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_wait_ready (handle);
15 guestfs_mount (handle, "/dev/sda1", "/");
16 guestfs_touch (handle, "/hello");
17 guestfs_sync (handle);
18 guestfs_close (handle);
22 Libguestfs is a library for accessing and modifying guest disk images.
23 Amongst the things this is good for: making batch configuration
24 changes to guests, getting disk used/free statistics (see also:
25 virt-df), migrating between virtualization systems (see also:
26 virt-p2v), performing partial backups, performing partial guest
27 clones, cloning guests and changing registry/UUID/hostname info, and
30 Libguestfs uses Linux kernel and qemu code, and can access any type of
31 guest filesystem that Linux and qemu can, including but not limited
32 to: ext2/3/4, btrfs, FAT and NTFS, LVM, many different disk partition
33 schemes, qcow, qcow2, vmdk.
35 Libguestfs provides ways to enumerate guest storage (eg. partitions,
36 LVs, what filesystem is in each LV, etc.). It can also run commands
37 in the context of the guest. Also you can access filesystems over FTP.
39 Libguestfs is a library that can be linked with C and C++ management
40 programs (or management programs written in OCaml, Perl, Python, Ruby, Java
41 or Haskell). You can also use it from shell scripts or the command line.
43 You don't need to be root to use libguestfs, although obviously you do
44 need enough permissions to access the disk images.
46 =head1 CONNECTION MANAGEMENT
48 If you are using the high-level API, then you should call the
49 functions in the following order:
51 guestfs_h *handle = guestfs_create ();
53 guestfs_add_drive (handle, "guest.img");
54 /* call guestfs_add_drive additional times if the guest has
58 guestfs_launch (handle);
59 guestfs_wait_ready (handle);
61 /* now you can examine what partitions, LVs etc are available
62 * you have to mount / at least
64 guestfs_mount (handle, "/dev/sda1", "/");
66 /* now you can perform actions on the guest disk image */
67 guestfs_touch (handle, "/hello");
69 /* you only need to call guestfs_sync if you have made
70 * changes to the guest image
72 guestfs_sync (handle);
74 guestfs_close (handle);
76 C<guestfs_wait_ready> and all of the actions including C<guestfs_sync>
77 are blocking calls. You can use the low-level event API to do
78 non-blocking operations instead.
80 All functions that return integers, return C<-1> on error. See
81 section ERROR HANDLING below for how to handle errors.
85 C<guestfs_h> is the opaque type representing a connection handle.
86 Create a handle by calling C<guestfs_create>. Call C<guestfs_close>
87 to free the handle and release all resources used.
89 For information on using multiple handles and threads, see the section
90 MULTIPLE HANDLES AND MULTIPLE THREADS below.
94 guestfs_h *guestfs_create (void);
96 Create a connection handle.
98 You have to call C<guestfs_add_drive> on the handle at least once.
100 This function returns a non-NULL pointer to a handle on success or
103 After configuring the handle, you have to call C<guestfs_launch> and
104 C<guestfs_wait_ready>.
106 You may also want to configure error handling for the handle. See
107 ERROR HANDLING section below.
111 void guestfs_close (guestfs_h *handle);
113 This closes the connection handle and frees up all resources used.
115 =head1 ERROR HANDLING
117 The convention in all functions that return C<int> is that they return
118 C<-1> to indicate an error. You can get additional information on
119 errors by calling C<guestfs_last_error> and/or by setting up an error
120 handler with C<guestfs_set_error_handler>.
122 The default error handler prints the information string to C<stderr>.
124 Out of memory errors are handled differently. The default action is
125 to call L<abort(3)>. If this is undesirable, then you can set a
126 handler using C<guestfs_set_out_of_memory_handler>.
128 =head2 guestfs_last_error
130 const char *guestfs_last_error (guestfs_h *handle);
132 This returns the last error message that happened on C<handle>. If
133 there has not been an error since the handle was created, then this
136 The lifetime of the returned string is until the next error occurs, or
137 C<guestfs_close> is called.
139 The error string is not localized (ie. is always in English), because
140 this makes searching for error messages in search engines give the
141 largest number of results.
143 =head2 guestfs_set_error_handler
145 typedef void (*guestfs_error_handler_cb) (guestfs_h *handle,
148 void guestfs_set_error_handler (guestfs_h *handle,
149 guestfs_error_handler_cb cb,
152 The callback C<cb> will be called if there is an error. The
153 parameters passed to the callback are an opaque data pointer and the
154 error message string.
156 Note that the message string C<msg> is freed as soon as the callback
157 function returns, so if you want to stash it somewhere you must make
160 The default handler prints messages on C<stderr>.
162 If you set C<cb> to C<NULL> then I<no> handler is called.
164 =head2 guestfs_get_error_handler
166 guestfs_error_handler_cb guestfs_get_error_handler (guestfs_h *handle,
169 Returns the current error handler callback.
171 =head2 guestfs_set_out_of_memory_handler
173 typedef void (*guestfs_abort_cb) (void);
174 int guestfs_set_out_of_memory_handler (guestfs_h *handle,
177 The callback C<cb> will be called if there is an out of memory
178 situation. I<Note this callback must not return>.
180 The default is to call L<abort(3)>.
182 You cannot set C<cb> to C<NULL>. You can't ignore out of memory
185 =head2 guestfs_get_out_of_memory_handler
187 guestfs_abort_fn guestfs_get_out_of_memory_handler (guestfs_h *handle);
189 This returns the current out of memory handler.
193 Libguestfs needs a kernel and initrd.img, which it finds by looking
194 along an internal path.
196 By default it looks for these in the directory C<$libdir/guestfs>
197 (eg. C</usr/local/lib/guestfs> or C</usr/lib64/guestfs>).
199 Use C<guestfs_set_path> or set the environment variable
200 C<LIBGUESTFS_PATH> to change the directories that libguestfs will
201 search in. The value is a colon-separated list of paths. The current
202 directory is I<not> searched unless the path contains an empty element
203 or C<.>. For example C<LIBGUESTFS_PATH=:/usr/lib/guestfs> would
204 search the current directory and then C</usr/lib/guestfs>.
208 This section provides additional documentation for groups of API
209 calls, which may not be obvious from reading about the individual
214 Libguestfs provides access to a large part of the LVM2 API. It won't
215 make much sense unless you familiarize yourself with the concepts of
216 physical volumes, volume groups and logical volumes.
218 This author strongly recommends reading the LVM HOWTO, online at
219 L<http://tldp.org/HOWTO/LVM-HOWTO/>.
223 To create MBR-style (ie. normal PC) partitions use one of the
224 C<guestfs_sfdisk*> variants. These calls use the external
225 L<sfdisk(8)> command.
227 The simplest call is:
229 char *lines[] = { ",", NULL };
230 guestfs_sfdiskM (g, "/dev/sda", lines);
232 This will create a single partition on C</dev/sda> called
233 C</dev/sda1> covering the whole disk.
235 In general MBR partitions are both unnecessarily complicated and
236 depend on archaic details, namely the Cylinder-Head-Sector (CHS)
237 geometry of the disk. C<guestfs_sfdiskM> allows you to specify sizes
238 in megabytes instead of cylinders, which is a small win.
239 C<guestfs_sfdiskM> will choose the nearest cylinder to approximate the
240 requested size. There's a lot of crazy stuff to do with IDE and
241 virtio disks having different, incompatible CHS geometries, that you
242 probably don't want to know about. My advice: make a single partition
243 to cover the whole disk, then use LVM on top.
245 In future we aim to provide access to libparted.
249 For small, single files, use C<guestfs_write_file>. In some versions
250 of libguestfs there was a bug which limited this call to text files
251 (not containing ASCII NUL characters).
253 To upload a single file, use C<guestfs_upload>. This call has no
254 limits on file content or size (even files larger than 4 GB).
256 To upload multiple files, see C<guestfs_tar_in> and C<guestfs_tgz_in>.
258 However the fastest way to upload I<large numbers of arbitrary files>
259 is to turn them into a squashfs or CD ISO (see L<mksquashfs(8)> and
260 L<mkisofs(8)>), then attach this using C<guestfs_add_drive_ro>. If
261 you add the drive in a predictable way (eg. adding it last after all
262 other drives) then you can get the device name from
263 C<guestfs_list_devices> and mount it directly using
264 C<guestfs_mount_ro>. Note that squashfs images are sometimes
265 non-portable between kernel versions, and they don't support labels or
266 UUIDs. If you want to pre-build an image or you need to mount it
267 using a label or UUID, use an ISO image instead.
271 Use C<guestfs_cat> to download small, text only files. This call
272 is limited to files which are less than 2 MB and which cannot contain
273 any ASCII NUL (C<\0>) characters. However it has a very simple
276 C<guestfs_read_file> can be used to read files which contain
277 arbitrary 8 bit data, since it returns a (pointer, size) pair.
278 However it is still limited to "small" files, less than 2 MB.
280 C<guestfs_download> can be used to download any file, with no
281 limits on content or size (even files larger than 4 GB).
283 To download multiple files, see C<guestfs_tar_out> and
286 =head2 RUNNING COMMANDS
288 Although libguestfs is a primarily an API for manipulating files
289 inside guest images, we also provide some limited facilities for
290 running commands inside guests.
292 There are many limitations to this:
298 The kernel version that the command runs under will be different
299 from what it expects.
303 If the command needs to communicate with daemons, then most likely
304 they won't be running.
308 The command will be running in limited memory.
312 Only supports Linux guests (not Windows, BSD, etc).
316 Architecture limitations (eg. won't work for a PPC guest on
321 The two main API calls to run commands are C<guestfs_command> and
322 C<guestfs_sh> (there are also variations).
324 The difference is that C<guestfs_sh> runs commands using the shell, so
325 any shell globs, redirections, etc will work.
329 C<guestfs_ll> is just designed for humans to read (mainly when using
330 the L<guestfish(1)>-equivalent command C<ll>).
332 C<guestfs_ls> is a quick way to get a list of files in a directory
335 C<guestfs_readdir> is a programmatic way to get a list of files in a
336 directory, plus additional information about each one.
338 C<guestfs_find> can be used to recursively list files.
340 =head1 HIGH-LEVEL API ACTIONS
344 We guarantee the libguestfs ABI (binary interface), for public,
345 high-level actions as outlined in this section. Although we will
346 deprecate some actions, for example if they get replaced by newer
347 calls, we will keep the old actions forever. This allows you the
348 developer to program in confidence against libguestfs.
356 =head1 STATE MACHINE AND LOW-LEVEL EVENT API
358 Internally, libguestfs is implemented by running a virtual machine
359 using L<qemu(1)>. QEmu runs as a child process of the main program,
360 and most of this discussion won't make sense unless you understand
361 that the complexity is dealing with the (asynchronous) actions of the
365 ___________________ _________________________
367 | main program | | qemu +-----------------+|
368 | | | | Linux kernel ||
369 +-------------------+ | +-----------------+|
370 | libguestfs <-------------->| guestfsd ||
371 | | | +-----------------+|
372 \___________________/ \_________________________/
374 The diagram above shows libguestfs communicating with the guestfsd
375 daemon running inside the qemu child process. There are several
376 points of failure here: qemu can fail to start, the virtual machine
377 inside qemu can fail to boot, guestfsd can fail to start or not
378 establish communication, any component can start successfully but fail
379 asynchronously later, and so on.
383 libguestfs uses a state machine to model the child process:
394 / | \ \ guestfs_launch
400 / | guestfs_wait_ready
405 \______/ <------ \________/
407 The normal transitions are (1) CONFIG (when the handle is created, but
408 there is no child process), (2) LAUNCHING (when the child process is
409 booting up), (3) alternating between READY and BUSY as commands are
410 issued to, and carried out by, the child process.
412 The guest may be killed by C<guestfs_kill_subprocess>, or may die
413 asynchronously at any time (eg. due to some internal error), and that
414 causes the state to transition back to CONFIG.
416 Configuration commands for qemu such as C<guestfs_add_drive> can only
417 be issued when in the CONFIG state.
419 The high-level API offers two calls that go from CONFIG through
420 LAUNCHING to READY. C<guestfs_launch> is a non-blocking call that
421 starts up the child process, immediately moving from CONFIG to
422 LAUNCHING. C<guestfs_wait_ready> blocks until the child process is
423 READY to accept commands (or until some failure or timeout). The
424 low-level event API described below provides a non-blocking way to
425 replace C<guestfs_wait_ready>.
427 High-level API actions such as C<guestfs_mount> can only be issued
428 when in the READY state. These high-level API calls block waiting for
429 the command to be carried out (ie. the state to transition to BUSY and
430 then back to READY). But using the low-level event API, you get
431 non-blocking versions. (But you can still only carry out one
432 operation per handle at a time - that is a limitation of the
433 communications protocol we use).
435 Finally, the child process sends asynchronous messages back to the
436 main program, such as kernel log messages. Mostly these are ignored
437 by the high-level API, but using the low-level event API you can
438 register to receive these messages.
440 =head2 SETTING CALLBACKS TO HANDLE EVENTS
442 The child process generates events in some situations. Current events
443 include: receiving a reply message after some action, receiving a log
444 message, the child process exits, &c.
446 Use the C<guestfs_set_*_callback> functions to set a callback for
447 different types of events.
449 Only I<one callback of each type> can be registered for each handle.
450 Calling C<guestfs_set_*_callback> again overwrites the previous
451 callback of that type. Cancel all callbacks of this type by calling
452 this function with C<cb> set to C<NULL>.
454 =head2 NON-BLOCKING ACTIONS
456 XXX This section was documented in previous versions but never
457 implemented in a way which matched the documentation. For now I have
458 removed the documentation, pending a working implementation. See also
459 C<src/guestfs-actions.c> in the source.
462 =head2 guestfs_set_send_callback
464 typedef void (*guestfs_send_cb) (guestfs_h *g, void *opaque);
465 void guestfs_set_send_callback (guestfs_h *handle,
469 The callback function C<cb> will be called whenever a message
470 which is queued for sending, has been sent.
472 =head2 guestfs_set_reply_callback
474 typedef void (*guestfs_reply_cb) (guestfs_h *g, void *opaque, XDR *xdr);
475 void guestfs_set_reply_callback (guestfs_h *handle,
479 The callback function C<cb> will be called whenever a reply is
480 received from the child process. (This corresponds to a transition
481 from the BUSY state to the READY state).
483 Note that the C<xdr> that you get in the callback is in C<XDR_DECODE>
484 mode, and you need to consume it before you return from the callback
485 function (since it gets destroyed after).
487 =head2 guestfs_set_log_message_callback
489 typedef void (*guestfs_log_message_cb) (guestfs_h *g, void *opaque,
491 void guestfs_set_log_message_callback (guestfs_h *handle,
492 guestfs_log_message_cb cb,
495 The callback function C<cb> will be called whenever qemu or the guest
496 writes anything to the console.
498 Use this function to capture kernel messages and similar.
500 Normally there is no log message handler, and log messages are just
503 =head2 guestfs_set_subprocess_quit_callback
505 typedef void (*guestfs_subprocess_quit_cb) (guestfs_h *g, void *opaque);
506 void guestfs_set_subprocess_quit_callback (guestfs_h *handle,
507 guestfs_subprocess_quit_cb cb,
510 The callback function C<cb> will be called when the child process
511 quits, either asynchronously or if killed by
512 C<guestfs_kill_subprocess>. (This corresponds to a transition from
513 any state to the CONFIG state).
515 =head2 guestfs_set_launch_done_callback
517 typedef void (*guestfs_launch_done_cb) (guestfs_h *g, void *opaque);
518 void guestfs_set_launch_done_callback (guestfs_h *handle,
522 The callback function C<cb> will be called when the child process
523 becomes ready first time after it has been launched. (This
524 corresponds to a transition from LAUNCHING to the READY state).
526 You can use this instead of C<guestfs_wait_ready> to implement a
527 non-blocking wait for the child process to finish booting up.
529 =head2 EVENT MAIN LOOP
531 To use the low-level event API and/or to use handles from multiple
532 threads, you have to provide an event "main loop". You can write your
533 own, but if you don't want to write one, two types are provided for
538 =item libguestfs-select
540 A simple main loop that is implemented using L<select(2)>.
542 This is the default main loop for new guestfs handles, unless you
543 call C<guestfs_set_main_loop> after a handle is created.
545 =item libguestfs-glib
547 An implementation which can be used with GLib and GTK+ programs. You
548 can use this to write graphical (GTK+) programs which use libguestfs
549 without hanging during long or slow operations.
553 =head2 MULTIPLE HANDLES AND MULTIPLE THREADS
555 The support for multiple handles and multiple threads is modelled
556 after glib (although doesn't require glib, if you use the select-based
559 L<http://library.gnome.org/devel/glib/unstable/glib-The-Main-Event-Loop.html>
561 You will need to create one main loop for each thread that wants to
562 use libguestfs. Each guestfs handle should be confined to one thread.
563 If you try to pass guestfs handles between threads, you will get
566 If you only want to use guestfs handles from one thread in your
567 program, but your program has other threads doing other things, then
568 you don't need to do anything special.
570 =head2 SINGLE THREAD CASE
572 In the single thread case, there is a single select-based main loop
573 created for you. All guestfs handles will use this main loop to
574 execute high level API actions.
576 =head2 MULTIPLE THREADS CASE
578 In the multiple threads case, you will need to create a main loop for
579 each thread that wants to use libguestfs.
581 To create main loops for other threads, use
582 C<guestfs_create_main_loop> or C<guestfs_glib_create_main_loop>.
584 Then you will need to attach each handle to the thread-specific main
587 handle = guestfs_create ();
588 guestfs_set_main_loop (handle, main_loop_of_current_thread);
590 =head2 guestfs_set_main_loop
592 void guestfs_set_main_loop (guestfs_h *handle,
593 guestfs_main_loop *main_loop);
595 Sets the main loop used by high level API actions for this handle. By
596 default, the select-based main loop is used (see
597 C<guestfs_get_default_main_loop>).
599 You only need to use this in multi-threaded programs, where multiple
600 threads want to use libguestfs. Create a main loop for each thread,
601 then call this function.
603 You cannot pass guestfs handles between threads.
605 =head2 guestfs_get_main_loop
607 guestfs_main_loop *guestfs_get_main_loop (guestfs_h *handle);
609 Return the main loop used by C<handle>.
611 =head2 guestfs_get_default_main_loop
613 guestfs_main_loop *guestfs_get_default_main_loop (void);
615 Return the default select-based main loop.
617 =head2 guestfs_create_main_loop
619 guestfs_main_loop *guestfs_create_main_loop (void);
621 This creates a select-based main loop. You should create one main
622 loop for each additional thread that needs to use libguestfs.
624 =head2 guestfs_free_main_loop
626 void guestfs_free_main_loop (guestfs_main_loop *);
628 Free the select-based main loop which was previously allocated with
629 C<guestfs_create_main_loop>.
631 =head2 WRITING A CUSTOM MAIN LOOP
633 This isn't documented. Please see the libguestfs-select and
634 libguestfs-glib implementations.
636 =head1 BLOCK DEVICE NAMING
638 In the kernel there is now quite a profusion of schemata for naming
639 block devices (in this context, by I<block device> I mean a physical
640 or virtual hard drive). The original Linux IDE driver used names
641 starting with C</dev/hd*>. SCSI devices have historically used a
642 different naming scheme, C</dev/sd*>. When the Linux kernel I<libata>
643 driver became a popular replacement for the old IDE driver
644 (particularly for SATA devices) those devices also used the
645 C</dev/sd*> scheme. Additionally we now have virtual machines with
646 paravirtualized drivers. This has created several different naming
647 systems, such as C</dev/vd*> for virtio disks and C</dev/xvd*> for Xen
650 As discussed above, libguestfs uses a qemu appliance running an
651 embedded Linux kernel to access block devices. We can run a variety
652 of appliances based on a variety of Linux kernels.
654 This causes a problem for libguestfs because many API calls use device
655 or partition names. Working scripts and the recipe (example) scripts
656 that we make available over the internet could fail if the naming
659 Therefore libguestfs defines C</dev/sd*> as the I<standard naming
660 scheme>. Internally C</dev/sd*> names are translated, if necessary,
661 to other names as required. For example, under RHEL 5 which uses the
662 C</dev/hd*> scheme, any device parameter C</dev/sda2> is translated to
663 C</dev/hda2> transparently.
665 Note that this I<only> applies to parameters. The
666 C<guestfs_list_devices>, C<guestfs_list_partitions> and similar calls
667 return the true names of the devices and partitions as known to the
670 =head2 ALGORITHM FOR BLOCK DEVICE NAME TRANSLATION
672 Usually this translation is transparent. However in some (very rare)
673 cases you may need to know the exact algorithm. Such cases include
674 where you use C<guestfs_config> to add a mixture of virtio and IDE
675 devices to the qemu-based appliance, so have a mixture of C</dev/sd*>
676 and C</dev/vd*> devices.
678 The algorithm is applied only to I<parameters> which are known to be
679 either device or partition names. Return values from functions such
680 as C<guestfs_list_devices> are never changed.
686 Is the string a parameter which is a device or partition name?
690 Does the string begin with C</dev/sd>?
694 Does the named device exist? If so, we use that device.
695 However if I<not> then we continue with this algorithm.
699 Replace initial C</dev/sd> string with C</dev/hd>.
701 For example, change C</dev/sda2> to C</dev/hda2>.
703 If that named device exists, use it. If not, continue.
707 Replace initial C</dev/sd> string with C</dev/vd>.
709 If that named device exists, use it. If not, return an error.
713 =head2 PORTABILITY CONCERNS
715 Although the standard naming scheme and automatic translation is
716 useful for simple programs and guestfish scripts, for larger programs
717 it is best not to rely on this mechanism.
719 Where possible for maximum future portability programs using
720 libguestfs should use these future-proof techniques:
726 Use C<guestfs_list_devices> or C<guestfs_list_partitions> to list
727 actual device names, and then use those names directly.
729 Since those device names exist by definition, they will never be
734 Use higher level ways to identify filesystems, such as LVM names,
735 UUIDs and filesystem labels.
741 =head2 COMMUNICATION PROTOCOL
743 Don't rely on using this protocol directly. This section documents
744 how it currently works, but it may change at any time.
746 The protocol used to talk between the library and the daemon running
747 inside the qemu virtual machine is a simple RPC mechanism built on top
748 of XDR (RFC 1014, RFC 1832, RFC 4506).
750 The detailed format of structures is in C<src/guestfs_protocol.x>
751 (note: this file is automatically generated).
753 There are two broad cases, ordinary functions that don't have any
754 C<FileIn> and C<FileOut> parameters, which are handled with very
755 simple request/reply messages. Then there are functions that have any
756 C<FileIn> or C<FileOut> parameters, which use the same request and
757 reply messages, but they may also be followed by files sent using a
760 =head3 ORDINARY FUNCTIONS (NO FILEIN/FILEOUT PARAMS)
762 For ordinary functions, the request message is:
764 total length (header + arguments,
765 but not including the length word itself)
766 struct guestfs_message_header (encoded as XDR)
767 struct guestfs_<foo>_args (encoded as XDR)
769 The total length field allows the daemon to allocate a fixed size
770 buffer into which it slurps the rest of the message. As a result, the
771 total length is limited to C<GUESTFS_MESSAGE_MAX> bytes (currently
772 4MB), which means the effective size of any request is limited to
773 somewhere under this size.
775 Note also that many functions don't take any arguments, in which case
776 the C<guestfs_I<foo>_args> is completely omitted.
778 The header contains the procedure number (C<guestfs_proc>) which is
779 how the receiver knows what type of args structure to expect, or none
782 The reply message for ordinary functions is:
784 total length (header + ret,
785 but not including the length word itself)
786 struct guestfs_message_header (encoded as XDR)
787 struct guestfs_<foo>_ret (encoded as XDR)
789 As above the C<guestfs_I<foo>_ret> structure may be completely omitted
790 for functions that return no formal return values.
792 As above the total length of the reply is limited to
793 C<GUESTFS_MESSAGE_MAX>.
795 In the case of an error, a flag is set in the header, and the reply
796 message is slightly changed:
798 total length (header + error,
799 but not including the length word itself)
800 struct guestfs_message_header (encoded as XDR)
801 struct guestfs_message_error (encoded as XDR)
803 The C<guestfs_message_error> structure contains the error message as a
806 =head3 FUNCTIONS THAT HAVE FILEIN PARAMETERS
808 A C<FileIn> parameter indicates that we transfer a file I<into> the
809 guest. The normal request message is sent (see above). However this
810 is followed by a sequence of file chunks.
812 total length (header + arguments,
813 but not including the length word itself,
814 and not including the chunks)
815 struct guestfs_message_header (encoded as XDR)
816 struct guestfs_<foo>_args (encoded as XDR)
817 sequence of chunks for FileIn param #0
818 sequence of chunks for FileIn param #1 etc.
820 The "sequence of chunks" is:
822 length of chunk (not including length word itself)
823 struct guestfs_chunk (encoded as XDR)
825 struct guestfs_chunk (encoded as XDR)
828 struct guestfs_chunk (with data.data_len == 0)
830 The final chunk has the C<data_len> field set to zero. Additionally a
831 flag is set in the final chunk to indicate either successful
832 completion or early cancellation.
834 At time of writing there are no functions that have more than one
835 FileIn parameter. However this is (theoretically) supported, by
836 sending the sequence of chunks for each FileIn parameter one after
837 another (from left to right).
839 Both the library (sender) I<and> the daemon (receiver) may cancel the
840 transfer. The library does this by sending a chunk with a special
841 flag set to indicate cancellation. When the daemon sees this, it
842 cancels the whole RPC, does I<not> send any reply, and goes back to
843 reading the next request.
845 The daemon may also cancel. It does this by writing a special word
846 C<GUESTFS_CANCEL_FLAG> to the socket. The library listens for this
847 during the transfer, and if it gets it, it will cancel the transfer
848 (it sends a cancel chunk). The special word is chosen so that even if
849 cancellation happens right at the end of the transfer (after the
850 library has finished writing and has started listening for the reply),
851 the "spurious" cancel flag will not be confused with the reply
854 This protocol allows the transfer of arbitrary sized files (no 32 bit
855 limit), and also files where the size is not known in advance
856 (eg. from pipes or sockets). However the chunks are rather small
857 (C<GUESTFS_MAX_CHUNK_SIZE>), so that neither the library nor the
858 daemon need to keep much in memory.
860 =head3 FUNCTIONS THAT HAVE FILEOUT PARAMETERS
862 The protocol for FileOut parameters is exactly the same as for FileIn
863 parameters, but with the roles of daemon and library reversed.
865 total length (header + ret,
866 but not including the length word itself,
867 and not including the chunks)
868 struct guestfs_message_header (encoded as XDR)
869 struct guestfs_<foo>_ret (encoded as XDR)
870 sequence of chunks for FileOut param #0
871 sequence of chunks for FileOut param #1 etc.
873 =head3 INITIAL MESSAGE
875 Because the underlying channel (QEmu -net channel) doesn't have any
876 sort of connection control, when the daemon launches it sends an
877 initial word (C<GUESTFS_LAUNCH_FLAG>) which indicates that the guest
878 and daemon is alive. This is what C<guestfs_wait_ready> waits for.
882 If you want to compile your own qemu, run qemu from a non-standard
883 location, or pass extra arguments to qemu, then you can write a
884 shell-script wrapper around qemu.
886 There is one important rule to remember: you I<must C<exec qemu>> as
887 the last command in the shell script (so that qemu replaces the shell
888 and becomes the direct child of the libguestfs-using program). If you
889 don't do this, then the qemu process won't be cleaned up correctly.
891 Here is an example of a wrapper, where I have built my own copy of
895 qemudir=/home/rjones/d/qemu
896 exec $qemudir/x86_64-softmmu/qemu-system-x86_64 -L $qemudir/pc-bios "$@"
898 Save this script as C</tmp/qemu.wrapper> (or wherever), C<chmod +x>,
899 and then use it by setting the LIBGUESTFS_QEMU environment variable.
902 LIBGUESTFS_QEMU=/tmp/qemu.wrapper guestfish
904 Note that libguestfs also calls qemu with the -help and -version
905 options in order to determine features.
907 =head1 ENVIRONMENT VARIABLES
911 =item LIBGUESTFS_APPEND
913 Pass additional options to the guest kernel.
915 =item LIBGUESTFS_DEBUG
917 Set C<LIBGUESTFS_DEBUG=1> to enable verbose messages. This
918 has the same effect as calling C<guestfs_set_verbose (handle, 1)>.
920 =item LIBGUESTFS_KERNEL
922 Override the ordinary selection of appliance kernel.
924 =item LIBGUESTFS_MEMSIZE
926 Set the memory allocated to the qemu process, in megabytes. For
929 LIBGUESTFS_MEMSIZE=700
931 =item LIBGUESTFS_PATH
933 Set the path that libguestfs uses to search for kernel and initrd.img.
934 See the discussion of paths in section PATH above.
936 =item LIBGUESTFS_QEMU
938 Set the default qemu binary that libguestfs uses. If not set, then
939 the qemu which was found at compile time by the configure script is
942 See also L<QEMU WRAPPERS> above.
946 Location of temporary directory, defaults to C</tmp>.
948 If libguestfs was compiled to use the supermin appliance then each
949 handle will require rather a large amount of space in this directory
950 for short periods of time (~ 80 MB). You can use C<$TMPDIR> to
951 configure another directory to use in case C</tmp> is not large
961 L<http://libguestfs.org/>.
965 To get a list of bugs against libguestfs use this link:
967 L<https://bugzilla.redhat.com/buglist.cgi?component=libguestfs&product=Virtualization+Tools>
969 To report a new bug against libguestfs use this link:
971 L<https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virtualization+Tools>
973 When reporting a bug, please check:
979 That the bug hasn't been reported already.
983 That you are testing a recent version.
987 Describe the bug accurately, and give a way to reproduce it.
993 Richard W.M. Jones (C<rjones at redhat dot com>)
997 Copyright (C) 2009 Red Hat Inc.
998 L<http://libguestfs.org/>
1000 This library is free software; you can redistribute it and/or
1001 modify it under the terms of the GNU Lesser General Public
1002 License as published by the Free Software Foundation; either
1003 version 2 of the License, or (at your option) any later version.
1005 This library is distributed in the hope that it will be useful,
1006 but WITHOUT ANY WARRANTY; without even the implied warranty of
1007 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1008 Lesser General Public License for more details.
1010 You should have received a copy of the GNU Lesser General Public
1011 License along with this library; if not, write to the Free Software
1012 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA