1 =head2 add-cdrom | cdrom
5 This function adds a virtual CD-ROM disk image to the guest.
7 This is equivalent to the qemu parameter C<-cdrom filename>.
13 This function adds a virtual machine disk image C<filename> to the
14 guest. The first time you call this function, the disk appears as IDE
15 disk 0 (C</dev/sda>) in the guest, the second time as C</dev/sdb>, and
18 You don't necessarily need to be root when using libguestfs. However
19 you obviously do need sufficient permissions to access the filename
20 for whatever operations you want to perform (ie. read access if you
21 just want to read the image or write access if you want to modify the
24 This is equivalent to the qemu parameter C<-drive file=filename>.
30 Close the current Augeas handle and free up any resources
31 used by it. After calling this, you have to call
32 C<aug-init> again before you can use any other
37 aug-defnode name expr val
39 Defines a variable C<name> whose value is the result of
42 If C<expr> evaluates to an empty nodeset, a node is created,
43 equivalent to calling C<aug-set> C<expr>, C<value>.
44 C<name> will be the nodeset containing that single node.
46 On success this returns a pair containing the
47 number of nodes in the nodeset, and a boolean flag
48 if a node was created.
54 Defines an Augeas variable C<name> whose value is the result
55 of evaluating C<expr>. If C<expr> is NULL, then C<name> is
58 On success this returns the number of nodes in C<expr>, or
59 C<0> if C<expr> evaluates to something which is not a nodeset.
65 Look up the value associated with C<path>. If C<path>
66 matches exactly one node, the C<value> is returned.
72 Create a new Augeas handle for editing configuration files.
73 If there was any previous Augeas handle associated with this
74 guestfs session, then it is closed.
76 You must call this before using any other C<aug-*>
79 C<root> is the filesystem root. C<root> must not be NULL,
82 The flags are the same as the flags defined in
83 E<lt>augeas.hE<gt>, the logical I<or> of the following
88 =item C<AUG_SAVE_BACKUP> = 1
90 Keep the original file with a C<.augsave> extension.
92 =item C<AUG_SAVE_NEWFILE> = 2
94 Save changes into a file with extension C<.augnew>, and
95 do not overwrite original. Overrides C<AUG_SAVE_BACKUP>.
97 =item C<AUG_TYPE_CHECK> = 4
99 Typecheck lenses (can be expensive).
101 =item C<AUG_NO_STDINC> = 8
103 Do not use standard load path for modules.
105 =item C<AUG_SAVE_NOOP> = 16
107 Make save a no-op, just record what would have been changed.
109 =item C<AUG_NO_LOAD> = 32
111 Do not load the tree in C<aug-init>.
115 To close the handle, you can call C<aug-close>.
117 To find out more about Augeas, see L<http://augeas.net/>.
121 aug-insert path label true|false
123 Create a new sibling C<label> for C<path>, inserting it into
124 the tree before or after C<path> (depending on the boolean
127 C<path> must match exactly one existing node in the tree, and
128 C<label> must be a label, ie. not contain C</>, C<*> or end
129 with a bracketed index C<[N]>.
135 Load files into the tree.
137 See C<aug_load> in the Augeas documentation for the full gory
144 This is just a shortcut for listing C<aug-match>
145 C<path/*> and sorting the resulting nodes into alphabetical order.
151 Returns a list of paths which match the path expression C<path>.
152 The returned paths are sufficiently qualified so that they match
153 exactly one node in the current tree.
159 Move the node C<src> to C<dest>. C<src> must match exactly
160 one node. C<dest> is overwritten if it exists.
166 Remove C<path> and all of its children.
168 On success this returns the number of entries which were removed.
174 This writes all pending changes to disk.
176 The flags which were passed to C<aug-init> affect exactly
183 Set the value associated with C<path> to C<value>.
185 =head2 blockdev-flushbufs
187 blockdev-flushbufs device
189 This tells the kernel to flush internal buffers associated
192 This uses the L<blockdev(8)> command.
194 =head2 blockdev-getbsz
196 blockdev-getbsz device
198 This returns the block size of a device.
200 (Note this is different from both I<size in blocks> and
201 I<filesystem block size>).
203 This uses the L<blockdev(8)> command.
205 =head2 blockdev-getro
207 blockdev-getro device
209 Returns a boolean indicating if the block device is read-only
210 (true if read-only, false if not).
212 This uses the L<blockdev(8)> command.
214 =head2 blockdev-getsize64
216 blockdev-getsize64 device
218 This returns the size of the device in bytes.
220 See also C<blockdev-getsz>.
222 This uses the L<blockdev(8)> command.
224 =head2 blockdev-getss
226 blockdev-getss device
228 This returns the size of sectors on a block device.
229 Usually 512, but can be larger for modern devices.
231 (Note, this is not the size in sectors, use C<blockdev-getsz>
234 This uses the L<blockdev(8)> command.
236 =head2 blockdev-getsz
238 blockdev-getsz device
240 This returns the size of the device in units of 512-byte sectors
241 (even if the sectorsize isn't 512 bytes ... weird).
243 See also C<blockdev-getss> for the real sector size of
244 the device, and C<blockdev-getsize64> for the more
245 useful I<size in bytes>.
247 This uses the L<blockdev(8)> command.
249 =head2 blockdev-rereadpt
251 blockdev-rereadpt device
253 Reread the partition table on C<device>.
255 This uses the L<blockdev(8)> command.
257 =head2 blockdev-setbsz
259 blockdev-setbsz device blocksize
261 This sets the block size of a device.
263 (Note this is different from both I<size in blocks> and
264 I<filesystem block size>).
266 This uses the L<blockdev(8)> command.
268 =head2 blockdev-setro
270 blockdev-setro device
272 Sets the block device named C<device> to read-only.
274 This uses the L<blockdev(8)> command.
276 =head2 blockdev-setrw
278 blockdev-setrw device
280 Sets the block device named C<device> to read-write.
282 This uses the L<blockdev(8)> command.
288 Return the contents of the file named C<path>.
290 Note that this function cannot correctly handle binary files
291 (specifically, files containing C<\0> character which is treated
292 as end of string). For those you need to use the C<download>
293 function which has a more complex interface.
295 Because of the message protocol, there is a transfer limit
296 of somewhere between 2MB and 4MB. To transfer large files you should use
301 checksum csumtype path
303 This call computes the MD5, SHAx or CRC checksum of the
306 The type of checksum to compute is given by the C<csumtype>
307 parameter which must have one of the following values:
313 Compute the cyclic redundancy check (CRC) specified by POSIX
314 for the C<cksum> command.
318 Compute the MD5 hash (using the C<md5sum> program).
322 Compute the SHA1 hash (using the C<sha1sum> program).
326 Compute the SHA224 hash (using the C<sha224sum> program).
330 Compute the SHA256 hash (using the C<sha256sum> program).
334 Compute the SHA384 hash (using the C<sha384sum> program).
338 Compute the SHA512 hash (using the C<sha512sum> program).
342 The checksum is returned as a printable string.
348 Change the mode (permissions) of C<path> to C<mode>. Only
349 numeric modes are supported.
353 chown owner group path
355 Change the file owner to C<owner> and group to C<group>.
357 Only numeric uid and gid are supported. If you want to use
358 names, you will need to locate and parse the password file
359 yourself (Augeas support makes this relatively easy).
363 command 'arguments ...'
365 This call runs a command from the guest filesystem. The
366 filesystem must be mounted, and must contain a compatible
367 operating system (ie. something Linux, with the same
368 or compatible processor architecture).
370 The single parameter is an argv-style list of arguments.
371 The first element is the name of the program to run.
372 Subsequent elements are parameters. The list must be
373 non-empty (ie. must contain a program name).
375 The C<$PATH> environment variable will contain at least
376 C</usr/bin> and C</bin>. If you require a program from
377 another location, you should provide the full path in the
380 Shared libraries and data files required by the program
381 must be available on filesystems which are mounted in the
382 correct places. It is the caller's responsibility to ensure
383 all filesystems that are needed are mounted at the right
388 command-lines 'arguments ...'
390 This is the same as C<command>, but splits the
391 result into a list of lines.
395 config qemuparam qemuvalue
397 This can be used to add arbitrary qemu command line parameters
398 of the form C<-param value>. Actually it's not quite arbitrary - we
399 prevent you from setting some parameters which would interfere with
400 parameters that we use.
402 The first character of C<param> string must be a C<-> (dash).
404 C<value> can be NULL.
410 This copies a file from C<src> to C<dest> where C<dest> is
411 either a destination filename or destination directory.
417 This copies a file or directory from C<src> to C<dest>
418 recursively using the C<cp -a> command.
422 debug subcmd 'extraargs ...'
424 The C<debug> command exposes some internals of
425 C<guestfsd> (the guestfs daemon) that runs inside the
428 There is no comprehensive help for this command. You have
429 to look at the file C<daemon/debug.c> in the libguestfs source
430 to find out what you can do.
436 This returns the kernel messages (C<dmesg> output) from
437 the guest kernel. This is sometimes useful for extended
438 debugging of problems.
440 Another way to get the same information is to enable
441 verbose messages with C<set-verbose> or by setting
442 the environment variable C<LIBGUESTFS_DEBUG=1> before
447 download remotefilename (filename|-)
449 Download file C<remotefilename> and save it as C<filename>
450 on the local machine.
452 C<filename> can also be a named pipe.
454 See also C<upload>, C<cat>.
456 Use C<-> instead of a filename to read/write from stdin/stdout.
460 drop-caches whattodrop
462 This instructs the guest kernel to drop its page cache,
463 and/or dentries and inode caches. The parameter C<whattodrop>
464 tells the kernel what precisely to drop, see
465 L<http://linux-mm.org/Drop_Caches>
467 Setting C<whattodrop> to 3 should drop everything.
469 This automatically calls L<sync(2)> before the operation,
470 so that the maximum guest memory is freed.
476 This returns C<true> if and only if there is a file, directory
477 (or anything) with the given C<path> name.
479 See also C<is-file>, C<is-dir>, C<stat>.
485 This call uses the standard L<file(1)> command to determine
486 the type or contents of the file. This also works on devices,
487 for example to find out whether a partition contains a filesystem.
489 The exact command which runs is C<file -bsL path>. Note in
490 particular that the filename is not prepended to the output
497 This runs the filesystem checker (fsck) on C<device> which
498 should have filesystem type C<fstype>.
500 The returned integer is the status. See L<fsck(8)> for the
501 list of status codes from C<fsck>.
509 Multiple status codes can be summed together.
513 A non-zero return code can mean "success", for example if
514 errors have been corrected on the filesystem.
518 Checking or repairing NTFS volumes is not supported
523 This command is entirely equivalent to running C<fsck -a -t fstype device>.
529 Get the autosync flag.
535 This returns the ext2/3/4 filesystem label of the filesystem on
542 This returns the ext2/3/4 filesystem UUID of the filesystem on
549 Return the current search path.
551 This is always non-NULL. If it wasn't set already, then this will
552 return the default path.
558 Return the current qemu binary.
560 This is always non-NULL. If it wasn't set already, then this will
561 return the default qemu binary name.
567 This returns the current state as an opaque integer. This is
568 only useful for printing debug and internal error messages.
570 For more information on states, see L<guestfs(3)>.
576 This returns the verbose messages flag.
580 grub-install root device
582 This command installs GRUB (the Grand Unified Bootloader) on
583 C<device>, with the root directory being C<root>.
589 This returns true iff this handle is busy processing a command
590 (in the C<BUSY> state).
592 For more information on states, see L<guestfs(3)>.
598 This returns true iff this handle is being configured
599 (in the C<CONFIG> state).
601 For more information on states, see L<guestfs(3)>.
607 This returns C<true> if and only if there is a directory
608 with the given C<path> name. Note that it returns false for
609 other objects like files.
617 This returns C<true> if and only if there is a file
618 with the given C<path> name. Note that it returns false for
619 other objects like directories.
627 This returns true iff this handle is launching the subprocess
628 (in the C<LAUNCHING> state).
630 For more information on states, see L<guestfs(3)>.
636 This returns true iff this handle is ready to accept commands
637 (in the C<READY> state).
639 For more information on states, see L<guestfs(3)>.
641 =head2 kill-subprocess
645 This kills the qemu subprocess. You should never need to call this.
651 Internally libguestfs is implemented by running a virtual machine
654 You should call this after configuring the handle
655 (eg. adding drives) but before performing any actions.
661 List all the block devices.
663 The full block device names are returned, eg. C</dev/sda>
665 =head2 list-partitions
669 List all the partitions detected on all block devices.
671 The full partition device names are returned, eg. C</dev/sda1>
673 This does not return logical volumes. For that you will need to
680 List the files in C<directory> (relative to the root directory,
681 there is no cwd) in the format of 'ls -la'.
683 This command is mostly useful for interactive sessions. It
684 is I<not> intended that you try to parse the output string.
690 List the files in C<directory> (relative to the root directory,
691 there is no cwd). The '.' and '..' entries are not returned, but
692 hidden files are shown.
694 This command is mostly useful for interactive sessions. Programs
695 should probably use C<readdir> instead.
701 Returns file information for the given C<path>.
703 This is the same as C<stat> except that if C<path>
704 is a symbolic link, then the link is stat-ed, not the file it
707 This is the same as the C<lstat(2)> system call.
711 lvcreate logvol volgroup mbytes
713 This creates an LVM volume group called C<logvol>
714 on the volume group C<volgroup>, with C<size> megabytes.
716 =head2 lvm-remove-all
720 This command removes all LVM logical volumes, volume groups
721 and physical volumes.
723 B<This command is dangerous. Without careful use you
724 can easily destroy all your data>.
730 Remove an LVM logical volume C<device>, where C<device> is
731 the path to the LV, such as C</dev/VG/LV>.
733 You can also remove all LVs in a volume group by specifying
734 the VG name, C</dev/VG>.
740 List all the logical volumes detected. This is the equivalent
741 of the L<lvs(8)> command.
743 This returns a list of the logical volume device names
744 (eg. C</dev/VolGroup00/LogVol00>).
746 See also C<lvs-full>.
752 List all the logical volumes detected. This is the equivalent
753 of the L<lvs(8)> command. The "full" version includes all fields.
759 Create a directory named C<path>.
765 Create a directory named C<path>, creating any parent directories
766 as necessary. This is like the C<mkdir -p> shell command.
772 This creates a filesystem on C<device> (usually a partition
773 of LVM logical volume). The filesystem type is C<fstype>, for
778 mount device mountpoint
780 Mount a guest disk at a position in the filesystem. Block devices
781 are named C</dev/sda>, C</dev/sdb> and so on, as they were added to
782 the guest. If those block devices contain partitions, they will have
783 the usual names (eg. C</dev/sda1>). Also LVM C</dev/VG/LV>-style
786 The rules are the same as for L<mount(2)>: A filesystem must
787 first be mounted on C</> before others can be mounted. Other
788 filesystems can only be mounted on directories which already
791 The mounted filesystem is writable, if we have sufficient permissions
792 on the underlying device.
794 The filesystem options C<sync> and C<noatime> are set with this
795 call, in order to improve reliability.
799 mount-options options device mountpoint
801 This is the same as the C<mount> command, but it
802 allows you to set the mount options as for the
803 L<mount(8)> I<-o> flag.
807 mount-ro device mountpoint
809 This is the same as the C<mount> command, but it
810 mounts the filesystem with the read-only (I<-o ro>) flag.
814 mount-vfs options vfstype device mountpoint
816 This is the same as the C<mount> command, but it
817 allows you to set both the mount options and the vfstype
818 as for the L<mount(8)> I<-o> and I<-t> flags.
824 This returns the list of currently mounted filesystems. It returns
825 the list of devices (eg. C</dev/sda1>, C</dev/VG/LV>).
827 Some internal mounts are not shown.
833 This moves a file from C<src> to C<dest> where C<dest> is
834 either a destination filename or destination directory.
840 This is a test probe into the guestfs daemon running inside
841 the qemu subprocess. Calling this function checks that the
842 daemon responds to the ping message, without affecting the daemon
843 or attached block device(s) in any other way.
849 This creates an LVM physical volume on the named C<device>,
850 where C<device> should usually be a partition name such
857 This wipes a physical volume C<device> so that LVM will no longer
860 The implementation uses the C<pvremove> command which refuses to
861 wipe physical volumes that contain any volume groups, so you have
862 to remove those first.
868 List all the physical volumes detected. This is the equivalent
869 of the L<pvs(8)> command.
871 This returns a list of just the device names that contain
872 PVs (eg. C</dev/sda2>).
874 See also C<pvs-full>.
880 List all the physical volumes detected. This is the equivalent
881 of the L<pvs(8)> command. The "full" version includes all fields.
887 Return the contents of the file named C<path>.
889 The file contents are returned as a list of lines. Trailing
890 C<LF> and C<CRLF> character sequences are I<not> returned.
892 Note that this function cannot correctly handle binary files
893 (specifically, files containing C<\0> character which is treated
894 as end of line). For those you need to use the C<read-file>
895 function which has a more complex interface.
901 Remove the single file C<path>.
907 Remove the file or directory C<path>, recursively removing the
908 contents if its a directory. This is like the C<rm -rf> shell
915 Remove the single directory C<path>.
917 =head2 set-autosync | autosync
919 set-autosync true|false
921 If C<autosync> is true, this enables autosync. Libguestfs will make a
922 best effort attempt to run C<umount-all> followed by
923 C<sync> when the handle is closed
924 (also if the program exits without closing handles).
926 This is disabled by default (except in guestfish where it is
931 set-e2label device label
933 This sets the ext2/3/4 filesystem label of the filesystem on
934 C<device> to C<label>. Filesystem labels are limited to
937 You can use either C<tune2fs-l> or C<get-e2label>
938 to return the existing label on a filesystem.
942 set-e2uuid device uuid
944 This sets the ext2/3/4 filesystem UUID of the filesystem on
945 C<device> to C<uuid>. The format of the UUID and alternatives
946 such as C<clear>, C<random> and C<time> are described in the
947 L<tune2fs(8)> manpage.
949 You can use either C<tune2fs-l> or C<get-e2uuid>
950 to return the existing UUID of a filesystem.
952 =head2 set-path | path
956 Set the path that libguestfs searches for kernel and initrd.img.
958 The default is C<$libdir/guestfs> unless overridden by setting
959 C<LIBGUESTFS_PATH> environment variable.
961 The string C<path> is stashed in the libguestfs handle, so the caller
962 must make sure it remains valid for the lifetime of the handle.
964 Setting C<path> to C<NULL> restores the default path.
966 =head2 set-qemu | qemu
970 Set the qemu binary that we will use.
972 The default is chosen when the library was compiled by the
975 You can also override this by setting the C<LIBGUESTFS_QEMU>
976 environment variable.
978 The string C<qemu> is stashed in the libguestfs handle, so the caller
979 must make sure it remains valid for the lifetime of the handle.
981 Setting C<qemu> to C<NULL> restores the default qemu binary.
983 =head2 set-verbose | verbose
985 set-verbose true|false
987 If C<verbose> is true, this turns on verbose messages (to C<stderr>).
989 Verbose messages are disabled unless the environment variable
990 C<LIBGUESTFS_DEBUG> is defined and set to C<1>.
994 sfdisk device cyls heads sectors 'lines ...'
996 This is a direct interface to the L<sfdisk(8)> program for creating
997 partitions on block devices.
999 C<device> should be a block device, for example C</dev/sda>.
1001 C<cyls>, C<heads> and C<sectors> are the number of cylinders, heads
1002 and sectors on the device, which are passed directly to sfdisk as
1003 the I<-C>, I<-H> and I<-S> parameters. If you pass C<0> for any
1004 of these, then the corresponding parameter is omitted. Usually for
1005 'large' disks, you can just pass C<0> for these, but for small
1006 (floppy-sized) disks, sfdisk (or rather, the kernel) cannot work
1007 out the right geometry and you will need to tell it.
1009 C<lines> is a list of lines that we feed to C<sfdisk>. For more
1010 information refer to the L<sfdisk(8)> manpage.
1012 To create a single partition occupying the whole disk, you would
1013 pass C<lines> as a single element list, when the single element being
1014 the string C<,> (comma).
1016 B<This command is dangerous. Without careful use you
1017 can easily destroy all your data>.
1023 Returns file information for the given C<path>.
1025 This is the same as the C<stat(2)> system call.
1031 Returns file system statistics for any mounted file system.
1032 C<path> should be a file or directory in the mounted file system
1033 (typically it is the mount point itself, but it doesn't need to be).
1035 This is the same as the C<statvfs(2)> system call.
1041 This syncs the disk, so that any writes are flushed through to the
1042 underlying disk image.
1044 You should always call this if you have modified a disk image, before
1049 tar-in (tarfile|-) directory
1051 This command uploads and unpacks local file C<tarfile> (an
1052 I<uncompressed> tar file) into C<directory>.
1054 To upload a compressed tarball, use C<tgz-in>.
1056 Use C<-> instead of a filename to read/write from stdin/stdout.
1060 tar-out directory (tarfile|-)
1062 This command packs the contents of C<directory> and downloads
1063 it to local file C<tarfile>.
1065 To download a compressed tarball, use C<tgz-out>.
1067 Use C<-> instead of a filename to read/write from stdin/stdout.
1071 tgz-in (tarball|-) directory
1073 This command uploads and unpacks local file C<tarball> (a
1074 I<gzip compressed> tar file) into C<directory>.
1076 To upload an uncompressed tarball, use C<tar-in>.
1078 Use C<-> instead of a filename to read/write from stdin/stdout.
1082 tgz-out directory (tarball|-)
1084 This command packs the contents of C<directory> and downloads
1085 it to local file C<tarball>.
1087 To download an uncompressed tarball, use C<tar-out>.
1089 Use C<-> instead of a filename to read/write from stdin/stdout.
1095 Touch acts like the L<touch(1)> command. It can be used to
1096 update the timestamps on a file, or, if the file does not exist,
1097 to create a new zero-length file.
1103 This returns the contents of the ext2, ext3 or ext4 filesystem
1104 superblock on C<device>.
1106 It is the same as running C<tune2fs -l device>. See L<tune2fs(8)>
1107 manpage for more details. The list of fields returned isn't
1108 clearly defined, and depends on both the version of C<tune2fs>
1109 that libguestfs was built against, and the filesystem itself.
1111 =head2 umount | unmount
1115 This unmounts the given filesystem. The filesystem may be
1116 specified either by its mountpoint (path) or the device which
1117 contains the filesystem.
1119 =head2 umount-all | unmount-all
1123 This unmounts all mounted filesystems.
1125 Some internal mounts are not unmounted by this call.
1129 upload (filename|-) remotefilename
1131 Upload local file C<filename> to C<remotefilename> on the
1134 C<filename> can also be a named pipe.
1136 See also C<download>.
1138 Use C<-> instead of a filename to read/write from stdin/stdout.
1142 vgcreate volgroup 'physvols ...'
1144 This creates an LVM volume group called C<volgroup>
1145 from the non-empty list of physical volumes C<physvols>.
1151 Remove an LVM volume group C<vgname>, (for example C<VG>).
1153 This also forcibly removes all logical volumes in the volume
1160 List all the volumes groups detected. This is the equivalent
1161 of the L<vgs(8)> command.
1163 This returns a list of just the volume group names that were
1164 detected (eg. C<VolGroup00>).
1166 See also C<vgs-full>.
1172 List all the volumes groups detected. This is the equivalent
1173 of the L<vgs(8)> command. The "full" version includes all fields.
1177 write-file path content size
1179 This call creates a file called C<path>. The contents of the
1180 file is the string C<content> (which can contain any 8 bit data),
1181 with length C<size>.
1183 As a special case, if C<size> is C<0>
1184 then the length is calculated using C<strlen> (so in this case
1185 the content cannot contain embedded ASCII NULs).
1187 Because of the message protocol, there is a transfer limit
1188 of somewhere between 2MB and 4MB. To transfer large files you should use
1195 This command writes zeroes over the first few blocks of C<device>.
1197 How many blocks are zeroed isn't specified (but it's I<not> enough
1198 to securely wipe the device). It should be sufficient to remove
1199 any partition tables, filesystem superblocks and so on.