / API call
796 in a format which is similar to guestfish commands.
797 "LIBGUESTFS_DEBUG"
799 Set this to 1 in order to enable massive amounts of debug messages.
800 If you think there is some problem inside the libguestfs appliance,
801 then you should use this option.
802 To set these from the shell, do this before running the program:
804 export LIBGUESTFS_TRACE=1
806 export LIBGUESTFS_DEBUG=1
807 For csh/tcsh the equivalent commands would be:
809 setenv LIBGUESTFS_TRACE 1
811 setenv LIBGUESTFS_DEBUG 1
812 For further information, see: "ENVIRONMENT VARIABLES" in guestfs(3).
814 How do I debug when using guestfish?
816 You can use the same environment variables above. Alternatively use
817 the guestfish options -x (to trace commands) or -v (to get the full
818 debug output), or both.
819 For further information, see: guestfish(1).
821 How do I debug when using the API?
823 Call "guestfs_set_trace" in guestfs(3) to enable command traces, and/or
824 "guestfs_set_verbose" in guestfs(3) to enable debug messages.
825 For best results, call these functions as early as possible, just after
827 creating the guestfs handle if you can, and definitely before calling
828 launch.
829 How do I capture debug output and put it into my logging system?
831 Use the event API. For examples, see: "SETTING CALLBACKS TO HANDLE
832 EVENTS" in guestfs(3) and the examples/debug-logging.c program in the
833 libguestfs sources.
834 Digging deeper into the appliance boot process.
836 Enable debugging and then read this documentation on the appliance boot
837 process: guestfs-internals(1).
838 libguestfs hangs or fails during run/launch.
840 Enable debugging and look at the full output. If you cannot work out
841 what is going on, file a bug report, including the complete output of
842 libguestfs-test-tool(1).
843 Debugging libvirt
845 If you are using the libvirt backend, and libvirt is failing, then you
846 can enable debugging by editing /etc/libvirt/libvirtd.conf.
847 If you are running as non-root, then you have to edit a different file.
849 Create ~/.config/libvirt/libvirtd.conf containing:
850 log_level=1
852 log_outputs="1:file:/tmp/libvirtd.log"
853 Kill any session (non-root) libvirtd that is running, and next time you
855 run the libguestfs command, you should see a large amount of useful
856 debugging information from libvirtd in /tmp/libvirtd.log
857 Broken kernel, or trying a different kernel.
859 You can choose a different kernel for the appliance by setting some
860 supermin environment variables:
861 export SUPERMIN_KERNEL_VERSION=4.8.0-1.fc25.x86_64
863 export SUPERMIN_KERNEL=/boot/vmlinuz-$SUPERMIN_KERNEL_VERSION
864 export SUPERMIN_MODULES=/lib/modules/$SUPERMIN_KERNEL_VERSION
865 rm -rf /var/tmp/.guestfs-*
866 libguestfs-test-tool
867 Broken qemu, or trying a different qemu.
869 You can choose a different qemu by setting the hypervisor environment
870 variable:
871 export LIBGUESTFS_HV=/path/to/qemu-system-x86_64
873 libguestfs-test-tool
874
876 See also guestfs-internals(1).
877 Why don’t you do everything through the FUSE / filesystem interface?
879 We offer a command called guestmount(1) which lets you mount guest
880 filesystems on the host. This is implemented as a FUSE module. Why
881 don't we just implement the whole of libguestfs using this mechanism,
882 instead of having the large and rather complicated API?
883 The reasons are twofold. Firstly, libguestfs offers API calls for
885 doing things like creating and deleting partitions and logical volumes,
886 which don't fit into a filesystem model very easily. Or rather, you
887 could fit them in: for example, creating a partition could be mapped to
888 "mkdir /fs/hda1" but then you'd have to specify some method to choose
889 the size of the partition (maybe "echo 100M > /fs/hda1/.size"), and the
890 partition type, start and end sectors etc., but once you've done that
891 the filesystem-based API starts to look more complicated than the call-
892 based API we currently have.
893 The second reason is for efficiency. FUSE itself is reasonably
895 efficient, but it does make lots of small, independent calls into the
896 FUSE module. In guestmount these have to be translated into messages
897 to the libguestfs appliance which has a big overhead (in time and round
898 trips). For example, reading a file in 64 KB chunks is inefficient
899 because each chunk would turn into a single round trip. In the
900 libguestfs API it is much more efficient to download an entire file or
901 directory through one of the streaming calls like "guestfs_download" or
902 "guestfs_tar_out".
903 Why don’t you do everything through GVFS?
905 The problems are similar to the problems with FUSE.
906 GVFS is a better abstraction than POSIX/FUSE. There is an FTP backend
908 for GVFS, which is encouraging because FTP is conceptually similar to
909 the libguestfs API. However the GVFS FTP backend makes multiple
910 simultaneous connections in order to keep interactivity, which we can't
911 easily do with libguestfs.
912 Why can I write to the disk, even though I added it read-only?
914 Why does "--ro" appear to have no effect?
915 When you add a disk read-only, libguestfs places a writable overlay on
916 top of the underlying disk. Writes go into this overlay, and are
917 discarded when the handle is closed (or "guestfish" etc. exits).
918 There are two reasons for doing it this way: Firstly read-only disks
920 aren't possible in many cases (eg. IDE simply doesn't support them, so
921 you couldn't have an IDE-emulated read-only disk, although this is not
922 common in real libguestfs installations).
923 Secondly and more importantly, even if read-only disks were possible,
925 you wouldn't want them. Mounting any filesystem that has a journal,
926 even "mount -o ro", causes writes to the filesystem because the journal
927 has to be replayed and metadata updated. If the disk was truly read-
928 only, you wouldn't be able to mount a dirty filesystem.
929 To make it usable, we create the overlay as a place to temporarily
931 store these writes, and then we discard it afterwards. This ensures
932 that the underlying disk is always untouched.
933 Note also that there is a regression test for this when building
935 libguestfs (in "tests/qemu"). This is one reason why it’s important
936 for packagers to run the test suite.
937 Does "--ro" make all disks read-only?
939 No! The "--ro" option only affects disks added on the command line,
940 ie. using "-a" and "-d" options.
941 In guestfish, if you use the "add" command, then disk is added read-
943 write (unless you specify the "readonly:true" flag explicitly with the
944 command).
945 Can I use "guestfish --ro" as a way to backup my virtual machines?
947 Usually this is not a good idea. The question is answered in more
948 detail in this mailing list posting:
949 https://www.redhat.com/archives/libguestfs/2010-August/msg00024.html
950 See also the next question.
952 Why can’t I run fsck on a live filesystem using "guestfish --ro"?
954 This command will usually not work:
955 guestfish --ro -a /dev/vg/my_root_fs run : fsck /dev/sda
957 The reason for this is that qemu creates a snapshot over the original
959 filesystem, but it doesn't create a strict point-in-time snapshot.
960 Blocks of data on the underlying filesystem are read by qemu at
961 different times as the fsck operation progresses, with host writes in
962 between. The result is that fsck sees massive corruption (imaginary,
963 not real!) and fails.
964 What you have to do is to create a point-in-time snapshot. If it’s a
966 logical volume, use an LVM2 snapshot. If the filesystem is located
967 inside something like a btrfs/ZFS file, use a btrfs/ZFS snapshot, and
968 then run the fsck on the snapshot. In practice you don't need to use
969 libguestfs for this -- just run /sbin/fsck directly.
970 Creating point-in-time snapshots of host devices and files is outside
972 the scope of libguestfs, although libguestfs can operate on them once
973 they are created.
974 What’s the difference between guestfish and virt-rescue?
976 A lot of people are confused by the two superficially similar tools we
977 provide:
978 $ guestfish --ro -a guest.img
980 ><fs> run
981 ><fs> fsck /dev/sda1
982 $ virt-rescue --ro guest.img
984 ><rescue> /sbin/fsck /dev/sda1
985 And the related question which then arises is why you can’t type in
987 full shell commands with all the --options in guestfish (but you can in
988 virt-rescue(1)).
989 guestfish(1) is a program providing structured access to the guestfs(3)
991 API. It happens to be a nice interactive shell too, but its primary
992 purpose is structured access from shell scripts. Think of it more like
993 a language binding, like Python and other bindings, but for shell. The
994 key differentiating factor of guestfish (and the libguestfs API in
995 general) is the ability to automate changes.
996 virt-rescue(1) is a free-for-all freeform way to boot the libguestfs
998 appliance and make arbitrary changes to your VM. It’s not structured,
999 you can't automate it, but for making quick ad-hoc fixes to your
1000 guests, it can be quite useful.
1001 But, libguestfs also has a "backdoor" into the appliance allowing you
1003 to send arbitrary shell commands. It’s not as flexible as virt-rescue,
1004 because you can't interact with the shell commands, but here it is
1005 anyway:
1006 ><fs> debug sh "cmd arg1 arg2 ..."
1008 Note that you should not rely on this. It could be removed or changed
1010 in future. If your program needs some operation, please add it to the
1011 libguestfs API instead.
1012 What’s the deal with "guestfish -i"?
1014 Why does virt-cat only work on a real VM image, but virt-df works on any
1015 disk image?
1016 What does "no root device found in this operating system image" mean?
1017 These questions are all related at a fundamental level which may not be
1018 immediately obvious.
1019 At the guestfs(3) API level, a "disk image" is just a pile of
1021 partitions and filesystems.
1022 In contrast, when the virtual machine boots, it mounts those
1024 filesystems into a consistent hierarchy such as:
1025 / (/dev/sda2)
1027 │
1028 ├── /boot (/dev/sda1)
1029 │
1030 ├── /home (/dev/vg_external/Homes)
1031 │
1032 ├── /usr (/dev/vg_os/lv_usr)
1033 │
1034 └── /var (/dev/vg_os/lv_var)
1035 (or drive letters on Windows).
1037 The API first of all sees the disk image at the "pile of filesystems"
1039 level. But it also has a way to inspect the disk image to see if it
1040 contains an operating system, and how the disks are mounted when the
1041 operating system boots: "INSPECTION" in guestfs(3).
1042 Users expect some tools (like virt-cat(1)) to work with VM paths:
1044 virt-cat fedora.img /var/log/messages
1046 How does virt-cat know that /var is a separate partition? The trick is
1048 that virt-cat performs inspection on the disk image, and uses that to
1049 translate the path correctly.
1050 Some tools (including virt-cat(1), virt-edit(1), virt-ls(1)) use
1052 inspection to map VM paths. Other tools, such as virt-df(1) and
1053 virt-filesystems(1) operate entirely at the raw "big pile of
1054 filesystems" level of the libguestfs API, and don't use inspection.
1055 guestfish(1) is in an interesting middle ground. If you use the -a and
1057 -m command line options, then you have to tell guestfish exactly how to
1058 add disk images and where to mount partitions. This is the raw API
1059 level.
1060 If you use the -i option, libguestfs performs inspection and mounts the
1062 filesystems for you.
1063 The error "no root device found in this operating system image" is
1065 related to this. It means inspection was unable to locate an operating
1066 system within the disk image you gave it. You might see this from
1067 programs like virt-cat if you try to run them on something which is
1068 just a disk image, not a virtual machine disk image.
1069 What do these "debug*" and "internal-*" functions do?
1071 There are some functions which are used for debugging and internal
1072 purposes which are not part of the stable API.
1073 The "debug*" (or "guestfs_debug*") functions, primarily "guestfs_debug"
1075 in guestfs(3) and a handful of others, are used for debugging
1076 libguestfs. Although they are not part of the stable API and thus may
1077 change or be removed at any time, some programs may want to call these
1078 while waiting for features to be added to libguestfs.
1079 The "internal-*" (or "guestfs_internal_*") functions are purely to be
1081 used by libguestfs itself. There is no reason for programs to call
1082 them, and programs should not try to use them. Using them will often
1083 cause bad things to happen, as well as not being part of the documented
1084 stable API.
1085
1087 Where do I send patches?
1088 Please send patches to the libguestfs mailing list
1089 https://www.redhat.com/mailman/listinfo/libguestfs. You don't have to
1090 be subscribed, but there will be a delay until your posting is manually
1091 approved.
1092 Please don’t use github pull requests - they will be ignored. The
1094 reasons are (a) we want to discuss and dissect patches on the mailing
1095 list, and (b) github pull requests turn into merge commits but we
1096 prefer to have a linear history.
1097 How do I propose a feature?
1099 Large new features that you intend to contribute should be discussed on
1100 the mailing list first
1101 (https://www.redhat.com/mailman/listinfo/libguestfs). This avoids
1102 disappointment and wasted work if we don't think the feature would fit
1103 into the libguestfs project.
1104 If you want to suggest a useful feature but don’t want to write the
1106 code, you can file a bug (see "GETTING HELP AND REPORTING BUGS") with
1107 "RFE: " at the beginning of the Summary line.
1108 Who can commit to libguestfs git?
1110 About 5 people have commit access to github. Patches should be posted
1111 on the list first and ACKed. The policy for ACKing and pushing patches
1112 is outlined here:
1113 https://www.redhat.com/archives/libguestfs/2012-January/msg00023.html
1115 Can I fork libguestfs?
1117 Of course you can. Git makes it easy to fork libguestfs. Github makes
1118 it even easier. It’s nice if you tell us on the mailing list about
1119 forks and the reasons for them.
1120
1122 Can I monitor the live disk activity of a virtual machine using libguestfs?
1123 A common request is to be able to use libguestfs to monitor the live
1124 disk activity of a guest, for example, to get notified every time a
1125 guest creates a new file. Libguestfs does not work in the way some
1126 people imagine, as you can see from this diagram:
1127 ┌─────────────────────────────────────┐
1129 │ monitoring program using libguestfs │
1130 └─────────────────────────────────────┘
1131 ↓
1132 ┌───────────┐ ┌──────────────────────┐
1133 │ live VM │ │ libguestfs appliance │
1134 ├───────────┤ ├──────────────────────┤
1135 │ kernel (1)│ │ appliance kernel (2) │
1136 └───────────┘ └──────────────────────┘
1137 ↓ ↓ (r/o connection)
1138 ┌──────────────────────┐
1139 | disk image |
1140 └──────────────────────┘
1141 This scenario is safe (as long as you set the "readonly" flag when
1143 adding the drive). However the libguestfs appliance kernel (2) does
1144 not see all the changes made to the disk image, for two reasons:
1145 i. The VM kernel (1) can cache data in memory, so it doesn't appear in
1147 the disk image.
1148 ii. The libguestfs appliance kernel (2) doesn't expect that the disk
1150 image is changing underneath it, so its own cache is not magically
1151 updated even when the VM kernel (1) does update the disk image.
1152 The only supported solution is to restart the entire libguestfs
1154 appliance whenever you want to look at changes in the disk image. At
1155 the API level that corresponds to calling "guestfs_shutdown" followed
1156 by "guestfs_launch", which is a heavyweight operation (see also
1157 guestfs-performance(3)).
1158 There are some unsupported hacks you can try if relaunching the
1160 appliance is really too costly:
1161 • Call "guestfs_drop_caches (g, 3)". This causes all cached data
1163 help by the libguestfs appliance kernel (2) to be discarded, so it
1164 goes back to the disk image.
1165 However this on its own is not sufficient, because qemu also caches
1167 some data. You will also need to patch libguestfs to (re-)enable
1168 the "cache=none" mode. See:
1169 https://rwmj.wordpress.com/2013/09/02/new-in-libguestfs-allow-cache-mode-to-be-selected/
1170 • Use a tool like virt-bmap instead.
1172 • Run an agent inside the guest.
1174 Nothing helps if the guest is making more fundamental changes (eg.
1176 deleting filesystems). For those kinds of things you must relaunch the
1177 appliance.
1178 (Note there is a third problem that you need to use consistent
1180 snapshots to really examine live disk images, but that’s a general
1181 problem with using libguestfs against any live disk image.)
1182
1184 guestfish(1), guestfs(3), http://libguestfs.org/.
1185
1187 Richard W.M. Jones ("rjones at redhat dot com")
1188
1190 Copyright (C) 2012-2020 Red Hat Inc.
1191
1193 This library is free software; you can redistribute it and/or modify it
1194 under the terms of the GNU Lesser General Public License as published
1195 by the Free Software Foundation; either version 2 of the License, or
1196 (at your option) any later version.
1197 This library is distributed in the hope that it will be useful, but
1199 WITHOUT ANY WARRANTY; without even the implied warranty of
1200 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1201 Lesser General Public License for more details.
1202 You should have received a copy of the GNU Lesser General Public
1204 License along with this library; if not, write to the Free Software
1205 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
1206 02110-1301 USA
1207
1209 To get a list of bugs against libguestfs, use this link:
1210 https://bugzilla.redhat.com/buglist.cgi?component=libguestfs&product=Virtualization+Tools
1211 To report a new bug against libguestfs, use this link:
1213 https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virtualization+Tools
1214 When reporting a bug, please supply:
1216 • The version of libguestfs.
1218 • Where you got libguestfs (eg. which Linux distro, compiled from
1220 source, etc)
1221 • Describe the bug accurately and give a way to reproduce it.
1223 • Run libguestfs-test-tool(1) and paste the complete, unedited output
1225 into the bug report.
1226libguestfs-1.49.9 2023-01-19 guestfs-faq(1)