5 guestfs - Library for accessing and modifying virtual machine images
11 guestfs_h *handle = guestfs_create ();
12 guestfs_add_drive (handle, "guest.img");
13 guestfs_launch (handle);
14 guestfs_mount (handle, "/dev/sda1", "/");
15 guestfs_touch (handle, "/hello");
16 guestfs_sync (handle);
17 guestfs_close (handle);
21 Libguestfs is a library for accessing and modifying guest disk images.
22 Amongst the things this is good for: making batch configuration
23 changes to guests, getting disk used/free statistics (see also:
24 virt-df), migrating between virtualization systems (see also:
25 virt-p2v), performing partial backups, performing partial guest
26 clones, cloning guests and changing registry/UUID/hostname info, and
29 Libguestfs uses Linux kernel and qemu code, and can access any type of
30 guest filesystem that Linux and qemu can, including but not limited
31 to: ext2/3/4, btrfs, FAT and NTFS, LVM, many different disk partition
32 schemes, qcow, qcow2, vmdk.
34 Libguestfs provides ways to enumerate guest storage (eg. partitions,
35 LVs, what filesystem is in each LV, etc.). It can also run commands
36 in the context of the guest. Also you can access filesystems over FTP.
38 Libguestfs is a library that can be linked with C and C++ management
39 programs (or management programs written in OCaml, Perl, Python, Ruby, Java
40 or Haskell). You can also use it from shell scripts or the command line.
42 You don't need to be root to use libguestfs, although obviously you do
43 need enough permissions to access the disk images.
45 Libguestfs is a large API because it can do many things. For a gentle
46 introduction, please read the L</API OVERVIEW> section next.
50 This section provides a gentler overview of the libguestfs API. We
51 also try to group API calls together, where that may not be obvious
52 from reading about the individual calls below.
56 Before you can use libguestfs calls, you have to create a handle.
57 Then you must add at least one disk image to the handle, followed by
58 launching the handle, then performing whatever operations you want,
59 and finally closing the handle. So the general structure of all
60 libguestfs-using programs looks like this:
62 guestfs_h *handle = guestfs_create ();
64 /* Call guestfs_add_drive additional times if there are
65 * multiple disk images.
67 guestfs_add_drive (handle, "guest.img");
69 /* Most manipulation calls won't work until you've launched
70 * the handle. You have to do this _after_ adding drives
71 * and _before_ other commands.
73 guestfs_launch (handle);
75 /* Now you can examine what partitions, LVs etc are available.
77 char **partitions = guestfs_list_partitions (handle);
78 char **logvols = guestfs_lvs (handle);
80 /* To access a filesystem in the image, you must mount it.
82 guestfs_mount (handle, "/dev/sda1", "/");
84 /* Now you can perform filesystem actions on the guest
87 guestfs_touch (handle, "/hello");
89 /* You only need to call guestfs_sync if you have made
90 * changes to the guest image.
92 guestfs_sync (handle);
94 /* Close the handle. */
95 guestfs_close (handle);
97 The code above doesn't include any error checking. In real code you
98 should check return values carefully for errors. In general all
99 functions that return integers return C<-1> on error, and all
100 functions that return pointers return C<NULL> on error. See section
101 L</ERROR HANDLING> below for how to handle errors, and consult the
102 documentation for each function call below to see precisely how they
103 return error indications.
107 The image filename (C<"guest.img"> in the example above) could be a
108 disk image from a virtual machine, a L<dd(1)> copy of a physical block
109 device, an actual block device, or simply an empty file of zeroes that
110 you have created through L<posix_fallocate(3)>. Libguestfs lets you
111 do useful things to all of these.
113 You can add a disk read-only using C<guestfs_add_drive_ro>, in which
114 case libguestfs won't modify the file.
116 Be extremely cautious if the disk image is in use, eg. if it is being
117 used by a virtual machine. Adding it read-write will almost certainly
118 cause disk corruption, but adding it read-only is safe.
120 You must add at least one disk image, and you may add multiple disk
121 images. In the API, the disk images are usually referred to as
122 C</dev/sda> (for the first one you added), C</dev/sdb> (for the second
125 Once C<guestfs_launch> has been called you cannot add any more images.
126 You can call C<guestfs_list_devices> to get a list of the device
127 names, in the order that you added them. See also L</BLOCK DEVICE
132 Before you can read or write files, create directories and so on in a
133 disk image that contains filesystems, you have to mount those
134 filesystems using C<guestfs_mount>. If you already know that a disk
135 image contains (for example) one partition with a filesystem on that
136 partition, then you can mount it directly:
138 guestfs_mount (handle, "/dev/sda1", "/");
140 where C</dev/sda1> means literally the first partition (C<1>) of the
141 first disk image that we added (C</dev/sda>). If the disk contains
142 Linux LVM2 logical volumes you could refer to those instead (eg. C</dev/VG/LV>).
144 If you are given a disk image and you don't know what it contains then
145 you have to find out. Libguestfs can also do that: use
146 C<guestfs_list_partitions> and C<guestfs_lvs> to list possible
147 partitions and LVs, and either try mounting each to see what is
148 mountable, or else examine them with C<guestfs_file>. But you might
149 find it easier to look at higher level programs built on top of
150 libguestfs, in particular L<virt-inspector(1)>.
152 To mount a disk image read-only, use C<guestfs_mount_ro>. There are
153 several other variations of the C<guestfs_mount_*> call.
155 =head2 FILESYSTEM ACCESS AND MODIFICATION
157 The majority of the libguestfs API consists of fairly low-level calls
158 for accessing and modifying the files, directories, symlinks etc on
159 mounted filesystems. There are over a hundred such calls which you
160 can find listed in detail below in this man page, and we don't even
161 pretend to cover them all in this overview.
163 Specify filenames as full paths including the mount point.
165 For example, if you mounted a filesystem at C<"/"> and you want to
166 read the file called C<"etc/passwd"> then you could do:
168 char *data = guestfs_cat (handle, "/etc/passwd");
170 This would return C<data> as a newly allocated buffer containing the
171 full content of that file (with some conditions: see also
172 L</DOWNLOADING> below), or C<NULL> if there was an error.
174 As another example, to create a top-level directory on that filesystem
175 called C<"var"> you would do:
177 guestfs_mkdir (handle, "/var");
179 To create a symlink you could do:
181 guestfs_ln_s (handle, "/etc/init.d/portmap",
182 "/etc/rc3.d/S30portmap");
184 Libguestfs will reject attempts to use relative paths. There is no
185 concept of a current working directory. Libguestfs can return errors
186 in many situations: for example if the filesystem isn't writable, or
187 if a file or directory that you requested doesn't exist. If you are
188 using the C API (documented here) you have to check for those error
189 conditions after each call. (Other language bindings turn these
190 errors into exceptions).
192 File writes are affected by the per-handle umask, set by calling
193 C<guestfs_umask> and defaulting to 022.
197 Libguestfs contains API calls to read, create and modify partition
198 tables on disk images.
200 In the common case where you want to create a single partition
201 covering the whole disk, you should use the C<guestfs_part_disk>
204 const char *parttype = "mbr";
205 if (disk_is_larger_than_2TB)
207 guestfs_part_disk (g, "/dev/sda", parttype);
209 Obviously this effectively wipes anything that was on that disk image
212 In general MBR partitions are both unnecessarily complicated and
213 depend on archaic details, namely the Cylinder-Head-Sector (CHS)
214 geometry of the disk. C<guestfs_sfdiskM> can be used to
215 create more complex arrangements where the relative sizes are
216 expressed in megabytes instead of cylinders, which is a small win.
217 C<guestfs_sfdiskM> will choose the nearest cylinder to approximate the
218 requested size. There's a lot of crazy stuff to do with IDE and
219 virtio disks having different, incompatible CHS geometries, that you
220 probably don't want to know about.
222 My advice: make a single partition to cover the whole disk, then use
227 Libguestfs provides access to a large part of the LVM2 API, such as
228 C<guestfs_lvcreate> and C<guestfs_vgremove>. It won't make much sense
229 unless you familiarize yourself with the concepts of physical volumes,
230 volume groups and logical volumes.
232 This author strongly recommends reading the LVM HOWTO, online at
233 L<http://tldp.org/HOWTO/LVM-HOWTO/>.
237 Use C<guestfs_cat> to download small, text only files. This call
238 is limited to files which are less than 2 MB and which cannot contain
239 any ASCII NUL (C<\0>) characters. However it has a very simple
242 C<guestfs_read_file> can be used to read files which contain
243 arbitrary 8 bit data, since it returns a (pointer, size) pair.
244 However it is still limited to "small" files, less than 2 MB.
246 C<guestfs_download> can be used to download any file, with no
247 limits on content or size (even files larger than 4 GB).
249 To download multiple files, see C<guestfs_tar_out> and
254 It's often the case that you want to write a file or files to the disk
257 For small, single files, use C<guestfs_write_file>. This call
258 currently contains a bug which limits the call to plain text files
259 (not containing ASCII NUL characters).
261 To upload a single file, use C<guestfs_upload>. This call has no
262 limits on file content or size (even files larger than 4 GB).
264 To upload multiple files, see C<guestfs_tar_in> and C<guestfs_tgz_in>.
266 However the fastest way to upload I<large numbers of arbitrary files>
267 is to turn them into a squashfs or CD ISO (see L<mksquashfs(8)> and
268 L<mkisofs(8)>), then attach this using C<guestfs_add_drive_ro>. If
269 you add the drive in a predictable way (eg. adding it last after all
270 other drives) then you can get the device name from
271 C<guestfs_list_devices> and mount it directly using
272 C<guestfs_mount_ro>. Note that squashfs images are sometimes
273 non-portable between kernel versions, and they don't support labels or
274 UUIDs. If you want to pre-build an image or you need to mount it
275 using a label or UUID, use an ISO image instead.
279 C<guestfs_ll> is just designed for humans to read (mainly when using
280 the L<guestfish(1)>-equivalent command C<ll>).
282 C<guestfs_ls> is a quick way to get a list of files in a directory
283 from programs, as a flat list of strings.
285 C<guestfs_readdir> is a programmatic way to get a list of files in a
286 directory, plus additional information about each one. It is more
287 equivalent to using the L<readdir(3)> call on a local filesystem.
289 C<guestfs_find> can be used to recursively list files.
291 =head2 RUNNING COMMANDS
293 Although libguestfs is a primarily an API for manipulating files
294 inside guest images, we also provide some limited facilities for
295 running commands inside guests.
297 There are many limitations to this:
303 The kernel version that the command runs under will be different
304 from what it expects.
308 If the command needs to communicate with daemons, then most likely
309 they won't be running.
313 The command will be running in limited memory.
317 Only supports Linux guests (not Windows, BSD, etc).
321 Architecture limitations (eg. won't work for a PPC guest on
326 For SELinux guests, you may need to enable SELinux and load policy
327 first. See L</SELINUX> in this manpage.
331 The two main API calls to run commands are C<guestfs_command> and
332 C<guestfs_sh> (there are also variations).
334 The difference is that C<guestfs_sh> runs commands using the shell, so
335 any shell globs, redirections, etc will work.
337 =head2 CONFIGURATION FILES
339 To read and write configuration files in Linux guest filesystems, we
340 strongly recommend using Augeas. For example, Augeas understands how
341 to read and write, say, a Linux shadow password file or X.org
342 configuration file, and so avoids you having to write that code.
344 The main Augeas calls are bound through the C<guestfs_aug_*> APIs. We
345 don't document Augeas itself here because there is excellent
346 documentation on the L<http://augeas.net/> website.
348 If you don't want to use Augeas (you fool!) then try calling
349 C<guestfs_read_lines> to get the file as a list of lines which
350 you can iterate over.
354 We support SELinux guests. To ensure that labeling happens correctly
355 in SELinux guests, you need to enable SELinux and load the guest's
362 Before launching, do:
364 guestfs_set_selinux (g, 1);
368 After mounting the guest's filesystem(s), load the policy. This
369 is best done by running the L<load_policy(8)> command in the
372 guestfs_sh (g, "/usr/sbin/load_policy");
374 (Older versions of C<load_policy> require you to specify the
375 name of the policy file).
379 Optionally, set the security context for the API. The correct
380 security context to use can only be known by inspecting the
381 guest. As an example:
383 guestfs_setcon (g, "unconfined_u:unconfined_r:unconfined_t:s0");
387 This will work for running commands and editing existing files.
389 When new files are created, you may need to label them explicitly,
390 for example by running the external command
391 C<restorecon pathname>.
393 =head2 SPECIAL CONSIDERATIONS FOR WINDOWS GUESTS
395 Libguestfs can mount NTFS partitions. It does this using the
396 L<http://www.ntfs-3g.org/> driver.
398 DOS and Windows still use drive letters, and the filesystems are
399 always treated as case insensitive by Windows itself, and therefore
400 you might find a Windows configuration file referring to a path like
401 C<c:\windows\system32>. When the filesystem is mounted in libguestfs,
402 that directory might be referred to as C</WINDOWS/System32>.
404 Drive letter mappings are outside the scope of libguestfs. You have
405 to use libguestfs to read the appropriate Windows Registry and
406 configuration files, to determine yourself how drives are mapped (see
407 also L<virt-inspector(1)>).
409 Replacing backslash characters with forward slash characters is also
410 outside the scope of libguestfs, but something that you can easily do.
412 Where we can help is in resolving the case insensitivity of paths.
413 For this, call C<guestfs_case_sensitive_path>.
415 Libguestfs also provides some help for decoding Windows Registry
416 "hive" files, through the library C<libhivex> which is part of
417 libguestfs. You have to locate and download the hive file(s)
418 yourself, and then pass them to C<libhivex> functions. See also the
419 programs L<hivexml(1)>, L<hivexget(1)> and L<virt-win-reg(1)> for more
422 =head1 CONNECTION MANAGEMENT
426 C<guestfs_h> is the opaque type representing a connection handle.
427 Create a handle by calling C<guestfs_create>. Call C<guestfs_close>
428 to free the handle and release all resources used.
430 For information on using multiple handles and threads, see the section
431 L</MULTIPLE HANDLES AND MULTIPLE THREADS> below.
433 =head2 guestfs_create
435 guestfs_h *guestfs_create (void);
437 Create a connection handle.
439 You have to call C<guestfs_add_drive> on the handle at least once.
441 This function returns a non-NULL pointer to a handle on success or
444 After configuring the handle, you have to call C<guestfs_launch>.
446 You may also want to configure error handling for the handle. See
447 L</ERROR HANDLING> section below.
451 void guestfs_close (guestfs_h *handle);
453 This closes the connection handle and frees up all resources used.
455 =head1 ERROR HANDLING
457 The convention in all functions that return C<int> is that they return
458 C<-1> to indicate an error. You can get additional information on
459 errors by calling C<guestfs_last_error> and/or by setting up an error
460 handler with C<guestfs_set_error_handler>.
462 The default error handler prints the information string to C<stderr>.
464 Out of memory errors are handled differently. The default action is
465 to call L<abort(3)>. If this is undesirable, then you can set a
466 handler using C<guestfs_set_out_of_memory_handler>.
468 =head2 guestfs_last_error
470 const char *guestfs_last_error (guestfs_h *handle);
472 This returns the last error message that happened on C<handle>. If
473 there has not been an error since the handle was created, then this
476 The lifetime of the returned string is until the next error occurs, or
477 C<guestfs_close> is called.
479 The error string is not localized (ie. is always in English), because
480 this makes searching for error messages in search engines give the
481 largest number of results.
483 =head2 guestfs_set_error_handler
485 typedef void (*guestfs_error_handler_cb) (guestfs_h *handle,
488 void guestfs_set_error_handler (guestfs_h *handle,
489 guestfs_error_handler_cb cb,
492 The callback C<cb> will be called if there is an error. The
493 parameters passed to the callback are an opaque data pointer and the
494 error message string.
496 Note that the message string C<msg> is freed as soon as the callback
497 function returns, so if you want to stash it somewhere you must make
500 The default handler prints messages on C<stderr>.
502 If you set C<cb> to C<NULL> then I<no> handler is called.
504 =head2 guestfs_get_error_handler
506 guestfs_error_handler_cb guestfs_get_error_handler (guestfs_h *handle,
509 Returns the current error handler callback.
511 =head2 guestfs_set_out_of_memory_handler
513 typedef void (*guestfs_abort_cb) (void);
514 int guestfs_set_out_of_memory_handler (guestfs_h *handle,
517 The callback C<cb> will be called if there is an out of memory
518 situation. I<Note this callback must not return>.
520 The default is to call L<abort(3)>.
522 You cannot set C<cb> to C<NULL>. You can't ignore out of memory
525 =head2 guestfs_get_out_of_memory_handler
527 guestfs_abort_fn guestfs_get_out_of_memory_handler (guestfs_h *handle);
529 This returns the current out of memory handler.
533 Libguestfs needs a kernel and initrd.img, which it finds by looking
534 along an internal path.
536 By default it looks for these in the directory C<$libdir/guestfs>
537 (eg. C</usr/local/lib/guestfs> or C</usr/lib64/guestfs>).
539 Use C<guestfs_set_path> or set the environment variable
540 C<LIBGUESTFS_PATH> to change the directories that libguestfs will
541 search in. The value is a colon-separated list of paths. The current
542 directory is I<not> searched unless the path contains an empty element
543 or C<.>. For example C<LIBGUESTFS_PATH=:/usr/lib/guestfs> would
544 search the current directory and then C</usr/lib/guestfs>.
546 =head1 HIGH-LEVEL API ACTIONS
550 We guarantee the libguestfs ABI (binary interface), for public,
551 high-level actions as outlined in this section. Although we will
552 deprecate some actions, for example if they get replaced by newer
553 calls, we will keep the old actions forever. This allows you the
554 developer to program in confidence against libguestfs.
562 =head1 STATE MACHINE AND LOW-LEVEL EVENT API
564 Internally, libguestfs is implemented by running a virtual machine
565 using L<qemu(1)>. QEmu runs as a child process of the main program,
566 and most of this discussion won't make sense unless you understand
567 that the complexity is dealing with the (asynchronous) actions of the
571 ___________________ _________________________
573 | main program | | qemu +-----------------+|
574 | | | | Linux kernel ||
575 +-------------------+ | +-----------------+|
576 | libguestfs <-------------->| guestfsd ||
577 | | | +-----------------+|
578 \___________________/ \_________________________/
580 The diagram above shows libguestfs communicating with the guestfsd
581 daemon running inside the qemu child process. There are several
582 points of failure here: qemu can fail to start, the virtual machine
583 inside qemu can fail to boot, guestfsd can fail to start or not
584 establish communication, any component can start successfully but fail
585 asynchronously later, and so on.
589 libguestfs uses a state machine to model the child process:
600 / | \ \ guestfs_launch
611 \______/ <------ \________/
613 The normal transitions are (1) CONFIG (when the handle is created, but
614 there is no child process), (2) LAUNCHING (when the child process is
615 booting up), (3) alternating between READY and BUSY as commands are
616 issued to, and carried out by, the child process.
618 The guest may be killed by C<guestfs_kill_subprocess>, or may die
619 asynchronously at any time (eg. due to some internal error), and that
620 causes the state to transition back to CONFIG.
622 Configuration commands for qemu such as C<guestfs_add_drive> can only
623 be issued when in the CONFIG state.
625 The high-level API offers two calls that go from CONFIG through
626 LAUNCHING to READY. C<guestfs_launch> blocks until the child process
627 is READY to accept commands (or until some failure or timeout).
628 C<guestfs_launch> internally moves the state from CONFIG to LAUNCHING
631 High-level API actions such as C<guestfs_mount> can only be issued
632 when in the READY state. These high-level API calls block waiting for
633 the command to be carried out (ie. the state to transition to BUSY and
634 then back to READY). But using the low-level event API, you get
635 non-blocking versions. (But you can still only carry out one
636 operation per handle at a time - that is a limitation of the
637 communications protocol we use).
639 Finally, the child process sends asynchronous messages back to the
640 main program, such as kernel log messages. Mostly these are ignored
641 by the high-level API, but using the low-level event API you can
642 register to receive these messages.
644 =head2 SETTING CALLBACKS TO HANDLE EVENTS
646 The child process generates events in some situations. Current events
647 include: receiving a log message, the child process exits.
649 Use the C<guestfs_set_*_callback> functions to set a callback for
650 different types of events.
652 Only I<one callback of each type> can be registered for each handle.
653 Calling C<guestfs_set_*_callback> again overwrites the previous
654 callback of that type. Cancel all callbacks of this type by calling
655 this function with C<cb> set to C<NULL>.
657 =head2 guestfs_set_log_message_callback
659 typedef void (*guestfs_log_message_cb) (guestfs_h *g, void *opaque,
661 void guestfs_set_log_message_callback (guestfs_h *handle,
662 guestfs_log_message_cb cb,
665 The callback function C<cb> will be called whenever qemu or the guest
666 writes anything to the console.
668 Use this function to capture kernel messages and similar.
670 Normally there is no log message handler, and log messages are just
673 =head2 guestfs_set_subprocess_quit_callback
675 typedef void (*guestfs_subprocess_quit_cb) (guestfs_h *g, void *opaque);
676 void guestfs_set_subprocess_quit_callback (guestfs_h *handle,
677 guestfs_subprocess_quit_cb cb,
680 The callback function C<cb> will be called when the child process
681 quits, either asynchronously or if killed by
682 C<guestfs_kill_subprocess>. (This corresponds to a transition from
683 any state to the CONFIG state).
685 =head2 guestfs_set_launch_done_callback
687 typedef void (*guestfs_launch_done_cb) (guestfs_h *g, void *opaque);
688 void guestfs_set_launch_done_callback (guestfs_h *handle,
692 The callback function C<cb> will be called when the child process
693 becomes ready first time after it has been launched. (This
694 corresponds to a transition from LAUNCHING to the READY state).
696 =head1 BLOCK DEVICE NAMING
698 In the kernel there is now quite a profusion of schemata for naming
699 block devices (in this context, by I<block device> I mean a physical
700 or virtual hard drive). The original Linux IDE driver used names
701 starting with C</dev/hd*>. SCSI devices have historically used a
702 different naming scheme, C</dev/sd*>. When the Linux kernel I<libata>
703 driver became a popular replacement for the old IDE driver
704 (particularly for SATA devices) those devices also used the
705 C</dev/sd*> scheme. Additionally we now have virtual machines with
706 paravirtualized drivers. This has created several different naming
707 systems, such as C</dev/vd*> for virtio disks and C</dev/xvd*> for Xen
710 As discussed above, libguestfs uses a qemu appliance running an
711 embedded Linux kernel to access block devices. We can run a variety
712 of appliances based on a variety of Linux kernels.
714 This causes a problem for libguestfs because many API calls use device
715 or partition names. Working scripts and the recipe (example) scripts
716 that we make available over the internet could fail if the naming
719 Therefore libguestfs defines C</dev/sd*> as the I<standard naming
720 scheme>. Internally C</dev/sd*> names are translated, if necessary,
721 to other names as required. For example, under RHEL 5 which uses the
722 C</dev/hd*> scheme, any device parameter C</dev/sda2> is translated to
723 C</dev/hda2> transparently.
725 Note that this I<only> applies to parameters. The
726 C<guestfs_list_devices>, C<guestfs_list_partitions> and similar calls
727 return the true names of the devices and partitions as known to the
730 =head2 ALGORITHM FOR BLOCK DEVICE NAME TRANSLATION
732 Usually this translation is transparent. However in some (very rare)
733 cases you may need to know the exact algorithm. Such cases include
734 where you use C<guestfs_config> to add a mixture of virtio and IDE
735 devices to the qemu-based appliance, so have a mixture of C</dev/sd*>
736 and C</dev/vd*> devices.
738 The algorithm is applied only to I<parameters> which are known to be
739 either device or partition names. Return values from functions such
740 as C<guestfs_list_devices> are never changed.
746 Is the string a parameter which is a device or partition name?
750 Does the string begin with C</dev/sd>?
754 Does the named device exist? If so, we use that device.
755 However if I<not> then we continue with this algorithm.
759 Replace initial C</dev/sd> string with C</dev/hd>.
761 For example, change C</dev/sda2> to C</dev/hda2>.
763 If that named device exists, use it. If not, continue.
767 Replace initial C</dev/sd> string with C</dev/vd>.
769 If that named device exists, use it. If not, return an error.
773 =head2 PORTABILITY CONCERNS
775 Although the standard naming scheme and automatic translation is
776 useful for simple programs and guestfish scripts, for larger programs
777 it is best not to rely on this mechanism.
779 Where possible for maximum future portability programs using
780 libguestfs should use these future-proof techniques:
786 Use C<guestfs_list_devices> or C<guestfs_list_partitions> to list
787 actual device names, and then use those names directly.
789 Since those device names exist by definition, they will never be
794 Use higher level ways to identify filesystems, such as LVM names,
795 UUIDs and filesystem labels.
801 =head2 COMMUNICATION PROTOCOL
803 Don't rely on using this protocol directly. This section documents
804 how it currently works, but it may change at any time.
806 The protocol used to talk between the library and the daemon running
807 inside the qemu virtual machine is a simple RPC mechanism built on top
808 of XDR (RFC 1014, RFC 1832, RFC 4506).
810 The detailed format of structures is in C<src/guestfs_protocol.x>
811 (note: this file is automatically generated).
813 There are two broad cases, ordinary functions that don't have any
814 C<FileIn> and C<FileOut> parameters, which are handled with very
815 simple request/reply messages. Then there are functions that have any
816 C<FileIn> or C<FileOut> parameters, which use the same request and
817 reply messages, but they may also be followed by files sent using a
820 =head3 ORDINARY FUNCTIONS (NO FILEIN/FILEOUT PARAMS)
822 For ordinary functions, the request message is:
824 total length (header + arguments,
825 but not including the length word itself)
826 struct guestfs_message_header (encoded as XDR)
827 struct guestfs_<foo>_args (encoded as XDR)
829 The total length field allows the daemon to allocate a fixed size
830 buffer into which it slurps the rest of the message. As a result, the
831 total length is limited to C<GUESTFS_MESSAGE_MAX> bytes (currently
832 4MB), which means the effective size of any request is limited to
833 somewhere under this size.
835 Note also that many functions don't take any arguments, in which case
836 the C<guestfs_I<foo>_args> is completely omitted.
838 The header contains the procedure number (C<guestfs_proc>) which is
839 how the receiver knows what type of args structure to expect, or none
842 The reply message for ordinary functions is:
844 total length (header + ret,
845 but not including the length word itself)
846 struct guestfs_message_header (encoded as XDR)
847 struct guestfs_<foo>_ret (encoded as XDR)
849 As above the C<guestfs_I<foo>_ret> structure may be completely omitted
850 for functions that return no formal return values.
852 As above the total length of the reply is limited to
853 C<GUESTFS_MESSAGE_MAX>.
855 In the case of an error, a flag is set in the header, and the reply
856 message is slightly changed:
858 total length (header + error,
859 but not including the length word itself)
860 struct guestfs_message_header (encoded as XDR)
861 struct guestfs_message_error (encoded as XDR)
863 The C<guestfs_message_error> structure contains the error message as a
866 =head3 FUNCTIONS THAT HAVE FILEIN PARAMETERS
868 A C<FileIn> parameter indicates that we transfer a file I<into> the
869 guest. The normal request message is sent (see above). However this
870 is followed by a sequence of file chunks.
872 total length (header + arguments,
873 but not including the length word itself,
874 and not including the chunks)
875 struct guestfs_message_header (encoded as XDR)
876 struct guestfs_<foo>_args (encoded as XDR)
877 sequence of chunks for FileIn param #0
878 sequence of chunks for FileIn param #1 etc.
880 The "sequence of chunks" is:
882 length of chunk (not including length word itself)
883 struct guestfs_chunk (encoded as XDR)
885 struct guestfs_chunk (encoded as XDR)
888 struct guestfs_chunk (with data.data_len == 0)
890 The final chunk has the C<data_len> field set to zero. Additionally a
891 flag is set in the final chunk to indicate either successful
892 completion or early cancellation.
894 At time of writing there are no functions that have more than one
895 FileIn parameter. However this is (theoretically) supported, by
896 sending the sequence of chunks for each FileIn parameter one after
897 another (from left to right).
899 Both the library (sender) I<and> the daemon (receiver) may cancel the
900 transfer. The library does this by sending a chunk with a special
901 flag set to indicate cancellation. When the daemon sees this, it
902 cancels the whole RPC, does I<not> send any reply, and goes back to
903 reading the next request.
905 The daemon may also cancel. It does this by writing a special word
906 C<GUESTFS_CANCEL_FLAG> to the socket. The library listens for this
907 during the transfer, and if it gets it, it will cancel the transfer
908 (it sends a cancel chunk). The special word is chosen so that even if
909 cancellation happens right at the end of the transfer (after the
910 library has finished writing and has started listening for the reply),
911 the "spurious" cancel flag will not be confused with the reply
914 This protocol allows the transfer of arbitrary sized files (no 32 bit
915 limit), and also files where the size is not known in advance
916 (eg. from pipes or sockets). However the chunks are rather small
917 (C<GUESTFS_MAX_CHUNK_SIZE>), so that neither the library nor the
918 daemon need to keep much in memory.
920 =head3 FUNCTIONS THAT HAVE FILEOUT PARAMETERS
922 The protocol for FileOut parameters is exactly the same as for FileIn
923 parameters, but with the roles of daemon and library reversed.
925 total length (header + ret,
926 but not including the length word itself,
927 and not including the chunks)
928 struct guestfs_message_header (encoded as XDR)
929 struct guestfs_<foo>_ret (encoded as XDR)
930 sequence of chunks for FileOut param #0
931 sequence of chunks for FileOut param #1 etc.
933 =head3 INITIAL MESSAGE
935 Because the underlying channel (QEmu -net channel) doesn't have any
936 sort of connection control, when the daemon launches it sends an
937 initial word (C<GUESTFS_LAUNCH_FLAG>) which indicates that the guest
938 and daemon is alive. This is what C<guestfs_launch> waits for.
940 =head1 MULTIPLE HANDLES AND MULTIPLE THREADS
942 All high-level libguestfs actions are synchronous. If you want
943 to use libguestfs asynchronously then you must create a thread.
945 Only use the handle from a single thread. Either use the handle
946 exclusively from one thread, or provide your own mutex so that two
947 threads cannot issue calls on the same handle at the same time.
951 If you want to compile your own qemu, run qemu from a non-standard
952 location, or pass extra arguments to qemu, then you can write a
953 shell-script wrapper around qemu.
955 There is one important rule to remember: you I<must C<exec qemu>> as
956 the last command in the shell script (so that qemu replaces the shell
957 and becomes the direct child of the libguestfs-using program). If you
958 don't do this, then the qemu process won't be cleaned up correctly.
960 Here is an example of a wrapper, where I have built my own copy of
964 qemudir=/home/rjones/d/qemu
965 exec $qemudir/x86_64-softmmu/qemu-system-x86_64 -L $qemudir/pc-bios "$@"
967 Save this script as C</tmp/qemu.wrapper> (or wherever), C<chmod +x>,
968 and then use it by setting the LIBGUESTFS_QEMU environment variable.
971 LIBGUESTFS_QEMU=/tmp/qemu.wrapper guestfish
973 Note that libguestfs also calls qemu with the -help and -version
974 options in order to determine features.
976 =head1 ENVIRONMENT VARIABLES
980 =item LIBGUESTFS_APPEND
982 Pass additional options to the guest kernel.
984 =item LIBGUESTFS_DEBUG
986 Set C<LIBGUESTFS_DEBUG=1> to enable verbose messages. This
987 has the same effect as calling C<guestfs_set_verbose (handle, 1)>.
989 =item LIBGUESTFS_MEMSIZE
991 Set the memory allocated to the qemu process, in megabytes. For
994 LIBGUESTFS_MEMSIZE=700
996 =item LIBGUESTFS_PATH
998 Set the path that libguestfs uses to search for kernel and initrd.img.
999 See the discussion of paths in section PATH above.
1001 =item LIBGUESTFS_QEMU
1003 Set the default qemu binary that libguestfs uses. If not set, then
1004 the qemu which was found at compile time by the configure script is
1007 See also L</QEMU WRAPPERS> above.
1009 =item LIBGUESTFS_TRACE
1011 Set C<LIBGUESTFS_TRACE=1> to enable command traces. This
1012 has the same effect as calling C<guestfs_set_trace (handle, 1)>.
1016 Location of temporary directory, defaults to C</tmp>.
1018 If libguestfs was compiled to use the supermin appliance then each
1019 handle will require rather a large amount of space in this directory
1020 for short periods of time (~ 80 MB). You can use C<$TMPDIR> to
1021 configure another directory to use in case C</tmp> is not large
1031 L<http://libguestfs.org/>.
1033 Tools with a similar purpose:
1042 To get a list of bugs against libguestfs use this link:
1044 L<https://bugzilla.redhat.com/buglist.cgi?component=libguestfs&product=Virtualization+Tools>
1046 To report a new bug against libguestfs use this link:
1048 L<https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virtualization+Tools>
1050 When reporting a bug, please check:
1056 That the bug hasn't been reported already.
1060 That you are testing a recent version.
1064 Describe the bug accurately, and give a way to reproduce it.
1068 Run libguestfs-test-tool and paste the B<complete, unedited>
1069 output into the bug report.
1075 Richard W.M. Jones (C<rjones at redhat dot com>)
1079 Copyright (C) 2009 Red Hat Inc.
1080 L<http://libguestfs.org/>
1082 This library is free software; you can redistribute it and/or
1083 modify it under the terms of the GNU Lesser General Public
1084 License as published by the Free Software Foundation; either
1085 version 2 of the License, or (at your option) any later version.
1087 This library is distributed in the hope that it will be useful,
1088 but WITHOUT ANY WARRANTY; without even the implied warranty of
1089 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1090 Lesser General Public License for more details.
1092 You should have received a copy of the GNU Lesser General Public
1093 License along with this library; if not, write to the Free Software
1094 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA