1 TODO list for libguestfs
2 ======================================================================
4 This list contains random ideas and musings on features we could add
5 to libguestfs in future.
12 Ideas for the Python bindings:
13 https://www.redhat.com/archives/fedora-virt/2009-April/msg00114.html
18 The API needs more test coverage, particularly lesser-used system
21 The big unresolved issue is UID/GID mapping between guest filesystem
22 IDs and the host. It's not easy to automate this because you need
23 extra details about the guest itself in order to get to its
24 UID->username map (eg. /etc/passwd from the guest).
29 BufferIn should turn into <char *, int> and simple strings in other
30 languages that can handle 8 bit clean strings.
32 Limit on transfers would still be 2MB for these types.
33 - then implement write-file properly
35 febootstrap / debootstrap inside appliance
36 ------------------------------------------
38 This was originally proposed as a way to install new operating systems
39 in the appliance. However no one has come up with a workable
45 Complete the Haskell bindings (see discussion on haskell-cafe).
50 Complete the bind tests - must test the return values and error cases.
52 virt-inspector - make libvirt XML
53 ---------------------------------
55 It should be possible to generate libvirt XML from virt-inspector
56 data, at least partially. This would be just another output type so:
58 virt-inspector --libvirt guest.img
60 Note that recent versions of libvirt/virt-install allow guests to be
61 imported, so this is not so useful any more.
63 "Standalone/local mode"
64 -----------------------
66 Instead of running guestfsd (the daemon) inside qemu, there should be
67 an option to just run guestfsd directly.
69 The architecture in this mode would look like:
84 (1) This only makes sense if we are running as root.
86 (2) There is no console / kernel messages in this configuration, but
87 we might consider capturing stderr from the daemon.
89 (3) guestfs_config and guestfs_add_drive become no-ops.
91 Obviously in this configuration, commands are run directly on the
92 local machine's disks. You could just run the commands themselves
93 directly, but libguestfs provides a convenient API and language
94 bindings. Also deals with tricky stuff like parsing the output of the
95 LVM commands. Also we get to leverage other code such as
98 This is mainly useful from live CDs, ie. virt-p2v.
100 Should we bother having the daemon at all and just link the guestfsd
101 code directly into libguestfs?
106 [This section should be filed as bugs, but no one seems to care for
107 PPC hosts and the hardware is rapidly becoming obsolete]
109 ppc (32 bit) works with qemu from git, however there is no serial console
111 ppc64 requires extra parameters:
113 however it still fails:
114 invalid/unsupported opcode: 01 - 01 - 1a (06301e83) 00000000018c2738 1
115 invalid bits: 00400000 for opcode: 0b - 19 - 15 (2d746572) 0000000000009230
117 no serial console in ppc or ppc64 because no one can tell us what
118 console=ttyXX option to use
120 Supermin appliance to febootstrap
121 ---------------------------------
123 Supermin appliance functionality should be moved into febootstrap.
125 Ideas for extra commands
126 ------------------------
128 General glibc / core programs:
155 Other initrd-* commands
156 -----------------------
163 Simple editing of configuration files
164 -------------------------------------
166 Some easy non-Augeas methods to edit configuration files.
169 replace /etc/file key value
171 which would look in /etc/file for any instances of
177 and replace them with
183 That would solve about 50% of reconfiguration needs, and for the
184 rest you'd use Augeas, 'download'+'upload' or 'edit'.
186 RWMJ: I had a go at implementing this, but it's quite error-prone to
187 do this sort of editing inside the C-based daemon code. It's far
188 better to do it with Augeas, or else to use an external language like
194 Currently we can't do Perl "one-liners". ie. The current syntax for
195 any short Perl one-liner would be:
197 perl -MSys::Guestfs -e '$g = Sys::Guestfs->new(); $g->add_drive ("foo"); $g->launch; $g->mount ("/dev/sda1", "/"); ....'
199 You can see we're well beyond a single line just getting to the point
200 of adding drives and mounting.
204 $h = create ($filename, \"/dev/sda1\" => \"/\");
206 $h = create ([$file1, $file2], \"/dev/sda1\" => \"/\");
208 To mount read-only, add C<ro =E<gt> 1> like this:
210 $h = create ($filename, \"/dev/sda1\" => \"/\", ro => 1);
212 which is equivalent to the following sequence of calls:
214 $h = Sys::Guestfs->new ();
215 $h->set_autosync (1);
216 $h->add_drive_ro ($filename);
218 $h->mount_ro (\"/dev/sda1\", \"/\");
220 Command-line form would be:
222 perl -MSys::Guestfs=:all -e '$_=create("guest.img", "/dev/sda1" => "/"); $_->cat ("/etc/fstab");'
224 That's not brief enough for one-liners, so we could have an extra
225 autogenerated module which creates a Sys::Guestfs handle singleton
226 (the handle is an implicit global variable as in guestfish), eg:
228 perl -MSys::Guestfs::One -e 'inspect("guest.img"); cat ("/etc/fstab");'
230 How would editing files work?
236 http://man.linux-ntfs.org/ntfsclone.8.html
241 Equip guestfish with some standard images that it can load
246 Maybe it's better to create these on the fly?
252 http://search.cpan.org/~rgiersig/IO-Tty-1.08/Pty.pm
253 http://www.perlmonks.org/index.pl?node_id=582185
255 Note that pty requires cooperation inside the C code too (there are
256 two sides to a pty, and one has to be handled after the fork).
261 Pass TERM from the library, through the kernel command line, to the
264 Windows-based daemon/appliance
265 ------------------------------
267 See discussion on list:
268 https://www.redhat.com/archives/libguestfs/2009-November/msg00165.html
270 virt-grow, virt-shrink
271 ----------------------
273 Grow and shrink existing guests. The main problem comes with
274 MBR-style partitions where you have to actually copy data around the
275 disk (unless you only want to change the final partition).
280 Add -drive file=...,lock=exclusive and -drive file=...,lock=shared
282 Change libguestfs and libvirt to do the right thing, so that multiple
283 instances of qemu cannot stomp on each other.
288 For multi-level disk images such as live CDs:
289 http://rwmj.wordpress.com/2009/07/15/unpack-the-russian-doll-of-a-f11-live-cd/
291 It's possible with libguestfs to recursively look for anything that
292 might be a filesystem, mount-{,loop} it and look in those, revealing
293 anything in a disk image.
295 However this won't work easily for VM disk images in the disk image.
296 One would have to download those to the host and launch another