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, 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 You can add a disk read-only using L</guestfs_add_drive_ro>, in which
128 case libguestfs won't modify the file.
130 Be extremely cautious if the disk image is in use, eg. if it is being
131 used by a virtual machine. Adding it read-write will almost certainly
132 cause disk corruption, but adding it read-only is safe.
134 You must add at least one disk image, and you may add multiple disk
135 images. In the API, the disk images are usually referred to as
136 C</dev/sda> (for the first one you added), C</dev/sdb> (for the second
139 Once L</guestfs_launch> has been called you cannot add any more images.
140 You can call L</guestfs_list_devices> to get a list of the device
141 names, in the order that you added them. See also L</BLOCK DEVICE
146 Before you can read or write files, create directories and so on in a
147 disk image that contains filesystems, you have to mount those
148 filesystems using L</guestfs_mount>. If you already know that a disk
149 image contains (for example) one partition with a filesystem on that
150 partition, then you can mount it directly:
152 guestfs_mount (g, "/dev/sda1", "/");
154 where C</dev/sda1> means literally the first partition (C<1>) of the
155 first disk image that we added (C</dev/sda>). If the disk contains
156 Linux LVM2 logical volumes you could refer to those instead (eg. C</dev/VG/LV>).
158 If you are given a disk image and you don't know what it contains then
159 you have to find out. Libguestfs can do that too: use
160 L</guestfs_list_partitions> and L</guestfs_lvs> to list possible
161 partitions and LVs, and either try mounting each to see what is
162 mountable, or else examine them with L</guestfs_vfs_type> or
163 L</guestfs_file>. But you might find it easier to look at higher level
164 programs built on top of libguestfs, in particular
165 L<virt-inspector(1)>.
167 To mount a disk image read-only, use L</guestfs_mount_ro>. There are
168 several other variations of the C<guestfs_mount_*> call.
170 =head2 FILESYSTEM ACCESS AND MODIFICATION
172 The majority of the libguestfs API consists of fairly low-level calls
173 for accessing and modifying the files, directories, symlinks etc on
174 mounted filesystems. There are over a hundred such calls which you
175 can find listed in detail below in this man page, and we don't even
176 pretend to cover them all in this overview.
178 Specify filenames as full paths, starting with C<"/"> and including
181 For example, if you mounted a filesystem at C<"/"> and you want to
182 read the file called C<"etc/passwd"> then you could do:
184 char *data = guestfs_cat (g, "/etc/passwd");
186 This would return C<data> as a newly allocated buffer containing the
187 full content of that file (with some conditions: see also
188 L</DOWNLOADING> below), or C<NULL> if there was an error.
190 As another example, to create a top-level directory on that filesystem
191 called C<"var"> you would do:
193 guestfs_mkdir (g, "/var");
195 To create a symlink you could do:
197 guestfs_ln_s (g, "/etc/init.d/portmap",
198 "/etc/rc3.d/S30portmap");
200 Libguestfs will reject attempts to use relative paths and there is no
201 concept of a current working directory.
203 Libguestfs can return errors in many situations: for example if the
204 filesystem isn't writable, or if a file or directory that you
205 requested doesn't exist. If you are using the C API (documented here)
206 you have to check for those error conditions after each call. (Other
207 language bindings turn these errors into exceptions).
209 File writes are affected by the per-handle umask, set by calling
210 L</guestfs_umask> and defaulting to 022. See L</UMASK>.
214 Libguestfs contains API calls to read, create and modify partition
215 tables on disk images.
217 In the common case where you want to create a single partition
218 covering the whole disk, you should use the L</guestfs_part_disk>
221 const char *parttype = "mbr";
222 if (disk_is_larger_than_2TB)
224 guestfs_part_disk (g, "/dev/sda", parttype);
226 Obviously this effectively wipes anything that was on that disk image
231 Libguestfs provides access to a large part of the LVM2 API, such as
232 L</guestfs_lvcreate> and L</guestfs_vgremove>. It won't make much sense
233 unless you familiarize yourself with the concepts of physical volumes,
234 volume groups and logical volumes.
236 This author strongly recommends reading the LVM HOWTO, online at
237 L<http://tldp.org/HOWTO/LVM-HOWTO/>.
241 Use L</guestfs_cat> to download small, text only files. This call
242 is limited to files which are less than 2 MB and which cannot contain
243 any ASCII NUL (C<\0>) characters. However it has a very simple
246 L</guestfs_read_file> can be used to read files which contain
247 arbitrary 8 bit data, since it returns a (pointer, size) pair.
248 However it is still limited to "small" files, less than 2 MB.
250 L</guestfs_download> can be used to download any file, with no
251 limits on content or size (even files larger than 4 GB).
253 To download multiple files, see L</guestfs_tar_out> and
258 It's often the case that you want to write a file or files to the disk
261 To write a small file with fixed content, use L</guestfs_write>. To
262 create a file of all zeroes, use L</guestfs_truncate_size> (sparse) or
263 L</guestfs_fallocate64> (with all disk blocks allocated). There are a
264 variety of other functions for creating test files, for example
265 L</guestfs_fill> and L</guestfs_fill_pattern>.
267 To upload a single file, use L</guestfs_upload>. This call has no
268 limits on file content or size (even files larger than 4 GB).
270 To upload multiple files, see L</guestfs_tar_in> and L</guestfs_tgz_in>.
272 However the fastest way to upload I<large numbers of arbitrary files>
273 is to turn them into a squashfs or CD ISO (see L<mksquashfs(8)> and
274 L<mkisofs(8)>), then attach this using L</guestfs_add_drive_ro>. If
275 you add the drive in a predictable way (eg. adding it last after all
276 other drives) then you can get the device name from
277 L</guestfs_list_devices> and mount it directly using
278 L</guestfs_mount_ro>. Note that squashfs images are sometimes
279 non-portable between kernel versions, and they don't support labels or
280 UUIDs. If you want to pre-build an image or you need to mount it
281 using a label or UUID, use an ISO image instead.
285 There are various different commands for copying between files and
286 devices and in and out of the guest filesystem. These are summarised
291 =item B<file> to B<file>
293 Use L</guestfs_cp> to copy a single file, or
294 L</guestfs_cp_a> to copy directories recursively.
296 =item B<file or device> to B<file or device>
298 Use L</guestfs_dd> which efficiently uses L<dd(1)>
299 to copy between files and devices in the guest.
301 Example: duplicate the contents of an LV:
303 guestfs_dd (g, "/dev/VG/Original", "/dev/VG/Copy");
305 The destination (C</dev/VG/Copy>) must be at least as large as the
306 source (C</dev/VG/Original>). To copy less than the whole
307 source device, use L</guestfs_copy_size>.
309 =item B<file on the host> to B<file or device>
311 Use L</guestfs_upload>. See L</UPLOADING> above.
313 =item B<file or device> to B<file on the host>
315 Use L</guestfs_download>. See L</DOWNLOADING> above.
321 L</guestfs_ll> is just designed for humans to read (mainly when using
322 the L<guestfish(1)>-equivalent command C<ll>).
324 L</guestfs_ls> is a quick way to get a list of files in a directory
325 from programs, as a flat list of strings.
327 L</guestfs_readdir> is a programmatic way to get a list of files in a
328 directory, plus additional information about each one. It is more
329 equivalent to using the L<readdir(3)> call on a local filesystem.
331 L</guestfs_find> and L</guestfs_find0> can be used to recursively list
334 =head2 RUNNING COMMANDS
336 Although libguestfs is a primarily an API for manipulating files
337 inside guest images, we also provide some limited facilities for
338 running commands inside guests.
340 There are many limitations to this:
346 The kernel version that the command runs under will be different
347 from what it expects.
351 If the command needs to communicate with daemons, then most likely
352 they won't be running.
356 The command will be running in limited memory.
360 Only supports Linux guests (not Windows, BSD, etc).
364 Architecture limitations (eg. won't work for a PPC guest on
369 For SELinux guests, you may need to enable SELinux and load policy
370 first. See L</SELINUX> in this manpage.
374 The two main API calls to run commands are L</guestfs_command> and
375 L</guestfs_sh> (there are also variations).
377 The difference is that L</guestfs_sh> runs commands using the shell, so
378 any shell globs, redirections, etc will work.
380 =head2 CONFIGURATION FILES
382 To read and write configuration files in Linux guest filesystems, we
383 strongly recommend using Augeas. For example, Augeas understands how
384 to read and write, say, a Linux shadow password file or X.org
385 configuration file, and so avoids you having to write that code.
387 The main Augeas calls are bound through the C<guestfs_aug_*> APIs. We
388 don't document Augeas itself here because there is excellent
389 documentation on the L<http://augeas.net/> website.
391 If you don't want to use Augeas (you fool!) then try calling
392 L</guestfs_read_lines> to get the file as a list of lines which
393 you can iterate over.
397 We support SELinux guests. To ensure that labeling happens correctly
398 in SELinux guests, you need to enable SELinux and load the guest's
405 Before launching, do:
407 guestfs_set_selinux (g, 1);
411 After mounting the guest's filesystem(s), load the policy. This
412 is best done by running the L<load_policy(8)> command in the
415 guestfs_sh (g, "/usr/sbin/load_policy");
417 (Older versions of C<load_policy> require you to specify the
418 name of the policy file).
422 Optionally, set the security context for the API. The correct
423 security context to use can only be known by inspecting the
424 guest. As an example:
426 guestfs_setcon (g, "unconfined_u:unconfined_r:unconfined_t:s0");
430 This will work for running commands and editing existing files.
432 When new files are created, you may need to label them explicitly,
433 for example by running the external command
434 C<restorecon pathname>.
438 Certain calls are affected by the current file mode creation mask (the
439 "umask"). In particular ones which create files or directories, such
440 as L</guestfs_touch>, L</guestfs_mknod> or L</guestfs_mkdir>. This
441 affects either the default mode that the file is created with or
442 modifies the mode that you supply.
444 The default umask is C<022>, so files are created with modes such as
445 C<0644> and directories with C<0755>.
447 There are two ways to avoid being affected by umask. Either set umask
448 to 0 (call C<guestfs_umask (g, 0)> early after launching). Or call
449 L</guestfs_chmod> after creating each file or directory.
451 For more information about umask, see L<umask(2)>.
453 =head2 SPECIAL CONSIDERATIONS FOR WINDOWS GUESTS
455 Libguestfs can mount NTFS partitions. It does this using the
456 L<http://www.ntfs-3g.org/> driver.
458 DOS and Windows still use drive letters, and the filesystems are
459 always treated as case insensitive by Windows itself, and therefore
460 you might find a Windows configuration file referring to a path like
461 C<c:\windows\system32>. When the filesystem is mounted in libguestfs,
462 that directory might be referred to as C</WINDOWS/System32>.
464 Drive letter mappings are outside the scope of libguestfs. You have
465 to use libguestfs to read the appropriate Windows Registry and
466 configuration files, to determine yourself how drives are mapped (see
467 also L<virt-inspector(1)>).
469 Replacing backslash characters with forward slash characters is also
470 outside the scope of libguestfs, but something that you can easily do.
472 Where we can help is in resolving the case insensitivity of paths.
473 For this, call L</guestfs_case_sensitive_path>.
475 Libguestfs also provides some help for decoding Windows Registry
476 "hive" files, through the library C<hivex> which is part of the
477 libguestfs project although ships as a separate tarball. You have to
478 locate and download the hive file(s) yourself, and then pass them to
479 C<hivex> functions. See also the programs L<hivexml(1)>,
480 L<hivexsh(1)>, L<hivexregedit(1)> and L<virt-win-reg(1)> for more help
483 =head2 USING LIBGUESTFS WITH OTHER PROGRAMMING LANGUAGES
485 Although we don't want to discourage you from using the C API, we will
486 mention here that the same API is also available in other languages.
488 The API is broadly identical in all supported languages. This means
489 that the C call C<guestfs_mount(g,path)> is
490 C<$g-E<gt>mount($path)> in Perl, C<g.mount(path)> in Python,
491 and C<Guestfs.mount g path> in OCaml. In other words, a
492 straightforward, predictable isomorphism between each language.
494 Error messages are automatically transformed
495 into exceptions if the language supports it.
497 We don't try to "object orientify" parts of the API in OO languages,
498 although contributors are welcome to write higher level APIs above
499 what we provide in their favourite languages if they wish.
505 You can use the I<guestfs.h> header file from C++ programs. The C++
506 API is identical to the C API. C++ classes and exceptions are not
511 The C# bindings are highly experimental. Please read the warnings
512 at the top of C<csharp/Libguestfs.cs>.
516 This is the only language binding that is working but incomplete.
517 Only calls which return simple integers have been bound in Haskell,
518 and we are looking for help to complete this binding.
522 Full documentation is contained in the Javadoc which is distributed
527 For documentation see the file C<guestfs.mli>.
531 For documentation see L<Sys::Guestfs(3)>.
535 For documentation do:
543 Use the Guestfs module. There is no Ruby-specific documentation, but
544 you can find examples written in Ruby in the libguestfs source.
546 =item B<shell scripts>
548 For documentation see L<guestfish(1)>.
552 =head2 LIBGUESTFS GOTCHAS
554 L<http://en.wikipedia.org/wiki/Gotcha_(programming)>: "A feature of a
555 system [...] that works in the way it is documented but is
556 counterintuitive and almost invites mistakes."
558 Since we developed libguestfs and the associated tools, there are
559 several things we would have designed differently, but are now stuck
560 with for backwards compatibility or other reasons. If there is ever a
561 libguestfs 2.0 release, you can expect these to change. Beware of
566 =item Autosync / forgetting to sync.
568 When modifying a filesystem from C or another language, you B<must>
569 unmount all filesystems and call L</guestfs_sync> explicitly before
570 you close the libguestfs handle. You can also call:
572 guestfs_set_autosync (g, 1);
574 to have the unmount/sync done automatically for you when the handle 'g'
575 is closed. (This feature is called "autosync", L</guestfs_set_autosync>
578 If you forget to do this, then it is entirely possible that your
579 changes won't be written out, or will be partially written, or (very
580 rarely) that you'll get disk corruption.
582 Note that in L<guestfish(3)> autosync is the default. So quick and
583 dirty guestfish scripts that forget to sync will work just fine, which
584 can make this very puzzling if you are trying to debug a problem.
586 =item Mount option C<-o sync> should not be the default.
588 If you use L</guestfs_mount>, then C<-o sync,noatime> are added
589 implicitly. However C<-o sync> does not add any reliability benefit,
590 but does have a very large performance impact.
592 The work around is to use L</guestfs_mount_options> and set the mount
593 options that you actually want to use.
595 =item Read-only should be the default.
597 In L<guestfish(3)>, I<--ro> should be the default, and you should
598 have to specify I<--rw> if you want to make changes to the image.
600 This would reduce the potential to corrupt live VM images.
602 Note that many filesystems change the disk when you just mount and
603 unmount, even if you didn't perform any writes. You need to use
604 L</guestfs_add_drive_ro> to guarantee that the disk is not changed.
606 =item guestfish command line is hard to use.
608 C<guestfish disk.img> doesn't do what people expect (open C<disk.img>
609 for examination). It tries to run a guestfish command C<disk.img>
610 which doesn't exist, so it fails. In earlier versions of guestfish
611 the error message was also unintuitive, but we have corrected this
612 since. Like the Bourne shell, we should have used C<guestfish -c
613 command> to run commands.
615 =item guestfish megabyte modifiers don't work right on all commands
617 In recent guestfish you can use C<1M> to mean 1 megabyte (and
618 similarly for other modifiers). What guestfish actually does is to
619 multiply the number part by the modifier part and pass the result to
620 the C API. However this doesn't work for a few APIs which aren't
621 expecting bytes, but are already expecting some other unit
624 The most common is L</guestfs_lvcreate>. The guestfish command:
628 does not do what you might expect. Instead because
629 L</guestfs_lvcreate> is already expecting megabytes, this tries to
630 create a 100 I<terabyte> (100 megabytes * megabytes) logical volume.
631 The error message you get from this is also a little obscure.
633 This could be fixed in the generator by specially marking parameters
634 and return values which take bytes or other units.
636 =item Protocol limit of 256 characters for error messages
638 This limit is both rather small and quite unnecessary. We should be
639 able to return error messages up to the length of the protocol message
642 Note that we cannot change the protocol without some breakage, because
643 there are distributions that repackage the Fedora appliance.
645 =item Protocol should return errno with error messages.
647 It would be a nice-to-have to be able to get the original value of
648 'errno' from inside the appliance along error paths (where set).
649 Currently L<guestmount(1)> goes through hoops to try to reverse the
650 error message string into an errno, see the function error() in
655 =head2 PROTOCOL LIMITS
657 Internally libguestfs uses a message-based protocol to pass API calls
658 and their responses to and from a small "appliance" (see L</INTERNALS>
659 for plenty more detail about this). The maximum message size used by
660 the protocol is slightly less than 4 MB. For some API calls you may
661 need to be aware of this limit. The API calls which may be affected
662 are individually documented, with a link back to this section of the
665 A simple call such as L</guestfs_cat> returns its result (the file
666 data) in a simple string. Because this string is at some point
667 internally encoded as a message, the maximum size that it can return
668 is slightly under 4 MB. If the requested file is larger than this
669 then you will get an error.
671 In order to transfer large files into and out of the guest filesystem,
672 you need to use particular calls that support this. The sections
673 L</UPLOADING> and L</DOWNLOADING> document how to do this.
675 You might also consider mounting the disk image using our FUSE
676 filesystem support (L<guestmount(1)>).
678 =head1 CONNECTION MANAGEMENT
682 C<guestfs_h> is the opaque type representing a connection handle.
683 Create a handle by calling L</guestfs_create>. Call L</guestfs_close>
684 to free the handle and release all resources used.
686 For information on using multiple handles and threads, see the section
687 L</MULTIPLE HANDLES AND MULTIPLE THREADS> below.
689 =head2 guestfs_create
691 guestfs_h *guestfs_create (void);
693 Create a connection handle.
695 You have to call L</guestfs_add_drive> on the handle at least once.
697 This function returns a non-NULL pointer to a handle on success or
700 After configuring the handle, you have to call L</guestfs_launch>.
702 You may also want to configure error handling for the handle. See
703 L</ERROR HANDLING> section below.
707 void guestfs_close (guestfs_h *g);
709 This closes the connection handle and frees up all resources used.
711 =head1 ERROR HANDLING
713 The convention in all functions that return C<int> is that they return
714 C<-1> to indicate an error. You can get additional information on
715 errors by calling L</guestfs_last_error> and/or by setting up an error
716 handler with L</guestfs_set_error_handler>.
718 The default error handler prints the information string to C<stderr>.
720 Out of memory errors are handled differently. The default action is
721 to call L<abort(3)>. If this is undesirable, then you can set a
722 handler using L</guestfs_set_out_of_memory_handler>.
724 =head2 guestfs_last_error
726 const char *guestfs_last_error (guestfs_h *g);
728 This returns the last error message that happened on C<g>. If
729 there has not been an error since the handle was created, then this
732 The lifetime of the returned string is until the next error occurs, or
733 L</guestfs_close> is called.
735 The error string is not localized (ie. is always in English), because
736 this makes searching for error messages in search engines give the
737 largest number of results.
739 =head2 guestfs_set_error_handler
741 typedef void (*guestfs_error_handler_cb) (guestfs_h *g,
744 void guestfs_set_error_handler (guestfs_h *g,
745 guestfs_error_handler_cb cb,
748 The callback C<cb> will be called if there is an error. The
749 parameters passed to the callback are an opaque data pointer and the
750 error message string.
752 Note that the message string C<msg> is freed as soon as the callback
753 function returns, so if you want to stash it somewhere you must make
756 The default handler prints messages on C<stderr>.
758 If you set C<cb> to C<NULL> then I<no> handler is called.
760 =head2 guestfs_get_error_handler
762 guestfs_error_handler_cb guestfs_get_error_handler (guestfs_h *g,
765 Returns the current error handler callback.
767 =head2 guestfs_set_out_of_memory_handler
769 typedef void (*guestfs_abort_cb) (void);
770 int guestfs_set_out_of_memory_handler (guestfs_h *g,
773 The callback C<cb> will be called if there is an out of memory
774 situation. I<Note this callback must not return>.
776 The default is to call L<abort(3)>.
778 You cannot set C<cb> to C<NULL>. You can't ignore out of memory
781 =head2 guestfs_get_out_of_memory_handler
783 guestfs_abort_fn guestfs_get_out_of_memory_handler (guestfs_h *g);
785 This returns the current out of memory handler.
789 Libguestfs needs a kernel and initrd.img, which it finds by looking
790 along an internal path.
792 By default it looks for these in the directory C<$libdir/guestfs>
793 (eg. C</usr/local/lib/guestfs> or C</usr/lib64/guestfs>).
795 Use L</guestfs_set_path> or set the environment variable
796 L</LIBGUESTFS_PATH> to change the directories that libguestfs will
797 search in. The value is a colon-separated list of paths. The current
798 directory is I<not> searched unless the path contains an empty element
799 or C<.>. For example C<LIBGUESTFS_PATH=:/usr/lib/guestfs> would
800 search the current directory and then C</usr/lib/guestfs>.
802 =head1 HIGH-LEVEL API ACTIONS
806 We guarantee the libguestfs ABI (binary interface), for public,
807 high-level actions as outlined in this section. Although we will
808 deprecate some actions, for example if they get replaced by newer
809 calls, we will keep the old actions forever. This allows you the
810 developer to program in confidence against the libguestfs API.
820 =head2 GROUPS OF FUNCTIONALITY IN THE APPLIANCE
822 Using L</guestfs_available> you can test availability of
823 the following groups of functions. This test queries the
824 appliance to see if the appliance you are currently using
825 supports the functionality.
829 =head2 GUESTFISH supported COMMAND
831 In L<guestfish(3)> there is a handy interactive command
832 C<supported> which prints out the available groups and
833 whether they are supported by this build of libguestfs.
834 Note however that you have to do C<run> first.
836 =head2 SINGLE CALLS AT COMPILE TIME
838 If you need to test whether a single libguestfs function is
839 available at compile time, we recommend using build tools
840 such as autoconf or cmake. For example in autotools you could
843 AC_CHECK_LIB([guestfs],[guestfs_create])
844 AC_CHECK_FUNCS([guestfs_dd])
846 which would result in C<HAVE_GUESTFS_DD> being either defined
847 or not defined in your program.
849 =head2 SINGLE CALLS AT RUN TIME
851 Testing at compile time doesn't guarantee that a function really
852 exists in the library. The reason is that you might be dynamically
853 linked against a previous I<libguestfs.so> (dynamic library)
854 which doesn't have the call. This situation unfortunately results
855 in a segmentation fault, which is a shortcoming of the C dynamic
856 linking system itself.
858 You can use L<dlopen(3)> to test if a function is available
859 at run time, as in this example program (note that you still
860 need the compile time check as well):
872 #ifdef HAVE_GUESTFS_DD
876 /* Test if the function guestfs_dd is really available. */
877 dl = dlopen (NULL, RTLD_LAZY);
879 fprintf (stderr, "dlopen: %s\n", dlerror ());
882 has_function = dlsym (dl, "guestfs_dd") != NULL;
886 printf ("this libguestfs.so does NOT have guestfs_dd function\n");
888 printf ("this libguestfs.so has guestfs_dd function\n");
889 /* Now it's safe to call
890 guestfs_dd (g, "foo", "bar");
894 printf ("guestfs_dd function was not found at compile time\n");
898 You may think the above is an awful lot of hassle, and it is.
899 There are other ways outside of the C linking system to ensure
900 that this kind of incompatibility never arises, such as using
903 Requires: libguestfs >= 1.0.80
907 <!-- old anchor for the next section -->
908 <a name="state_machine_and_low_level_event_api"/>
914 Internally, libguestfs is implemented by running an appliance (a
915 special type of small virtual machine) using L<qemu(1)>. Qemu runs as
916 a child process of the main program.
922 | | child process / appliance
923 | | __________________________
925 +-------------------+ RPC | +-----------------+ |
926 | libguestfs <--------------------> guestfsd | |
927 | | | +-----------------+ |
928 \___________________/ | | Linux kernel | |
929 | +--^--------------+ |
930 \_________|________________/
938 The library, linked to the main program, creates the child process and
939 hence the appliance in the L</guestfs_launch> function.
941 Inside the appliance is a Linux kernel and a complete stack of
942 userspace tools (such as LVM and ext2 programs) and a small
943 controlling daemon called L</guestfsd>. The library talks to
944 L</guestfsd> using remote procedure calls (RPC). There is a mostly
945 one-to-one correspondence between libguestfs API calls and RPC calls
946 to the daemon. Lastly the disk image(s) are attached to the qemu
947 process which translates device access by the appliance's Linux kernel
948 into accesses to the image.
950 A common misunderstanding is that the appliance "is" the virtual
951 machine. Although the disk image you are attached to might also be
952 used by some virtual machine, libguestfs doesn't know or care about
953 this. (But you will care if both libguestfs's qemu process and your
954 virtual machine are trying to update the disk image at the same time,
955 since these usually results in massive disk corruption).
959 libguestfs uses a state machine to model the child process:
970 / | \ \ guestfs_launch
981 \______/ <------ \________/
983 The normal transitions are (1) CONFIG (when the handle is created, but
984 there is no child process), (2) LAUNCHING (when the child process is
985 booting up), (3) alternating between READY and BUSY as commands are
986 issued to, and carried out by, the child process.
988 The guest may be killed by L</guestfs_kill_subprocess>, or may die
989 asynchronously at any time (eg. due to some internal error), and that
990 causes the state to transition back to CONFIG.
992 Configuration commands for qemu such as L</guestfs_add_drive> can only
993 be issued when in the CONFIG state.
995 The high-level API offers two calls that go from CONFIG through
996 LAUNCHING to READY. L</guestfs_launch> blocks until the child process
997 is READY to accept commands (or until some failure or timeout).
998 L</guestfs_launch> internally moves the state from CONFIG to LAUNCHING
1001 High-level API actions such as L</guestfs_mount> can only be issued
1002 when in the READY state. These high-level API calls block waiting for
1003 the command to be carried out (ie. the state to transition to BUSY and
1004 then back to READY). But using the low-level event API, you get
1005 non-blocking versions. (But you can still only carry out one
1006 operation per handle at a time - that is a limitation of the
1007 communications protocol we use).
1009 Finally, the child process sends asynchronous messages back to the
1010 main program, such as kernel log messages. Mostly these are ignored
1011 by the high-level API, but using the low-level event API you can
1012 register to receive these messages.
1014 =head2 SETTING CALLBACKS TO HANDLE EVENTS
1016 The child process generates events in some situations. Current events
1017 include: receiving a log message, the child process exits.
1019 Use the C<guestfs_set_*_callback> functions to set a callback for
1020 different types of events.
1022 Only I<one callback of each type> can be registered for each handle.
1023 Calling C<guestfs_set_*_callback> again overwrites the previous
1024 callback of that type. Cancel all callbacks of this type by calling
1025 this function with C<cb> set to C<NULL>.
1027 =head2 guestfs_set_log_message_callback
1029 typedef void (*guestfs_log_message_cb) (guestfs_h *g, void *opaque,
1030 char *buf, int len);
1031 void guestfs_set_log_message_callback (guestfs_h *g,
1032 guestfs_log_message_cb cb,
1035 The callback function C<cb> will be called whenever qemu or the guest
1036 writes anything to the console.
1038 Use this function to capture kernel messages and similar.
1040 Normally there is no log message handler, and log messages are just
1043 =head2 guestfs_set_subprocess_quit_callback
1045 typedef void (*guestfs_subprocess_quit_cb) (guestfs_h *g, void *opaque);
1046 void guestfs_set_subprocess_quit_callback (guestfs_h *g,
1047 guestfs_subprocess_quit_cb cb,
1050 The callback function C<cb> will be called when the child process
1051 quits, either asynchronously or if killed by
1052 L</guestfs_kill_subprocess>. (This corresponds to a transition from
1053 any state to the CONFIG state).
1055 =head2 guestfs_set_launch_done_callback
1057 typedef void (*guestfs_launch_done_cb) (guestfs_h *g, void *opaque);
1058 void guestfs_set_launch_done_callback (guestfs_h *g,
1059 guestfs_launch_done_cb cb,
1062 The callback function C<cb> will be called when the child process
1063 becomes ready first time after it has been launched. (This
1064 corresponds to a transition from LAUNCHING to the READY state).
1066 =head2 guestfs_set_close_callback
1068 typedef void (*guestfs_close_cb) (guestfs_h *g, void *opaque);
1069 void guestfs_set_close_callback (guestfs_h *g,
1070 guestfs_close_cb cb,
1073 The callback function C<cb> will be called while the handle
1074 is being closed (synchronously from L</guestfs_close>).
1076 Note that libguestfs installs an L<atexit(3)> handler to try to
1077 clean up handles that are open when the program exits. This
1078 means that this callback might be called indirectly from
1079 L<exit(3)>, which can cause unexpected problems in higher-level
1080 languages (eg. if your HLL interpreter has already been cleaned
1081 up by the time this is called, and if your callback then jumps
1082 into some HLL function).
1084 =head1 BLOCK DEVICE NAMING
1086 In the kernel there is now quite a profusion of schemata for naming
1087 block devices (in this context, by I<block device> I mean a physical
1088 or virtual hard drive). The original Linux IDE driver used names
1089 starting with C</dev/hd*>. SCSI devices have historically used a
1090 different naming scheme, C</dev/sd*>. When the Linux kernel I<libata>
1091 driver became a popular replacement for the old IDE driver
1092 (particularly for SATA devices) those devices also used the
1093 C</dev/sd*> scheme. Additionally we now have virtual machines with
1094 paravirtualized drivers. This has created several different naming
1095 systems, such as C</dev/vd*> for virtio disks and C</dev/xvd*> for Xen
1098 As discussed above, libguestfs uses a qemu appliance running an
1099 embedded Linux kernel to access block devices. We can run a variety
1100 of appliances based on a variety of Linux kernels.
1102 This causes a problem for libguestfs because many API calls use device
1103 or partition names. Working scripts and the recipe (example) scripts
1104 that we make available over the internet could fail if the naming
1107 Therefore libguestfs defines C</dev/sd*> as the I<standard naming
1108 scheme>. Internally C</dev/sd*> names are translated, if necessary,
1109 to other names as required. For example, under RHEL 5 which uses the
1110 C</dev/hd*> scheme, any device parameter C</dev/sda2> is translated to
1111 C</dev/hda2> transparently.
1113 Note that this I<only> applies to parameters. The
1114 L</guestfs_list_devices>, L</guestfs_list_partitions> and similar calls
1115 return the true names of the devices and partitions as known to the
1118 =head2 ALGORITHM FOR BLOCK DEVICE NAME TRANSLATION
1120 Usually this translation is transparent. However in some (very rare)
1121 cases you may need to know the exact algorithm. Such cases include
1122 where you use L</guestfs_config> to add a mixture of virtio and IDE
1123 devices to the qemu-based appliance, so have a mixture of C</dev/sd*>
1124 and C</dev/vd*> devices.
1126 The algorithm is applied only to I<parameters> which are known to be
1127 either device or partition names. Return values from functions such
1128 as L</guestfs_list_devices> are never changed.
1134 Is the string a parameter which is a device or partition name?
1138 Does the string begin with C</dev/sd>?
1142 Does the named device exist? If so, we use that device.
1143 However if I<not> then we continue with this algorithm.
1147 Replace initial C</dev/sd> string with C</dev/hd>.
1149 For example, change C</dev/sda2> to C</dev/hda2>.
1151 If that named device exists, use it. If not, continue.
1155 Replace initial C</dev/sd> string with C</dev/vd>.
1157 If that named device exists, use it. If not, return an error.
1161 =head2 PORTABILITY CONCERNS
1163 Although the standard naming scheme and automatic translation is
1164 useful for simple programs and guestfish scripts, for larger programs
1165 it is best not to rely on this mechanism.
1167 Where possible for maximum future portability programs using
1168 libguestfs should use these future-proof techniques:
1174 Use L</guestfs_list_devices> or L</guestfs_list_partitions> to list
1175 actual device names, and then use those names directly.
1177 Since those device names exist by definition, they will never be
1182 Use higher level ways to identify filesystems, such as LVM names,
1183 UUIDs and filesystem labels.
1189 =head2 COMMUNICATION PROTOCOL
1191 Don't rely on using this protocol directly. This section documents
1192 how it currently works, but it may change at any time.
1194 The protocol used to talk between the library and the daemon running
1195 inside the qemu virtual machine is a simple RPC mechanism built on top
1196 of XDR (RFC 1014, RFC 1832, RFC 4506).
1198 The detailed format of structures is in C<src/guestfs_protocol.x>
1199 (note: this file is automatically generated).
1201 There are two broad cases, ordinary functions that don't have any
1202 C<FileIn> and C<FileOut> parameters, which are handled with very
1203 simple request/reply messages. Then there are functions that have any
1204 C<FileIn> or C<FileOut> parameters, which use the same request and
1205 reply messages, but they may also be followed by files sent using a
1208 =head3 ORDINARY FUNCTIONS (NO FILEIN/FILEOUT PARAMS)
1210 For ordinary functions, the request message is:
1212 total length (header + arguments,
1213 but not including the length word itself)
1214 struct guestfs_message_header (encoded as XDR)
1215 struct guestfs_<foo>_args (encoded as XDR)
1217 The total length field allows the daemon to allocate a fixed size
1218 buffer into which it slurps the rest of the message. As a result, the
1219 total length is limited to C<GUESTFS_MESSAGE_MAX> bytes (currently
1220 4MB), which means the effective size of any request is limited to
1221 somewhere under this size.
1223 Note also that many functions don't take any arguments, in which case
1224 the C<guestfs_I<foo>_args> is completely omitted.
1226 The header contains the procedure number (C<guestfs_proc>) which is
1227 how the receiver knows what type of args structure to expect, or none
1230 The reply message for ordinary functions is:
1232 total length (header + ret,
1233 but not including the length word itself)
1234 struct guestfs_message_header (encoded as XDR)
1235 struct guestfs_<foo>_ret (encoded as XDR)
1237 As above the C<guestfs_I<foo>_ret> structure may be completely omitted
1238 for functions that return no formal return values.
1240 As above the total length of the reply is limited to
1241 C<GUESTFS_MESSAGE_MAX>.
1243 In the case of an error, a flag is set in the header, and the reply
1244 message is slightly changed:
1246 total length (header + error,
1247 but not including the length word itself)
1248 struct guestfs_message_header (encoded as XDR)
1249 struct guestfs_message_error (encoded as XDR)
1251 The C<guestfs_message_error> structure contains the error message as a
1254 =head3 FUNCTIONS THAT HAVE FILEIN PARAMETERS
1256 A C<FileIn> parameter indicates that we transfer a file I<into> the
1257 guest. The normal request message is sent (see above). However this
1258 is followed by a sequence of file chunks.
1260 total length (header + arguments,
1261 but not including the length word itself,
1262 and not including the chunks)
1263 struct guestfs_message_header (encoded as XDR)
1264 struct guestfs_<foo>_args (encoded as XDR)
1265 sequence of chunks for FileIn param #0
1266 sequence of chunks for FileIn param #1 etc.
1268 The "sequence of chunks" is:
1270 length of chunk (not including length word itself)
1271 struct guestfs_chunk (encoded as XDR)
1273 struct guestfs_chunk (encoded as XDR)
1276 struct guestfs_chunk (with data.data_len == 0)
1278 The final chunk has the C<data_len> field set to zero. Additionally a
1279 flag is set in the final chunk to indicate either successful
1280 completion or early cancellation.
1282 At time of writing there are no functions that have more than one
1283 FileIn parameter. However this is (theoretically) supported, by
1284 sending the sequence of chunks for each FileIn parameter one after
1285 another (from left to right).
1287 Both the library (sender) I<and> the daemon (receiver) may cancel the
1288 transfer. The library does this by sending a chunk with a special
1289 flag set to indicate cancellation. When the daemon sees this, it
1290 cancels the whole RPC, does I<not> send any reply, and goes back to
1291 reading the next request.
1293 The daemon may also cancel. It does this by writing a special word
1294 C<GUESTFS_CANCEL_FLAG> to the socket. The library listens for this
1295 during the transfer, and if it gets it, it will cancel the transfer
1296 (it sends a cancel chunk). The special word is chosen so that even if
1297 cancellation happens right at the end of the transfer (after the
1298 library has finished writing and has started listening for the reply),
1299 the "spurious" cancel flag will not be confused with the reply
1302 This protocol allows the transfer of arbitrary sized files (no 32 bit
1303 limit), and also files where the size is not known in advance
1304 (eg. from pipes or sockets). However the chunks are rather small
1305 (C<GUESTFS_MAX_CHUNK_SIZE>), so that neither the library nor the
1306 daemon need to keep much in memory.
1308 =head3 FUNCTIONS THAT HAVE FILEOUT PARAMETERS
1310 The protocol for FileOut parameters is exactly the same as for FileIn
1311 parameters, but with the roles of daemon and library reversed.
1313 total length (header + ret,
1314 but not including the length word itself,
1315 and not including the chunks)
1316 struct guestfs_message_header (encoded as XDR)
1317 struct guestfs_<foo>_ret (encoded as XDR)
1318 sequence of chunks for FileOut param #0
1319 sequence of chunks for FileOut param #1 etc.
1321 =head3 INITIAL MESSAGE
1323 Because the underlying channel (QEmu -net channel) doesn't have any
1324 sort of connection control, when the daemon launches it sends an
1325 initial word (C<GUESTFS_LAUNCH_FLAG>) which indicates that the guest
1326 and daemon is alive. This is what L</guestfs_launch> waits for.
1328 =head1 MULTIPLE HANDLES AND MULTIPLE THREADS
1330 All high-level libguestfs actions are synchronous. If you want
1331 to use libguestfs asynchronously then you must create a thread.
1333 Only use the handle from a single thread. Either use the handle
1334 exclusively from one thread, or provide your own mutex so that two
1335 threads cannot issue calls on the same handle at the same time.
1337 =head1 QEMU WRAPPERS
1339 If you want to compile your own qemu, run qemu from a non-standard
1340 location, or pass extra arguments to qemu, then you can write a
1341 shell-script wrapper around qemu.
1343 There is one important rule to remember: you I<must C<exec qemu>> as
1344 the last command in the shell script (so that qemu replaces the shell
1345 and becomes the direct child of the libguestfs-using program). If you
1346 don't do this, then the qemu process won't be cleaned up correctly.
1348 Here is an example of a wrapper, where I have built my own copy of
1352 qemudir=/home/rjones/d/qemu
1353 exec $qemudir/x86_64-softmmu/qemu-system-x86_64 -L $qemudir/pc-bios "$@"
1355 Save this script as C</tmp/qemu.wrapper> (or wherever), C<chmod +x>,
1356 and then use it by setting the LIBGUESTFS_QEMU environment variable.
1359 LIBGUESTFS_QEMU=/tmp/qemu.wrapper guestfish
1361 Note that libguestfs also calls qemu with the -help and -version
1362 options in order to determine features.
1364 =head1 LIBGUESTFS VERSION NUMBERS
1366 Since April 2010, libguestfs has started to make separate development
1367 and stable releases, along with corresponding branches in our git
1368 repository. These separate releases can be identified by version
1371 even numbers for stable: 1.2.x, 1.4.x, ...
1372 .-------- odd numbers for development: 1.3.x, 1.5.x, ...
1378 | `-------- sub-version
1380 `------ always '1' because we don't change the ABI
1382 Thus "1.3.5" is the 5th update to the development branch "1.3".
1384 As time passes we cherry pick fixes from the development branch and
1385 backport those into the stable branch, the effect being that the
1386 stable branch should get more stable and less buggy over time. So the
1387 stable releases are ideal for people who don't need new features but
1388 would just like the software to work.
1390 Our criteria for backporting changes are:
1396 Documentation changes which don't affect any code are
1397 backported unless the documentation refers to a future feature
1398 which is not in stable.
1402 Bug fixes which are not controversial, fix obvious problems, and
1403 have been well tested are backported.
1407 Simple rearrangements of code which shouldn't affect how it works get
1408 backported. This is so that the code in the two branches doesn't get
1409 too far out of step, allowing us to backport future fixes more easily.
1413 We I<don't> backport new features, new APIs, new tools etc, except in
1414 one exceptional case: the new feature is required in order to
1415 implement an important bug fix.
1419 A new stable branch starts when we think the new features in
1420 development are substantial and compelling enough over the current
1421 stable branch to warrant it. When that happens we create new stable
1422 and development versions 1.N.0 and 1.(N+1).0 [N is even]. The new
1423 dot-oh release won't necessarily be so stable at this point, but by
1424 backporting fixes from development, that branch will stabilize over
1427 =head1 ENVIRONMENT VARIABLES
1431 =item LIBGUESTFS_APPEND
1433 Pass additional options to the guest kernel.
1435 =item LIBGUESTFS_DEBUG
1437 Set C<LIBGUESTFS_DEBUG=1> to enable verbose messages. This
1438 has the same effect as calling C<guestfs_set_verbose (g, 1)>.
1440 =item LIBGUESTFS_MEMSIZE
1442 Set the memory allocated to the qemu process, in megabytes. For
1445 LIBGUESTFS_MEMSIZE=700
1447 =item LIBGUESTFS_PATH
1449 Set the path that libguestfs uses to search for kernel and initrd.img.
1450 See the discussion of paths in section PATH above.
1452 =item LIBGUESTFS_QEMU
1454 Set the default qemu binary that libguestfs uses. If not set, then
1455 the qemu which was found at compile time by the configure script is
1458 See also L</QEMU WRAPPERS> above.
1460 =item LIBGUESTFS_TRACE
1462 Set C<LIBGUESTFS_TRACE=1> to enable command traces. This
1463 has the same effect as calling C<guestfs_set_trace (g, 1)>.
1467 Location of temporary directory, defaults to C</tmp>.
1469 If libguestfs was compiled to use the supermin appliance then each
1470 handle will require rather a large amount of space in this directory
1471 for short periods of time (~ 80 MB). You can use C<$TMPDIR> to
1472 configure another directory to use in case C</tmp> is not large
1484 L<virt-inspector(1)>,
1485 L<virt-list-filesystems(1)>,
1486 L<virt-list-partitions(1)>,
1495 L<http://libguestfs.org/>.
1497 Tools with a similar purpose:
1506 To get a list of bugs against libguestfs use this link:
1508 L<https://bugzilla.redhat.com/buglist.cgi?component=libguestfs&product=Virtualization+Tools>
1510 To report a new bug against libguestfs use this link:
1512 L<https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virtualization+Tools>
1514 When reporting a bug, please check:
1520 That the bug hasn't been reported already.
1524 That you are testing a recent version.
1528 Describe the bug accurately, and give a way to reproduce it.
1532 Run libguestfs-test-tool and paste the B<complete, unedited>
1533 output into the bug report.
1539 Richard W.M. Jones (C<rjones at redhat dot com>)
1543 Copyright (C) 2009-2010 Red Hat Inc.
1544 L<http://libguestfs.org/>
1546 This library is free software; you can redistribute it and/or
1547 modify it under the terms of the GNU Lesser General Public
1548 License as published by the Free Software Foundation; either
1549 version 2 of the License, or (at your option) any later version.
1551 This library is distributed in the hope that it will be useful,
1552 but WITHOUT ANY WARRANTY; without even the implied warranty of
1553 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1554 Lesser General Public License for more details.
1556 You should have received a copy of the GNU Lesser General Public
1557 License along with this library; if not, write to the Free Software
1558 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA