5 guestfs - Library for accessing and modifying virtual machine images
11 guestfs_h *g = guestfs_create ();
12 guestfs_add_drive (g, "guest.img");
14 guestfs_mount (g, "/dev/sda1", "/");
15 guestfs_touch (g, "/hello");
16 guestfs_umount (g, "/");
20 cc prog.c -o prog -lguestfs
22 cc prog.c -o prog `pkg-config libguestfs --cflags --libs`
26 Libguestfs is a library for accessing and modifying guest disk images.
27 Amongst the things this is good for: making batch configuration
28 changes to guests, getting disk used/free statistics (see also:
29 virt-df), migrating between virtualization systems (see also:
30 virt-p2v), performing partial backups, performing partial guest
31 clones, cloning guests and changing registry/UUID/hostname info, and
34 Libguestfs uses Linux kernel and qemu code, and can access any type of
35 guest filesystem that Linux and qemu can, including but not limited
36 to: ext2/3/4, btrfs, FAT and NTFS, LVM, many different disk partition
37 schemes, qcow, qcow2, vmdk.
39 Libguestfs provides ways to enumerate guest storage (eg. partitions,
40 LVs, what filesystem is in each LV, etc.). It can also run commands
41 in the context of the guest. Also you can access filesystems over
44 Libguestfs is a library that can be linked with C and C++ management
45 programs (or management programs written in OCaml, Perl, Python, Ruby,
46 Java, PHP, Haskell or C#). You can also use it from shell scripts or the
49 You don't need to be root to use libguestfs, although obviously you do
50 need enough permissions to access the disk images.
52 Libguestfs is a large API because it can do many things. For a gentle
53 introduction, please read the L</API OVERVIEW> section next.
57 This section provides a gentler overview of the libguestfs API. We
58 also try to group API calls together, where that may not be obvious
59 from reading about the individual calls in the main section of this
64 Before you can use libguestfs calls, you have to create a handle.
65 Then you must add at least one disk image to the handle, followed by
66 launching the handle, then performing whatever operations you want,
67 and finally closing the handle. By convention we use the single
68 letter C<g> for the name of the handle variable, although of course
69 you can use any name you want.
71 The general structure of all libguestfs-using programs looks like
74 guestfs_h *g = guestfs_create ();
76 /* Call guestfs_add_drive additional times if there are
77 * multiple disk images.
79 guestfs_add_drive (g, "guest.img");
81 /* Most manipulation calls won't work until you've launched
82 * the handle 'g'. You have to do this _after_ adding drives
83 * and _before_ other commands.
87 /* Now you can examine what partitions, LVs etc are available.
89 char **partitions = guestfs_list_partitions (g);
90 char **logvols = guestfs_lvs (g);
92 /* To access a filesystem in the image, you must mount it.
94 guestfs_mount (g, "/dev/sda1", "/");
96 /* Now you can perform filesystem actions on the guest
99 guestfs_touch (g, "/hello");
101 /* You only need to call guestfs_sync if you have made
102 * changes to the guest image. (But if you've made changes
103 * then you *must* sync). See also: guestfs_umount and
104 * guestfs_umount_all calls.
108 /* Close the handle 'g'. */
111 The code above doesn't include any error checking. In real code you
112 should check return values carefully for errors. In general all
113 functions that return integers return C<-1> on error, and all
114 functions that return pointers return C<NULL> on error. See section
115 L</ERROR HANDLING> below for how to handle errors, and consult the
116 documentation for each function call below to see precisely how they
117 return error indications.
121 The image filename (C<"guest.img"> in the example above) could be a
122 disk image from a virtual machine, a L<dd(1)> copy of a physical hard
123 disk, an actual block device, or simply an empty file of zeroes that
124 you have created through L<posix_fallocate(3)>. Libguestfs lets you
125 do useful things to all of these.
127 The call you should use in modern code for adding drives is
128 L</guestfs_add_drive_opts>. To add a disk image, allowing writes, and
129 specifying that the format is raw, do:
131 guestfs_add_drive_opts (g, filename,
132 GUESTFS_ADD_DRIVE_OPTS_FORMAT, "raw",
135 You can add a disk read-only using:
137 guestfs_add_drive_opts (g, filename,
138 GUESTFS_ADD_DRIVE_OPTS_FORMAT, "raw",
139 GUESTFS_ADD_DRIVE_OPTS_READONLY, 1,
142 or by calling the older function L</guestfs_add_drive_ro>. In either
143 case libguestfs won't modify the file.
145 Be extremely cautious if the disk image is in use, eg. if it is being
146 used by a virtual machine. Adding it read-write will almost certainly
147 cause disk corruption, but adding it read-only is safe.
149 You must add at least one disk image, and you may add multiple disk
150 images. In the API, the disk images are usually referred to as
151 C</dev/sda> (for the first one you added), C</dev/sdb> (for the second
154 Once L</guestfs_launch> has been called you cannot add any more images.
155 You can call L</guestfs_list_devices> to get a list of the device
156 names, in the order that you added them. See also L</BLOCK DEVICE
161 Before you can read or write files, create directories and so on in a
162 disk image that contains filesystems, you have to mount those
163 filesystems using L</guestfs_mount>. If you already know that a disk
164 image contains (for example) one partition with a filesystem on that
165 partition, then you can mount it directly:
167 guestfs_mount (g, "/dev/sda1", "/");
169 where C</dev/sda1> means literally the first partition (C<1>) of the
170 first disk image that we added (C</dev/sda>). If the disk contains
171 Linux LVM2 logical volumes you could refer to those instead (eg. C</dev/VG/LV>).
173 If you are given a disk image and you don't know what it contains then
174 you have to find out. Libguestfs can do that too: use
175 L</guestfs_list_partitions> and L</guestfs_lvs> to list possible
176 partitions and LVs, and either try mounting each to see what is
177 mountable, or else examine them with L</guestfs_vfs_type> or
178 L</guestfs_file>. Libguestfs also has a set of APIs for inspection of
179 disk images (see L</INSPECTION> below). But you might find it easier
180 to look at higher level programs built on top of libguestfs, in
181 particular L<virt-inspector(1)>.
183 To mount a disk image read-only, use L</guestfs_mount_ro>. There are
184 several other variations of the C<guestfs_mount_*> call.
186 =head2 FILESYSTEM ACCESS AND MODIFICATION
188 The majority of the libguestfs API consists of fairly low-level calls
189 for accessing and modifying the files, directories, symlinks etc on
190 mounted filesystems. There are over a hundred such calls which you
191 can find listed in detail below in this man page, and we don't even
192 pretend to cover them all in this overview.
194 Specify filenames as full paths, starting with C<"/"> and including
197 For example, if you mounted a filesystem at C<"/"> and you want to
198 read the file called C<"etc/passwd"> then you could do:
200 char *data = guestfs_cat (g, "/etc/passwd");
202 This would return C<data> as a newly allocated buffer containing the
203 full content of that file (with some conditions: see also
204 L</DOWNLOADING> below), or C<NULL> if there was an error.
206 As another example, to create a top-level directory on that filesystem
207 called C<"var"> you would do:
209 guestfs_mkdir (g, "/var");
211 To create a symlink you could do:
213 guestfs_ln_s (g, "/etc/init.d/portmap",
214 "/etc/rc3.d/S30portmap");
216 Libguestfs will reject attempts to use relative paths and there is no
217 concept of a current working directory.
219 Libguestfs can return errors in many situations: for example if the
220 filesystem isn't writable, or if a file or directory that you
221 requested doesn't exist. If you are using the C API (documented here)
222 you have to check for those error conditions after each call. (Other
223 language bindings turn these errors into exceptions).
225 File writes are affected by the per-handle umask, set by calling
226 L</guestfs_umask> and defaulting to 022. See L</UMASK>.
230 Libguestfs contains API calls to read, create and modify partition
231 tables on disk images.
233 In the common case where you want to create a single partition
234 covering the whole disk, you should use the L</guestfs_part_disk>
237 const char *parttype = "mbr";
238 if (disk_is_larger_than_2TB)
240 guestfs_part_disk (g, "/dev/sda", parttype);
242 Obviously this effectively wipes anything that was on that disk image
247 Libguestfs provides access to a large part of the LVM2 API, such as
248 L</guestfs_lvcreate> and L</guestfs_vgremove>. It won't make much sense
249 unless you familiarize yourself with the concepts of physical volumes,
250 volume groups and logical volumes.
252 This author strongly recommends reading the LVM HOWTO, online at
253 L<http://tldp.org/HOWTO/LVM-HOWTO/>.
257 Use L</guestfs_cat> to download small, text only files. This call
258 is limited to files which are less than 2 MB and which cannot contain
259 any ASCII NUL (C<\0>) characters. However it has a very simple
262 L</guestfs_read_file> can be used to read files which contain
263 arbitrary 8 bit data, since it returns a (pointer, size) pair.
264 However it is still limited to "small" files, less than 2 MB.
266 L</guestfs_download> can be used to download any file, with no
267 limits on content or size (even files larger than 4 GB).
269 To download multiple files, see L</guestfs_tar_out> and
274 It's often the case that you want to write a file or files to the disk
277 To write a small file with fixed content, use L</guestfs_write>. To
278 create a file of all zeroes, use L</guestfs_truncate_size> (sparse) or
279 L</guestfs_fallocate64> (with all disk blocks allocated). There are a
280 variety of other functions for creating test files, for example
281 L</guestfs_fill> and L</guestfs_fill_pattern>.
283 To upload a single file, use L</guestfs_upload>. This call has no
284 limits on file content or size (even files larger than 4 GB).
286 To upload multiple files, see L</guestfs_tar_in> and L</guestfs_tgz_in>.
288 However the fastest way to upload I<large numbers of arbitrary files>
289 is to turn them into a squashfs or CD ISO (see L<mksquashfs(8)> and
290 L<mkisofs(8)>), then attach this using L</guestfs_add_drive_ro>. If
291 you add the drive in a predictable way (eg. adding it last after all
292 other drives) then you can get the device name from
293 L</guestfs_list_devices> and mount it directly using
294 L</guestfs_mount_ro>. Note that squashfs images are sometimes
295 non-portable between kernel versions, and they don't support labels or
296 UUIDs. If you want to pre-build an image or you need to mount it
297 using a label or UUID, use an ISO image instead.
301 There are various different commands for copying between files and
302 devices and in and out of the guest filesystem. These are summarised
307 =item B<file> to B<file>
309 Use L</guestfs_cp> to copy a single file, or
310 L</guestfs_cp_a> to copy directories recursively.
312 =item B<file or device> to B<file or device>
314 Use L</guestfs_dd> which efficiently uses L<dd(1)>
315 to copy between files and devices in the guest.
317 Example: duplicate the contents of an LV:
319 guestfs_dd (g, "/dev/VG/Original", "/dev/VG/Copy");
321 The destination (C</dev/VG/Copy>) must be at least as large as the
322 source (C</dev/VG/Original>). To copy less than the whole
323 source device, use L</guestfs_copy_size>.
325 =item B<file on the host> to B<file or device>
327 Use L</guestfs_upload>. See L</UPLOADING> above.
329 =item B<file or device> to B<file on the host>
331 Use L</guestfs_download>. See L</DOWNLOADING> above.
337 L</guestfs_ll> is just designed for humans to read (mainly when using
338 the L<guestfish(1)>-equivalent command C<ll>).
340 L</guestfs_ls> is a quick way to get a list of files in a directory
341 from programs, as a flat list of strings.
343 L</guestfs_readdir> is a programmatic way to get a list of files in a
344 directory, plus additional information about each one. It is more
345 equivalent to using the L<readdir(3)> call on a local filesystem.
347 L</guestfs_find> and L</guestfs_find0> can be used to recursively list
350 =head2 RUNNING COMMANDS
352 Although libguestfs is primarily an API for manipulating files
353 inside guest images, we also provide some limited facilities for
354 running commands inside guests.
356 There are many limitations to this:
362 The kernel version that the command runs under will be different
363 from what it expects.
367 If the command needs to communicate with daemons, then most likely
368 they won't be running.
372 The command will be running in limited memory.
376 The network may not be available unless you enable it
377 (see L</guestfs_set_network>).
381 Only supports Linux guests (not Windows, BSD, etc).
385 Architecture limitations (eg. won't work for a PPC guest on
390 For SELinux guests, you may need to enable SELinux and load policy
391 first. See L</SELINUX> in this manpage.
395 I<Security:> It is not safe to run commands from untrusted, possibly
396 malicious guests. These commands may attempt to exploit your program
397 by sending unexpected output. They could also try to exploit the
398 Linux kernel or qemu provided by the libguestfs appliance. They could
399 use the network provided by the libguestfs appliance to bypass
400 ordinary network partitions and firewalls. They could use the
401 elevated privileges or different SELinux context of your program
404 A secure alternative is to use libguestfs to install a "firstboot"
405 script (a script which runs when the guest next boots normally), and
406 to have this script run the commands you want in the normal context of
407 the running guest, network security and so on.
411 The two main API calls to run commands are L</guestfs_command> and
412 L</guestfs_sh> (there are also variations).
414 The difference is that L</guestfs_sh> runs commands using the shell, so
415 any shell globs, redirections, etc will work.
417 =head2 CONFIGURATION FILES
419 To read and write configuration files in Linux guest filesystems, we
420 strongly recommend using Augeas. For example, Augeas understands how
421 to read and write, say, a Linux shadow password file or X.org
422 configuration file, and so avoids you having to write that code.
424 The main Augeas calls are bound through the C<guestfs_aug_*> APIs. We
425 don't document Augeas itself here because there is excellent
426 documentation on the L<http://augeas.net/> website.
428 If you don't want to use Augeas (you fool!) then try calling
429 L</guestfs_read_lines> to get the file as a list of lines which
430 you can iterate over.
434 We support SELinux guests. To ensure that labeling happens correctly
435 in SELinux guests, you need to enable SELinux and load the guest's
442 Before launching, do:
444 guestfs_set_selinux (g, 1);
448 After mounting the guest's filesystem(s), load the policy. This
449 is best done by running the L<load_policy(8)> command in the
452 guestfs_sh (g, "/usr/sbin/load_policy");
454 (Older versions of C<load_policy> require you to specify the
455 name of the policy file).
459 Optionally, set the security context for the API. The correct
460 security context to use can only be known by inspecting the
461 guest. As an example:
463 guestfs_setcon (g, "unconfined_u:unconfined_r:unconfined_t:s0");
467 This will work for running commands and editing existing files.
469 When new files are created, you may need to label them explicitly,
470 for example by running the external command
471 C<restorecon pathname>.
475 Certain calls are affected by the current file mode creation mask (the
476 "umask"). In particular ones which create files or directories, such
477 as L</guestfs_touch>, L</guestfs_mknod> or L</guestfs_mkdir>. This
478 affects either the default mode that the file is created with or
479 modifies the mode that you supply.
481 The default umask is C<022>, so files are created with modes such as
482 C<0644> and directories with C<0755>.
484 There are two ways to avoid being affected by umask. Either set umask
485 to 0 (call C<guestfs_umask (g, 0)> early after launching). Or call
486 L</guestfs_chmod> after creating each file or directory.
488 For more information about umask, see L<umask(2)>.
490 =head2 ENCRYPTED DISKS
492 Libguestfs allows you to access Linux guests which have been
493 encrypted using whole disk encryption that conforms to the
494 Linux Unified Key Setup (LUKS) standard. This includes
495 nearly all whole disk encryption systems used by modern
498 Use L</guestfs_vfs_type> to identify LUKS-encrypted block
499 devices (it returns the string C<crypto_LUKS>).
501 Then open these devices by calling L</guestfs_luks_open>.
502 Obviously you will require the passphrase!
504 Opening a LUKS device creates a new device mapper device
505 called C</dev/mapper/mapname> (where C<mapname> is the
506 string you supply to L</guestfs_luks_open>).
507 Reads and writes to this mapper device are decrypted from and
508 encrypted to the underlying block device respectively.
510 LVM volume groups on the device can be made visible by calling
511 L</guestfs_vgscan> followed by L</guestfs_vg_activate_all>.
512 The logical volume(s) can now be mounted in the usual way.
514 Use the reverse process to close a LUKS device. Unmount
515 any logical volumes on it, deactivate the volume groups
516 by caling C<guestfs_vg_activate (g, 0, ["/dev/VG"])>.
517 Then close the mapper device by calling
518 L</guestfs_luks_close> on the C</dev/mapper/mapname>
519 device (I<not> the underlying encrypted block device).
523 Libguestfs has APIs for inspecting an unknown disk image to find out
524 if it contains operating systems. (These APIs used to be in a
525 separate Perl-only library called L<Sys::Guestfs::Lib(3)> but since
526 version 1.5.3 the most frequently used part of this library has been
527 rewritten in C and moved into the core code).
529 Add all disks belonging to the unknown virtual machine and call
530 L</guestfs_launch> in the usual way.
532 Then call L</guestfs_inspect_os>. This function uses other libguestfs
533 calls and certain heuristics, and returns a list of operating systems
534 that were found. An empty list means none were found. A single
535 element is the root filesystem of the operating system. For dual- or
536 multi-boot guests, multiple roots can be returned, each one
537 corresponding to a separate operating system. (Multi-boot virtual
538 machines are extremely rare in the world of virtualization, but since
539 this scenario can happen, we have built libguestfs to deal with it.)
541 For each root, you can then call various C<guestfs_inspect_get_*>
542 functions to get additional details about that operating system. For
543 example, call L</guestfs_inspect_get_type> to return the string
544 C<windows> or C<linux> for Windows and Linux-based operating systems
547 Un*x-like and Linux-based operating systems usually consist of several
548 filesystems which are mounted at boot time (for example, a separate
549 boot partition mounted on C</boot>). The inspection rules are able to
550 detect how filesystems correspond to mount points. Call
551 C<guestfs_inspect_get_mountpoints> to get this mapping. It might
552 return a hash table like this example:
555 / => /dev/vg_guest/lv_root
556 /usr => /dev/vg_guest/lv_usr
558 The caller can then make calls to L</guestfs_mount_options> to
559 mount the filesystems as suggested.
561 Be careful to mount filesystems in the right order (eg. C</> before
562 C</usr>). Sorting the keys of the hash by length, shortest first,
565 Inspection currently only works for some common operating systems.
566 Contributors are welcome to send patches for other operating systems
567 that we currently cannot detect.
569 Encrypted disks must be opened before inspection. See
570 L</ENCRYPTED DISKS> for more details. The L</guestfs_inspect_os>
571 function just ignores any encrypted devices.
573 A note on the implementation: The call L</guestfs_inspect_os> performs
574 inspection and caches the results in the guest handle. Subsequent
575 calls to C<guestfs_inspect_get_*> return this cached information, but
576 I<do not> re-read the disks. If you change the content of the guest
577 disks, you can redo inspection by calling L</guestfs_inspect_os>
580 =head2 SPECIAL CONSIDERATIONS FOR WINDOWS GUESTS
582 Libguestfs can mount NTFS partitions. It does this using the
583 L<http://www.ntfs-3g.org/> driver.
585 DOS and Windows still use drive letters, and the filesystems are
586 always treated as case insensitive by Windows itself, and therefore
587 you might find a Windows configuration file referring to a path like
588 C<c:\windows\system32>. When the filesystem is mounted in libguestfs,
589 that directory might be referred to as C</WINDOWS/System32>.
591 Drive letter mappings are outside the scope of libguestfs. You have
592 to use libguestfs to read the appropriate Windows Registry and
593 configuration files, to determine yourself how drives are mapped (see
594 also L<hivex(3)> and L<virt-inspector(1)>).
596 Replacing backslash characters with forward slash characters is also
597 outside the scope of libguestfs, but something that you can easily do.
599 Where we can help is in resolving the case insensitivity of paths.
600 For this, call L</guestfs_case_sensitive_path>.
602 Libguestfs also provides some help for decoding Windows Registry
603 "hive" files, through the library C<hivex> which is part of the
604 libguestfs project although ships as a separate tarball. You have to
605 locate and download the hive file(s) yourself, and then pass them to
606 C<hivex> functions. See also the programs L<hivexml(1)>,
607 L<hivexsh(1)>, L<hivexregedit(1)> and L<virt-win-reg(1)> for more help
610 =head2 USING LIBGUESTFS WITH OTHER PROGRAMMING LANGUAGES
612 Although we don't want to discourage you from using the C API, we will
613 mention here that the same API is also available in other languages.
615 The API is broadly identical in all supported languages. This means
616 that the C call C<guestfs_mount(g,path)> is
617 C<$g-E<gt>mount($path)> in Perl, C<g.mount(path)> in Python,
618 and C<Guestfs.mount g path> in OCaml. In other words, a
619 straightforward, predictable isomorphism between each language.
621 Error messages are automatically transformed
622 into exceptions if the language supports it.
624 We don't try to "object orientify" parts of the API in OO languages,
625 although contributors are welcome to write higher level APIs above
626 what we provide in their favourite languages if they wish.
632 You can use the I<guestfs.h> header file from C++ programs. The C++
633 API is identical to the C API. C++ classes and exceptions are not
638 The C# bindings are highly experimental. Please read the warnings
639 at the top of C<csharp/Libguestfs.cs>.
643 This is the only language binding that is working but incomplete.
644 Only calls which return simple integers have been bound in Haskell,
645 and we are looking for help to complete this binding.
649 Full documentation is contained in the Javadoc which is distributed
654 For documentation see the file C<guestfs.mli>.
658 For documentation see L<Sys::Guestfs(3)>.
662 For documentation see C<README-PHP> supplied with libguestfs
663 sources or in the php-libguestfs package for your distribution.
665 The PHP binding only works correctly on 64 bit machines.
669 For documentation do:
677 Use the Guestfs module. There is no Ruby-specific documentation, but
678 you can find examples written in Ruby in the libguestfs source.
680 =item B<shell scripts>
682 For documentation see L<guestfish(1)>.
686 =head2 LIBGUESTFS GOTCHAS
688 L<http://en.wikipedia.org/wiki/Gotcha_(programming)>: "A feature of a
689 system [...] that works in the way it is documented but is
690 counterintuitive and almost invites mistakes."
692 Since we developed libguestfs and the associated tools, there are
693 several things we would have designed differently, but are now stuck
694 with for backwards compatibility or other reasons. If there is ever a
695 libguestfs 2.0 release, you can expect these to change. Beware of
700 =item Autosync / forgetting to sync.
702 When modifying a filesystem from C or another language, you B<must>
703 unmount all filesystems and call L</guestfs_sync> explicitly before
704 you close the libguestfs handle. You can also call:
706 guestfs_set_autosync (g, 1);
708 to have the unmount/sync done automatically for you when the handle 'g'
709 is closed. (This feature is called "autosync", L</guestfs_set_autosync>
712 If you forget to do this, then it is entirely possible that your
713 changes won't be written out, or will be partially written, or (very
714 rarely) that you'll get disk corruption.
716 Note that in L<guestfish(3)> autosync is the default. So quick and
717 dirty guestfish scripts that forget to sync will work just fine, which
718 can make this very puzzling if you are trying to debug a problem.
720 Update: Autosync is enabled by default for all API users starting from
723 =item Mount option C<-o sync> should not be the default.
725 If you use L</guestfs_mount>, then C<-o sync,noatime> are added
726 implicitly. However C<-o sync> does not add any reliability benefit,
727 but does have a very large performance impact.
729 The work around is to use L</guestfs_mount_options> and set the mount
730 options that you actually want to use.
732 =item Read-only should be the default.
734 In L<guestfish(3)>, I<--ro> should be the default, and you should
735 have to specify I<--rw> if you want to make changes to the image.
737 This would reduce the potential to corrupt live VM images.
739 Note that many filesystems change the disk when you just mount and
740 unmount, even if you didn't perform any writes. You need to use
741 L</guestfs_add_drive_ro> to guarantee that the disk is not changed.
743 =item guestfish command line is hard to use.
745 C<guestfish disk.img> doesn't do what people expect (open C<disk.img>
746 for examination). It tries to run a guestfish command C<disk.img>
747 which doesn't exist, so it fails. In earlier versions of guestfish
748 the error message was also unintuitive, but we have corrected this
749 since. Like the Bourne shell, we should have used C<guestfish -c
750 command> to run commands.
752 =item guestfish megabyte modifiers don't work right on all commands
754 In recent guestfish you can use C<1M> to mean 1 megabyte (and
755 similarly for other modifiers). What guestfish actually does is to
756 multiply the number part by the modifier part and pass the result to
757 the C API. However this doesn't work for a few APIs which aren't
758 expecting bytes, but are already expecting some other unit
761 The most common is L</guestfs_lvcreate>. The guestfish command:
765 does not do what you might expect. Instead because
766 L</guestfs_lvcreate> is already expecting megabytes, this tries to
767 create a 100 I<terabyte> (100 megabytes * megabytes) logical volume.
768 The error message you get from this is also a little obscure.
770 This could be fixed in the generator by specially marking parameters
771 and return values which take bytes or other units.
773 =item Library should return errno with error messages.
775 It would be a nice-to-have to be able to get the original value of
776 'errno' from inside the appliance along error paths (where set).
777 Currently L<guestmount(1)> goes through hoops to try to reverse the
778 error message string into an errno, see the function error() in
781 In libguestfs 1.5.4, the protocol was changed so that the
782 Linux errno is sent back from the daemon.
784 =item Ambiguity between devices and paths
786 There is a subtle ambiguity in the API between a device name
787 (eg. C</dev/sdb2>) and a similar pathname. A file might just happen
788 to be called C<sdb2> in the directory C</dev> (consider some non-Unix
791 In the current API we usually resolve this ambiguity by having two
792 separate calls, for example L</guestfs_checksum> and
793 L</guestfs_checksum_device>. Some API calls are ambiguous and
794 (incorrectly) resolve the problem by detecting if the path supplied
795 begins with C</dev/>.
797 To avoid both the ambiguity and the need to duplicate some calls, we
798 could make paths/devices into structured names. One way to do this
799 would be to use a notation like grub (C<hd(0,0)>), although nobody
800 really likes this aspect of grub. Another way would be to use a
801 structured type, equivalent to this OCaml type:
803 type path = Path of string | Device of int | Partition of int * int
805 which would allow you to pass arguments like:
808 Device 1 (* /dev/sdb, or perhaps /dev/sda *)
809 Partition (1, 2) (* /dev/sdb2 (or is it /dev/sda2 or /dev/sdb3?) *)
810 Path "/dev/sdb2" (* not a device *)
812 As you can see there are still problems to resolve even with this
813 representation. Also consider how it might work in guestfish.
817 =head2 PROTOCOL LIMITS
819 Internally libguestfs uses a message-based protocol to pass API calls
820 and their responses to and from a small "appliance" (see L</INTERNALS>
821 for plenty more detail about this). The maximum message size used by
822 the protocol is slightly less than 4 MB. For some API calls you may
823 need to be aware of this limit. The API calls which may be affected
824 are individually documented, with a link back to this section of the
827 A simple call such as L</guestfs_cat> returns its result (the file
828 data) in a simple string. Because this string is at some point
829 internally encoded as a message, the maximum size that it can return
830 is slightly under 4 MB. If the requested file is larger than this
831 then you will get an error.
833 In order to transfer large files into and out of the guest filesystem,
834 you need to use particular calls that support this. The sections
835 L</UPLOADING> and L</DOWNLOADING> document how to do this.
837 You might also consider mounting the disk image using our FUSE
838 filesystem support (L<guestmount(1)>).
840 =head2 KEYS AND PASSPHRASES
842 Certain libguestfs calls take a parameter that contains sensitive key
843 material, passed in as a C string.
845 In the future we would hope to change the libguestfs implementation so
846 that keys are L<mlock(2)>-ed into physical RAM, and thus can never end
847 up in swap. However this is I<not> done at the moment, because of the
848 complexity of such an implementation.
850 Therefore you should be aware that any key parameter you pass to
851 libguestfs might end up being written out to the swap partition. If
852 this is a concern, scrub the swap partition or don't use libguestfs on
855 =head1 CONNECTION MANAGEMENT
859 C<guestfs_h> is the opaque type representing a connection handle.
860 Create a handle by calling L</guestfs_create>. Call L</guestfs_close>
861 to free the handle and release all resources used.
863 For information on using multiple handles and threads, see the section
864 L</MULTIPLE HANDLES AND MULTIPLE THREADS> below.
866 =head2 guestfs_create
868 guestfs_h *guestfs_create (void);
870 Create a connection handle.
872 You have to call L</guestfs_add_drive_opts> (or one of the equivalent
873 calls) on the handle at least once.
875 This function returns a non-NULL pointer to a handle on success or
878 After configuring the handle, you have to call L</guestfs_launch>.
880 You may also want to configure error handling for the handle. See
881 L</ERROR HANDLING> section below.
885 void guestfs_close (guestfs_h *g);
887 This closes the connection handle and frees up all resources used.
889 =head1 ERROR HANDLING
891 API functions can return errors. For example, almost all functions
892 that return C<int> will return C<-1> to indicate an error. You can
893 get additional information on
894 errors by calling L</guestfs_last_error> and/or by setting up an error
895 handler with L</guestfs_set_error_handler>.
897 When the handle is created, a default error handler is installed which
898 prints the error message string to C<stderr>. For small short-running
899 command line programs it is sufficient to do:
901 if (guestfs_launch (g) == -1)
904 since the default error handler will ensure that an error message has
905 been printed to C<stderr> before the program exits.
907 For other programs the caller will almost certainly want to install an
908 alternate error handler or do error handling in-line like this:
910 g = guestfs_create ();
912 /* This disables the default behaviour of printing errors
914 guestfs_set_error_handler (g, NULL, NULL);
916 if (guestfs_launch (g) == -1) {
917 /* Examine the error message and print it etc. */
918 char *msg = guestfs_last_error (g);
919 fprintf (stderr, "%s\n", msg);
923 Out of memory errors are handled differently. The default action is
924 to call L<abort(3)>. If this is undesirable, then you can set a
925 handler using L</guestfs_set_out_of_memory_handler>.
927 L</guestfs_create> returns C<NULL> if the handle cannot be created,
928 and because there is no handle if this happens there is no way to get
929 additional error information. However L</guestfs_create> is supposed
930 to be a lightweight operation which can only fail because of
931 insufficient memory (it returns NULL in this case).
933 =head2 guestfs_last_error
935 const char *guestfs_last_error (guestfs_h *g);
937 This returns the last error message that happened on C<g>. If
938 there has not been an error since the handle was created, then this
941 The lifetime of the returned string is until the next error occurs, or
942 L</guestfs_close> is called.
944 =head2 guestfs_set_error_handler
946 typedef void (*guestfs_error_handler_cb) (guestfs_h *g,
949 void guestfs_set_error_handler (guestfs_h *g,
950 guestfs_error_handler_cb cb,
953 The callback C<cb> will be called if there is an error. The
954 parameters passed to the callback are an opaque data pointer and the
955 error message string.
957 Note that the message string C<msg> is freed as soon as the callback
958 function returns, so if you want to stash it somewhere you must make
961 The default handler prints messages on C<stderr>.
963 If you set C<cb> to C<NULL> then I<no> handler is called.
965 =head2 guestfs_get_error_handler
967 guestfs_error_handler_cb guestfs_get_error_handler (guestfs_h *g,
970 Returns the current error handler callback.
972 =head2 guestfs_set_out_of_memory_handler
974 typedef void (*guestfs_abort_cb) (void);
975 int guestfs_set_out_of_memory_handler (guestfs_h *g,
978 The callback C<cb> will be called if there is an out of memory
979 situation. I<Note this callback must not return>.
981 The default is to call L<abort(3)>.
983 You cannot set C<cb> to C<NULL>. You can't ignore out of memory
986 =head2 guestfs_get_out_of_memory_handler
988 guestfs_abort_fn guestfs_get_out_of_memory_handler (guestfs_h *g);
990 This returns the current out of memory handler.
994 Libguestfs needs a kernel and initrd.img, which it finds by looking
995 along an internal path.
997 By default it looks for these in the directory C<$libdir/guestfs>
998 (eg. C</usr/local/lib/guestfs> or C</usr/lib64/guestfs>).
1000 Use L</guestfs_set_path> or set the environment variable
1001 L</LIBGUESTFS_PATH> to change the directories that libguestfs will
1002 search in. The value is a colon-separated list of paths. The current
1003 directory is I<not> searched unless the path contains an empty element
1004 or C<.>. For example C<LIBGUESTFS_PATH=:/usr/lib/guestfs> would
1005 search the current directory and then C</usr/lib/guestfs>.
1007 =head1 HIGH-LEVEL API ACTIONS
1009 =head2 ABI GUARANTEE
1011 We guarantee the libguestfs ABI (binary interface), for public,
1012 high-level actions as outlined in this section. Although we will
1013 deprecate some actions, for example if they get replaced by newer
1014 calls, we will keep the old actions forever. This allows you the
1015 developer to program in confidence against the libguestfs API.
1025 =head2 GROUPS OF FUNCTIONALITY IN THE APPLIANCE
1027 Using L</guestfs_available> you can test availability of
1028 the following groups of functions. This test queries the
1029 appliance to see if the appliance you are currently using
1030 supports the functionality.
1034 =head2 GUESTFISH supported COMMAND
1036 In L<guestfish(3)> there is a handy interactive command
1037 C<supported> which prints out the available groups and
1038 whether they are supported by this build of libguestfs.
1039 Note however that you have to do C<run> first.
1041 =head2 SINGLE CALLS AT COMPILE TIME
1043 Since version 1.5.8, C<E<lt>guestfs.hE<gt>> defines symbols
1044 for each C API function, such as:
1046 #define LIBGUESTFS_HAVE_DD 1
1048 if L</guestfs_dd> is available.
1050 Before version 1.5.8, if you needed to test whether a single
1051 libguestfs function is available at compile time, we recommended using
1052 build tools such as autoconf or cmake. For example in autotools you
1055 AC_CHECK_LIB([guestfs],[guestfs_create])
1056 AC_CHECK_FUNCS([guestfs_dd])
1058 which would result in C<HAVE_GUESTFS_DD> being either defined
1059 or not defined in your program.
1061 =head2 SINGLE CALLS AT RUN TIME
1063 Testing at compile time doesn't guarantee that a function really
1064 exists in the library. The reason is that you might be dynamically
1065 linked against a previous I<libguestfs.so> (dynamic library)
1066 which doesn't have the call. This situation unfortunately results
1067 in a segmentation fault, which is a shortcoming of the C dynamic
1068 linking system itself.
1070 You can use L<dlopen(3)> to test if a function is available
1071 at run time, as in this example program (note that you still
1072 need the compile time check as well):
1078 #include <guestfs.h>
1082 #ifdef LIBGUESTFS_HAVE_DD
1086 /* Test if the function guestfs_dd is really available. */
1087 dl = dlopen (NULL, RTLD_LAZY);
1089 fprintf (stderr, "dlopen: %s\n", dlerror ());
1090 exit (EXIT_FAILURE);
1092 has_function = dlsym (dl, "guestfs_dd") != NULL;
1096 printf ("this libguestfs.so does NOT have guestfs_dd function\n");
1098 printf ("this libguestfs.so has guestfs_dd function\n");
1099 /* Now it's safe to call
1100 guestfs_dd (g, "foo", "bar");
1104 printf ("guestfs_dd function was not found at compile time\n");
1108 You may think the above is an awful lot of hassle, and it is.
1109 There are other ways outside of the C linking system to ensure
1110 that this kind of incompatibility never arises, such as using
1113 Requires: libguestfs >= 1.0.80
1115 =head1 CALLS WITH OPTIONAL ARGUMENTS
1117 A recent feature of the API is the introduction of calls which take
1118 optional arguments. In C these are declared 3 ways. The main way is
1119 as a call which takes variable arguments (ie. C<...>), as in this
1122 int guestfs_add_drive_opts (guestfs_h *g, const char *filename, ...);
1124 Call this with a list of optional arguments, terminated by C<-1>.
1125 So to call with no optional arguments specified:
1127 guestfs_add_drive_opts (g, filename, -1);
1129 With a single optional argument:
1131 guestfs_add_drive_opts (g, filename,
1132 GUESTFS_ADD_DRIVE_OPTS_FORMAT, "qcow2",
1137 guestfs_add_drive_opts (g, filename,
1138 GUESTFS_ADD_DRIVE_OPTS_FORMAT, "qcow2",
1139 GUESTFS_ADD_DRIVE_OPTS_READONLY, 1,
1142 and so forth. Don't forget the terminating C<-1> otherwise
1143 Bad Things will happen!
1145 =head2 USING va_list FOR OPTIONAL ARGUMENTS
1147 The second variant has the same name with the suffix C<_va>, which
1148 works the same way but takes a C<va_list>. See the C manual for
1149 details. For the example function, this is declared:
1151 int guestfs_add_drive_opts_va (guestfs_h *g, const char *filename,
1154 =head2 CONSTRUCTING OPTIONAL ARGUMENTS
1156 The third variant is useful where you need to construct these
1157 calls. You pass in a structure where you fill in the optional
1158 fields. The structure has a bitmask as the first element which
1159 you must set to indicate which fields you have filled in. For
1160 our example function the structure and call are declared:
1162 struct guestfs_add_drive_opts_argv {
1168 int guestfs_add_drive_opts_argv (guestfs_h *g, const char *filename,
1169 const struct guestfs_add_drive_opts_argv *optargs);
1171 You could call it like this:
1173 struct guestfs_add_drive_opts_argv optargs = {
1174 .bitmask = GUESTFS_ADD_DRIVE_OPTS_READONLY_BITMASK |
1175 GUESTFS_ADD_DRIVE_OPTS_FORMAT_BITMASK,
1180 guestfs_add_drive_opts_argv (g, filename, &optargs);
1188 The C<_BITMASK> suffix on each option name when specifying the
1193 You do not need to fill in all fields of the structure.
1197 There must be a one-to-one correspondence between fields of the
1198 structure that are filled in, and bits set in the bitmask.
1202 =head2 OPTIONAL ARGUMENTS IN OTHER LANGUAGES
1204 In other languages, optional arguments are expressed in the
1205 way that is natural for that language. We refer you to the
1206 language-specific documentation for more details on that.
1208 For guestfish, see L<guestfish(1)/OPTIONAL ARGUMENTS>.
1212 <!-- old anchor for the next section -->
1213 <a name="state_machine_and_low_level_event_api"/>
1219 Internally, libguestfs is implemented by running an appliance (a
1220 special type of small virtual machine) using L<qemu(1)>. Qemu runs as
1221 a child process of the main program.
1227 | | child process / appliance
1228 | | __________________________
1230 +-------------------+ RPC | +-----------------+ |
1231 | libguestfs <--------------------> guestfsd | |
1232 | | | +-----------------+ |
1233 \___________________/ | | Linux kernel | |
1234 | +--^--------------+ |
1235 \_________|________________/
1243 The library, linked to the main program, creates the child process and
1244 hence the appliance in the L</guestfs_launch> function.
1246 Inside the appliance is a Linux kernel and a complete stack of
1247 userspace tools (such as LVM and ext2 programs) and a small
1248 controlling daemon called L</guestfsd>. The library talks to
1249 L</guestfsd> using remote procedure calls (RPC). There is a mostly
1250 one-to-one correspondence between libguestfs API calls and RPC calls
1251 to the daemon. Lastly the disk image(s) are attached to the qemu
1252 process which translates device access by the appliance's Linux kernel
1253 into accesses to the image.
1255 A common misunderstanding is that the appliance "is" the virtual
1256 machine. Although the disk image you are attached to might also be
1257 used by some virtual machine, libguestfs doesn't know or care about
1258 this. (But you will care if both libguestfs's qemu process and your
1259 virtual machine are trying to update the disk image at the same time,
1260 since these usually results in massive disk corruption).
1262 =head1 STATE MACHINE
1264 libguestfs uses a state machine to model the child process:
1275 / | \ \ guestfs_launch
1286 \______/ <------ \________/
1288 The normal transitions are (1) CONFIG (when the handle is created, but
1289 there is no child process), (2) LAUNCHING (when the child process is
1290 booting up), (3) alternating between READY and BUSY as commands are
1291 issued to, and carried out by, the child process.
1293 The guest may be killed by L</guestfs_kill_subprocess>, or may die
1294 asynchronously at any time (eg. due to some internal error), and that
1295 causes the state to transition back to CONFIG.
1297 Configuration commands for qemu such as L</guestfs_add_drive> can only
1298 be issued when in the CONFIG state.
1300 The API offers one call that goes from CONFIG through LAUNCHING to
1301 READY. L</guestfs_launch> blocks until the child process is READY to
1302 accept commands (or until some failure or timeout).
1303 L</guestfs_launch> internally moves the state from CONFIG to LAUNCHING
1304 while it is running.
1306 API actions such as L</guestfs_mount> can only be issued when in the
1307 READY state. These API calls block waiting for the command to be
1308 carried out (ie. the state to transition to BUSY and then back to
1309 READY). There are no non-blocking versions, and no way to issue more
1310 than one command per handle at the same time.
1312 Finally, the child process sends asynchronous messages back to the
1313 main program, such as kernel log messages. You can register a
1314 callback to receive these messages.
1316 =head2 SETTING CALLBACKS TO HANDLE EVENTS
1318 The child process generates events in some situations. Current events
1319 include: receiving a log message, the child process exits.
1321 Use the C<guestfs_set_*_callback> functions to set a callback for
1322 different types of events.
1324 Only I<one callback of each type> can be registered for each handle.
1325 Calling C<guestfs_set_*_callback> again overwrites the previous
1326 callback of that type. Cancel all callbacks of this type by calling
1327 this function with C<cb> set to C<NULL>.
1329 =head2 guestfs_set_log_message_callback
1331 typedef void (*guestfs_log_message_cb) (guestfs_h *g, void *opaque,
1332 char *buf, int len);
1333 void guestfs_set_log_message_callback (guestfs_h *g,
1334 guestfs_log_message_cb cb,
1337 The callback function C<cb> will be called whenever qemu or the guest
1338 writes anything to the console.
1340 Use this function to capture kernel messages and similar.
1342 Normally there is no log message handler, and log messages are just
1345 =head2 guestfs_set_subprocess_quit_callback
1347 typedef void (*guestfs_subprocess_quit_cb) (guestfs_h *g, void *opaque);
1348 void guestfs_set_subprocess_quit_callback (guestfs_h *g,
1349 guestfs_subprocess_quit_cb cb,
1352 The callback function C<cb> will be called when the child process
1353 quits, either asynchronously or if killed by
1354 L</guestfs_kill_subprocess>. (This corresponds to a transition from
1355 any state to the CONFIG state).
1357 =head2 guestfs_set_launch_done_callback
1359 typedef void (*guestfs_launch_done_cb) (guestfs_h *g, void *opaque);
1360 void guestfs_set_launch_done_callback (guestfs_h *g,
1361 guestfs_launch_done_cb cb,
1364 The callback function C<cb> will be called when the child process
1365 becomes ready first time after it has been launched. (This
1366 corresponds to a transition from LAUNCHING to the READY state).
1368 =head2 guestfs_set_close_callback
1370 typedef void (*guestfs_close_cb) (guestfs_h *g, void *opaque);
1371 void guestfs_set_close_callback (guestfs_h *g,
1372 guestfs_close_cb cb,
1375 The callback function C<cb> will be called while the handle
1376 is being closed (synchronously from L</guestfs_close>).
1378 Note that libguestfs installs an L<atexit(3)> handler to try to
1379 clean up handles that are open when the program exits. This
1380 means that this callback might be called indirectly from
1381 L<exit(3)>, which can cause unexpected problems in higher-level
1382 languages (eg. if your HLL interpreter has already been cleaned
1383 up by the time this is called, and if your callback then jumps
1384 into some HLL function).
1386 =head2 guestfs_set_progress_callback
1388 typedef void (*guestfs_progress_cb) (guestfs_h *g, void *opaque,
1389 int proc_nr, int serial,
1390 uint64_t position, uint64_t total);
1391 void guestfs_set_progress_callback (guestfs_h *g,
1392 guestfs_progress_cb cb,
1395 Some long-running operations can generate progress messages. If
1396 this callback is registered, then it will be called each time a
1397 progress message is generated (usually two seconds after the
1398 operation started, and three times per second thereafter until
1399 it completes, although the frequency may change in future versions).
1401 The callback receives two numbers: C<position> and C<total>.
1402 The units of C<total> are not defined, although for some
1403 operations C<total> may relate in some way to the amount of
1404 data to be transferred (eg. in bytes or megabytes), and
1405 C<position> may be the portion which has been transferred.
1407 The only defined and stable parts of the API are:
1413 The callback can display to the user some type of progress bar or
1414 indicator which shows the ratio of C<position>:C<total>.
1418 0 E<lt>= C<position> E<lt>= C<total>
1422 If any progress notification is sent during a call, then a final
1423 progress notification is always sent when C<position> = C<total>.
1425 This is to simplify caller code, so callers can easily set the
1426 progress indicator to "100%" at the end of the operation, without
1427 requiring special code to detect this case.
1431 The callback also receives the procedure number and serial number of
1432 the call. These are only useful for debugging protocol issues, and
1433 the callback can normally ignore them. The callback may want to
1434 print these numbers in error messages or debugging messages.
1436 =head1 PRIVATE DATA AREA
1438 You can attach named pieces of private data to the libguestfs handle,
1439 and fetch them by name for the lifetime of the handle. This is called
1440 the private data area and is only available from the C API.
1442 To attach a named piece of data, use the following call:
1444 void guestfs_set_private (guestfs_h *g, const char *key, void *data);
1446 C<key> is the name to associate with this data, and C<data> is an
1447 arbitrary pointer (which can be C<NULL>). Any previous item with the
1448 same name is overwritten.
1450 You can use any C<key> you want, but names beginning with an
1451 underscore character are reserved for internal libguestfs purposes
1452 (for implementing language bindings). It is recommended to prefix the
1453 name with some unique string to avoid collisions with other users.
1455 To retrieve the pointer, use:
1457 void *guestfs_get_private (guestfs_h *g, const char *key);
1459 This function returns C<NULL> if either no data is found associated
1460 with C<key>, or if the user previously set the C<key>'s C<data>
1463 Libguestfs does not try to look at or interpret the C<data> pointer in
1464 any way. As far as libguestfs is concerned, it need not be a valid
1465 pointer at all. In particular, libguestfs does I<not> try to free the
1466 data when the handle is closed. If the data must be freed, then the
1467 caller must either free it before calling L</guestfs_close> or must
1468 set up a close callback to do it (see L</guestfs_set_close_callback>,
1469 and note that only one callback can be registered for a handle).
1471 The private data area is implemented using a hash table, and should be
1472 reasonably efficient for moderate numbers of keys.
1474 =head1 BLOCK DEVICE NAMING
1476 In the kernel there is now quite a profusion of schemata for naming
1477 block devices (in this context, by I<block device> I mean a physical
1478 or virtual hard drive). The original Linux IDE driver used names
1479 starting with C</dev/hd*>. SCSI devices have historically used a
1480 different naming scheme, C</dev/sd*>. When the Linux kernel I<libata>
1481 driver became a popular replacement for the old IDE driver
1482 (particularly for SATA devices) those devices also used the
1483 C</dev/sd*> scheme. Additionally we now have virtual machines with
1484 paravirtualized drivers. This has created several different naming
1485 systems, such as C</dev/vd*> for virtio disks and C</dev/xvd*> for Xen
1488 As discussed above, libguestfs uses a qemu appliance running an
1489 embedded Linux kernel to access block devices. We can run a variety
1490 of appliances based on a variety of Linux kernels.
1492 This causes a problem for libguestfs because many API calls use device
1493 or partition names. Working scripts and the recipe (example) scripts
1494 that we make available over the internet could fail if the naming
1497 Therefore libguestfs defines C</dev/sd*> as the I<standard naming
1498 scheme>. Internally C</dev/sd*> names are translated, if necessary,
1499 to other names as required. For example, under RHEL 5 which uses the
1500 C</dev/hd*> scheme, any device parameter C</dev/sda2> is translated to
1501 C</dev/hda2> transparently.
1503 Note that this I<only> applies to parameters. The
1504 L</guestfs_list_devices>, L</guestfs_list_partitions> and similar calls
1505 return the true names of the devices and partitions as known to the
1508 =head2 ALGORITHM FOR BLOCK DEVICE NAME TRANSLATION
1510 Usually this translation is transparent. However in some (very rare)
1511 cases you may need to know the exact algorithm. Such cases include
1512 where you use L</guestfs_config> to add a mixture of virtio and IDE
1513 devices to the qemu-based appliance, so have a mixture of C</dev/sd*>
1514 and C</dev/vd*> devices.
1516 The algorithm is applied only to I<parameters> which are known to be
1517 either device or partition names. Return values from functions such
1518 as L</guestfs_list_devices> are never changed.
1524 Is the string a parameter which is a device or partition name?
1528 Does the string begin with C</dev/sd>?
1532 Does the named device exist? If so, we use that device.
1533 However if I<not> then we continue with this algorithm.
1537 Replace initial C</dev/sd> string with C</dev/hd>.
1539 For example, change C</dev/sda2> to C</dev/hda2>.
1541 If that named device exists, use it. If not, continue.
1545 Replace initial C</dev/sd> string with C</dev/vd>.
1547 If that named device exists, use it. If not, return an error.
1551 =head2 PORTABILITY CONCERNS
1553 Although the standard naming scheme and automatic translation is
1554 useful for simple programs and guestfish scripts, for larger programs
1555 it is best not to rely on this mechanism.
1557 Where possible for maximum future portability programs using
1558 libguestfs should use these future-proof techniques:
1564 Use L</guestfs_list_devices> or L</guestfs_list_partitions> to list
1565 actual device names, and then use those names directly.
1567 Since those device names exist by definition, they will never be
1572 Use higher level ways to identify filesystems, such as LVM names,
1573 UUIDs and filesystem labels.
1579 =head2 COMMUNICATION PROTOCOL
1581 Don't rely on using this protocol directly. This section documents
1582 how it currently works, but it may change at any time.
1584 The protocol used to talk between the library and the daemon running
1585 inside the qemu virtual machine is a simple RPC mechanism built on top
1586 of XDR (RFC 1014, RFC 1832, RFC 4506).
1588 The detailed format of structures is in C<src/guestfs_protocol.x>
1589 (note: this file is automatically generated).
1591 There are two broad cases, ordinary functions that don't have any
1592 C<FileIn> and C<FileOut> parameters, which are handled with very
1593 simple request/reply messages. Then there are functions that have any
1594 C<FileIn> or C<FileOut> parameters, which use the same request and
1595 reply messages, but they may also be followed by files sent using a
1598 =head3 ORDINARY FUNCTIONS (NO FILEIN/FILEOUT PARAMS)
1600 For ordinary functions, the request message is:
1602 total length (header + arguments,
1603 but not including the length word itself)
1604 struct guestfs_message_header (encoded as XDR)
1605 struct guestfs_<foo>_args (encoded as XDR)
1607 The total length field allows the daemon to allocate a fixed size
1608 buffer into which it slurps the rest of the message. As a result, the
1609 total length is limited to C<GUESTFS_MESSAGE_MAX> bytes (currently
1610 4MB), which means the effective size of any request is limited to
1611 somewhere under this size.
1613 Note also that many functions don't take any arguments, in which case
1614 the C<guestfs_I<foo>_args> is completely omitted.
1616 The header contains the procedure number (C<guestfs_proc>) which is
1617 how the receiver knows what type of args structure to expect, or none
1620 The reply message for ordinary functions is:
1622 total length (header + ret,
1623 but not including the length word itself)
1624 struct guestfs_message_header (encoded as XDR)
1625 struct guestfs_<foo>_ret (encoded as XDR)
1627 As above the C<guestfs_I<foo>_ret> structure may be completely omitted
1628 for functions that return no formal return values.
1630 As above the total length of the reply is limited to
1631 C<GUESTFS_MESSAGE_MAX>.
1633 In the case of an error, a flag is set in the header, and the reply
1634 message is slightly changed:
1636 total length (header + error,
1637 but not including the length word itself)
1638 struct guestfs_message_header (encoded as XDR)
1639 struct guestfs_message_error (encoded as XDR)
1641 The C<guestfs_message_error> structure contains the error message as a
1644 =head3 FUNCTIONS THAT HAVE FILEIN PARAMETERS
1646 A C<FileIn> parameter indicates that we transfer a file I<into> the
1647 guest. The normal request message is sent (see above). However this
1648 is followed by a sequence of file chunks.
1650 total length (header + arguments,
1651 but not including the length word itself,
1652 and not including the chunks)
1653 struct guestfs_message_header (encoded as XDR)
1654 struct guestfs_<foo>_args (encoded as XDR)
1655 sequence of chunks for FileIn param #0
1656 sequence of chunks for FileIn param #1 etc.
1658 The "sequence of chunks" is:
1660 length of chunk (not including length word itself)
1661 struct guestfs_chunk (encoded as XDR)
1663 struct guestfs_chunk (encoded as XDR)
1666 struct guestfs_chunk (with data.data_len == 0)
1668 The final chunk has the C<data_len> field set to zero. Additionally a
1669 flag is set in the final chunk to indicate either successful
1670 completion or early cancellation.
1672 At time of writing there are no functions that have more than one
1673 FileIn parameter. However this is (theoretically) supported, by
1674 sending the sequence of chunks for each FileIn parameter one after
1675 another (from left to right).
1677 Both the library (sender) I<and> the daemon (receiver) may cancel the
1678 transfer. The library does this by sending a chunk with a special
1679 flag set to indicate cancellation. When the daemon sees this, it
1680 cancels the whole RPC, does I<not> send any reply, and goes back to
1681 reading the next request.
1683 The daemon may also cancel. It does this by writing a special word
1684 C<GUESTFS_CANCEL_FLAG> to the socket. The library listens for this
1685 during the transfer, and if it gets it, it will cancel the transfer
1686 (it sends a cancel chunk). The special word is chosen so that even if
1687 cancellation happens right at the end of the transfer (after the
1688 library has finished writing and has started listening for the reply),
1689 the "spurious" cancel flag will not be confused with the reply
1692 This protocol allows the transfer of arbitrary sized files (no 32 bit
1693 limit), and also files where the size is not known in advance
1694 (eg. from pipes or sockets). However the chunks are rather small
1695 (C<GUESTFS_MAX_CHUNK_SIZE>), so that neither the library nor the
1696 daemon need to keep much in memory.
1698 =head3 FUNCTIONS THAT HAVE FILEOUT PARAMETERS
1700 The protocol for FileOut parameters is exactly the same as for FileIn
1701 parameters, but with the roles of daemon and library reversed.
1703 total length (header + ret,
1704 but not including the length word itself,
1705 and not including the chunks)
1706 struct guestfs_message_header (encoded as XDR)
1707 struct guestfs_<foo>_ret (encoded as XDR)
1708 sequence of chunks for FileOut param #0
1709 sequence of chunks for FileOut param #1 etc.
1711 =head3 INITIAL MESSAGE
1713 When the daemon launches it sends an initial word
1714 (C<GUESTFS_LAUNCH_FLAG>) which indicates that the guest and daemon is
1715 alive. This is what L</guestfs_launch> waits for.
1717 =head3 PROGRESS NOTIFICATION MESSAGES
1719 The daemon may send progress notification messages at any time. These
1720 are distinguished by the normal length word being replaced by
1721 C<GUESTFS_PROGRESS_FLAG>, followed by a fixed size progress message.
1723 The library turns them into progress callbacks (see
1724 C<guestfs_set_progress_callback>) if there is a callback registered,
1725 or discards them if not.
1727 The daemon self-limits the frequency of progress messages it sends
1728 (see C<daemon/proto.c:notify_progress>). Not all calls generate
1731 =head1 MULTIPLE HANDLES AND MULTIPLE THREADS
1733 All high-level libguestfs actions are synchronous. If you want
1734 to use libguestfs asynchronously then you must create a thread.
1736 Only use the handle from a single thread. Either use the handle
1737 exclusively from one thread, or provide your own mutex so that two
1738 threads cannot issue calls on the same handle at the same time.
1740 See the graphical program guestfs-browser for one possible
1741 architecture for multithreaded programs using libvirt and libguestfs.
1743 =head1 QEMU WRAPPERS
1745 If you want to compile your own qemu, run qemu from a non-standard
1746 location, or pass extra arguments to qemu, then you can write a
1747 shell-script wrapper around qemu.
1749 There is one important rule to remember: you I<must C<exec qemu>> as
1750 the last command in the shell script (so that qemu replaces the shell
1751 and becomes the direct child of the libguestfs-using program). If you
1752 don't do this, then the qemu process won't be cleaned up correctly.
1754 Here is an example of a wrapper, where I have built my own copy of
1758 qemudir=/home/rjones/d/qemu
1759 exec $qemudir/x86_64-softmmu/qemu-system-x86_64 -L $qemudir/pc-bios "$@"
1761 Save this script as C</tmp/qemu.wrapper> (or wherever), C<chmod +x>,
1762 and then use it by setting the LIBGUESTFS_QEMU environment variable.
1765 LIBGUESTFS_QEMU=/tmp/qemu.wrapper guestfish
1767 Note that libguestfs also calls qemu with the -help and -version
1768 options in order to determine features.
1770 =head1 LIBGUESTFS VERSION NUMBERS
1772 Since April 2010, libguestfs has started to make separate development
1773 and stable releases, along with corresponding branches in our git
1774 repository. These separate releases can be identified by version
1777 even numbers for stable: 1.2.x, 1.4.x, ...
1778 .-------- odd numbers for development: 1.3.x, 1.5.x, ...
1784 | `-------- sub-version
1786 `------ always '1' because we don't change the ABI
1788 Thus "1.3.5" is the 5th update to the development branch "1.3".
1790 As time passes we cherry pick fixes from the development branch and
1791 backport those into the stable branch, the effect being that the
1792 stable branch should get more stable and less buggy over time. So the
1793 stable releases are ideal for people who don't need new features but
1794 would just like the software to work.
1796 Our criteria for backporting changes are:
1802 Documentation changes which don't affect any code are
1803 backported unless the documentation refers to a future feature
1804 which is not in stable.
1808 Bug fixes which are not controversial, fix obvious problems, and
1809 have been well tested are backported.
1813 Simple rearrangements of code which shouldn't affect how it works get
1814 backported. This is so that the code in the two branches doesn't get
1815 too far out of step, allowing us to backport future fixes more easily.
1819 We I<don't> backport new features, new APIs, new tools etc, except in
1820 one exceptional case: the new feature is required in order to
1821 implement an important bug fix.
1825 A new stable branch starts when we think the new features in
1826 development are substantial and compelling enough over the current
1827 stable branch to warrant it. When that happens we create new stable
1828 and development versions 1.N.0 and 1.(N+1).0 [N is even]. The new
1829 dot-oh release won't necessarily be so stable at this point, but by
1830 backporting fixes from development, that branch will stabilize over
1833 =head1 ENVIRONMENT VARIABLES
1837 =item LIBGUESTFS_APPEND
1839 Pass additional options to the guest kernel.
1841 =item LIBGUESTFS_DEBUG
1843 Set C<LIBGUESTFS_DEBUG=1> to enable verbose messages. This
1844 has the same effect as calling C<guestfs_set_verbose (g, 1)>.
1846 =item LIBGUESTFS_MEMSIZE
1848 Set the memory allocated to the qemu process, in megabytes. For
1851 LIBGUESTFS_MEMSIZE=700
1853 =item LIBGUESTFS_PATH
1855 Set the path that libguestfs uses to search for kernel and initrd.img.
1856 See the discussion of paths in section PATH above.
1858 =item LIBGUESTFS_QEMU
1860 Set the default qemu binary that libguestfs uses. If not set, then
1861 the qemu which was found at compile time by the configure script is
1864 See also L</QEMU WRAPPERS> above.
1866 =item LIBGUESTFS_TRACE
1868 Set C<LIBGUESTFS_TRACE=1> to enable command traces. This
1869 has the same effect as calling C<guestfs_set_trace (g, 1)>.
1873 Location of temporary directory, defaults to C</tmp>.
1875 If libguestfs was compiled to use the supermin appliance then the
1876 real appliance is cached in this directory, shared between all
1877 handles belonging to the same EUID. You can use C<$TMPDIR> to
1878 configure another directory to use in case C</tmp> is not large
1890 L<virt-inspector(1)>,
1891 L<virt-list-filesystems(1)>,
1892 L<virt-list-partitions(1)>,
1901 L<http://libguestfs.org/>.
1903 Tools with a similar purpose:
1912 To get a list of bugs against libguestfs use this link:
1914 L<https://bugzilla.redhat.com/buglist.cgi?component=libguestfs&product=Virtualization+Tools>
1916 To report a new bug against libguestfs use this link:
1918 L<https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virtualization+Tools>
1920 When reporting a bug, please check:
1926 That the bug hasn't been reported already.
1930 That you are testing a recent version.
1934 Describe the bug accurately, and give a way to reproduce it.
1938 Run libguestfs-test-tool and paste the B<complete, unedited>
1939 output into the bug report.
1945 Richard W.M. Jones (C<rjones at redhat dot com>)
1949 Copyright (C) 2009-2010 Red Hat Inc.
1950 L<http://libguestfs.org/>
1952 This library is free software; you can redistribute it and/or
1953 modify it under the terms of the GNU Lesser General Public
1954 License as published by the Free Software Foundation; either
1955 version 2 of the License, or (at your option) any later version.
1957 This library is distributed in the hope that it will be useful,
1958 but WITHOUT ANY WARRANTY; without even the implied warranty of
1959 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1960 Lesser General Public License for more details.
1962 You should have received a copy of the GNU Lesser General Public
1963 License along with this library; if not, write to the Free Software
1964 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA