Rather than me explaining over again what it does, I've included the man page below. Rich. ---------------------------------------------------------------------- virt-format(1) Virtualization Support virt-format(1) NAME virt-format - Erase and make a blank disk SYNOPSIS virt-format [--options] -a disk.img [-a disk.img ...] DESCRIPTION Virt-format takes an existing disk file (or it can be a host partition, LV etc), erases all data on it, and formats it as a blank disk. It can optionally create partition tables, empty filesystems, logical volumes and more. To create a disk containing data, you may be better to use virt-make-fs(1). If you are creating a blank disk to use in guestfish(1), you should instead use the guestfish -N option. Normal usage would be something like this: virt-format -a disk.qcow or this: virt-format -a /dev/VG/LV "disk.qcow" or "/dev/VG/LV" must exist already. Any data on these disks will be erased by these commands. These commands will create a single empty MBR partition covering the whole disk, with no filesystem inside it. Additional parameters can be used to control the creation of partitions, filesystems, etc. The most commonly used options are: --filesystem=[ext3|ntfs|vfat|...] Create an empty filesystem ("ext3", "ntfs" etc) inside the partition. --lvm[=/dev/VG/LV] Create a Linux LVM2 logical volume on the disk. When used with --filesystem, the filesystem is created inside the LV. For more information about these and other options, see "OPTIONS" below. The format of the disk is normally auto-detected, but you can also force it by using the --format option (q.v.). In situations where you do not trust the existing content of the disk, then it is advisable to use this option to avoid possible exploits. OPTIONS --help Display brief help. -a file --add file Add file, a disk image, host partition, LV, external USB disk, etc. The format of the disk image is auto-detected. To override this and force a particular format use the --format=.. option. Any existing data on the disk is erased. --filesystem=ext3|ntfs|vfat|... Create an empty filesystem of the specified type. Many filesystem types are supported by libguestfs. --filesystem=none Create no filesystem. This is the default. --format=raw|qcow2|.. --format The default for the -a option is to auto-detect the format of the disk image. Using this forces the disk format for -a options which follow on the command line. Using --format with no argument switches back to auto-detection for subsequent -a options. For example: virt-format --format=raw -a disk.img forces raw format (no auto-detection) for "disk.img". virt-format --format=raw -a disk.img --format -a another.img forces raw format (no auto-detection) for "disk.img" and reverts to auto-detection for "another.img". If you have untrusted raw-format guest disk images, you should use this option to specify the disk format. This avoids a possible security problem with malicious guests (CVE-2010-3851). --lvm=/dev/VG/LV Create a Linux LVM2 logical volume called "/dev/VG/LV". You can change the name of the volume group and logical volume. --lvm Create a Linux LVM2 logical volume with the default name ("/dev/VG/LV"). --lvm=none Create no logical volume. This is the default. --partition=gpt Create a GPT partition instead of MBR. This is useful for disks which are 2TB or larger in size. --partition --partition=mbr Create an MBR partition. This is the default. --partition=none Create no partition table. Note that Windows may not be able to see these disks. -v --verbose Enable verbose messages for debugging. -V --version Display version number and exit. --wipe Normally virt-format does not wipe data from the disk (because that takes a long time). Thus if there is data on the disk, it is only hidden and partially overwritten by virt-format, and it might be recovered by disk editing tools. If you use this option, virt-format writes zeroes over the whole disk so that previous data is not recoverable. -x Enable tracing of libguestfs API calls. EXIT STATUS This program returns 0 on success, or 1 on failure. SEE ALSO guestfs(3), guestfish(1), virt-filesystems(1), virt-make-fs(1), virt-rescue(1), virt-resize(1), <http://libguestfs.org/>. AUTHOR Richard W.M. Jones <http://people.redhat.com/~rjones/> COPYRIGHT Copyright (C) 2012 Red Hat Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. libguestfs-1.15.16 2012-01-17 virt-format(1)
From: "Richard W.M. Jones" <rjones@redhat.com> This tool allows you to easily reformat a disk, creating a blank disk with optional partition, LVM and empty filesystem. --- .gitignore | 4 + Makefile.am | 3 +- configure.ac | 1 + format/Makefile.am | 78 +++++++++ format/format.c | 438 +++++++++++++++++++++++++++++++++++++++++++++++ format/virt-format.pod | 203 ++++++++++++++++++++++ po/POTFILES.in | 1 + src/guestfs.pod | 5 + tests/extra/Makefile.am | 1 + tools/virt-make-fs | 3 + 10 files changed, 736 insertions(+), 1 deletions(-) create mode 100644 format/Makefile.am create mode 100644 format/format.c create mode 100755 format/virt-format.pod diff --git a/.gitignore b/.gitignore index 0611975..36411aa 100644 --- a/.gitignore +++ b/.gitignore @@ -105,6 +105,9 @@ fish/virt-copy-in.1 fish/virt-copy-out.1 fish/virt-tar-in.1 fish/virt-tar-out.1 +format/stamp-virt-format.pod +format/virt-format +format/virt-format.1 fuse/guestmount fuse/guestmount.1 fuse/stamp-guestmount.pod @@ -145,6 +148,7 @@ html/virt-copy-out.1.html html/virt-df.1.html html/virt-edit.1.html html/virt-filesystems.1.html +html/virt-format.1.html html/virt-inspector.1.html html/virt-list-filesystems.1.html html/virt-list-partitions.1.html diff --git a/Makefile.am b/Makefile.am index d14cc12..22ae04e 100644 --- a/Makefile.am +++ b/Makefile.am @@ -52,7 +52,7 @@ SUBDIRS += test-tool SUBDIRS += fish # virt-tools in C. -SUBDIRS += align cat df edit inspector rescue +SUBDIRS += align cat df edit format inspector rescue # Language bindings. if HAVE_PERL @@ -187,6 +187,7 @@ HTMLFILES = \ html/virt-df.1.html \ html/virt-edit.1.html \ html/virt-filesystems.1.html \ + html/virt-format.1.html \ html/virt-inspector.1.html \ html/virt-list-filesystems.1.html \ html/virt-list-partitions.1.html \ diff --git a/configure.ac b/configure.ac index d8abf71..fa97479 100644 --- a/configure.ac +++ b/configure.ac @@ -1018,6 +1018,7 @@ AC_CONFIG_FILES([Makefile erlang/examples/Makefile examples/Makefile fish/Makefile + format/Makefile fuse/Makefile generator/Makefile gnulib/lib/Makefile diff --git a/format/Makefile.am b/format/Makefile.am new file mode 100644 index 0000000..b3ff216 --- /dev/null +++ b/format/Makefile.am @@ -0,0 +1,78 @@ +# libguestfs virt format tool +# Copyright (C) 2012 Red Hat Inc. +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + +include $(top_srcdir)/subdir-rules.mk + +EXTRA_DIST = \ + virt-format.pod + +CLEANFILES = stamp-virt-format.pod + +bin_PROGRAMS = virt-format + +SHARED_SOURCE_FILES = \ + ../fish/config.c \ + ../fish/inspect.c \ + ../fish/keys.c \ + ../fish/options.h \ + ../fish/options.c \ + ../fish/virt.c + +virt_format_SOURCES = \ + $(SHARED_SOURCE_FILES) \ + format.c + +virt_format_CFLAGS = \ + -DGUESTFS_WARN_DEPRECATED=1 \ + -I$(top_srcdir)/src -I$(top_builddir)/src \ + -I$(top_srcdir)/fish \ + -I$(srcdir)/../gnulib/lib -I../gnulib/lib \ + -DLOCALEBASEDIR=\""$(datadir)/locale"\" \ + $(WARN_CFLAGS) $(WERROR_CFLAGS) \ + $(LIBCONFIG_CFLAGS) \ + $(LIBVIRT_CFLAGS) + +virt_format_LDADD = \ + $(LIBCONFIG_LIBS) \ + $(top_builddir)/src/libguestfs.la \ + ../gnulib/lib/libgnu.la \ + $(LIBVIRT_LIBS) + +# Manual pages and HTML files for the website. +man_MANS = virt-format.1 +noinst_DATA = $(top_builddir)/html/virt-format.1.html + +virt-format.1 $(top_builddir)/html/virt-format.1.html: stamp-virt-format.pod + +stamp-virt-format.pod: virt-format.pod + $(top_builddir)/podwrapper.sh \ + --man virt-format.1 \ + --html $(top_builddir)/html/virt-format.1.html \ + $< + touch $@ + +# Tests. + +# random_val := $(shell awk 'BEGIN{srand(); print 1+int(255*rand())}' < /dev/null) + +# TESTS_ENVIRONMENT = \ +# MALLOC_PERTURB_=$(random_val) \ +# LD_LIBRARY_PATH=$(top_builddir)/src/.libs \ +# LIBGUESTFS_PATH=$(top_builddir)/appliance \ +# TMPDIR=$(top_builddir) + +# TESTS = test-virt-format.sh diff --git a/format/format.c b/format/format.c new file mode 100644 index 0000000..c264026 --- /dev/null +++ b/format/format.c @@ -0,0 +1,438 @@ +/* virt-format + * Copyright (C) 2012 Red Hat Inc. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <config.h> + +#include <stdio.h> +#include <stdlib.h> +#include <stdint.h> +#include <string.h> +#include <inttypes.h> +#include <unistd.h> +#include <getopt.h> +#include <locale.h> +#include <assert.h> +#include <libintl.h> + +#include "progname.h" +#include "c-ctype.h" + +#include "guestfs.h" +#include "options.h" + +/* These globals are shared with options.c. */ +guestfs_h *g; + +int read_only = 0; +int live = 0; +int verbose = 0; +int keys_from_stdin = 0; +int echo_keys = 0; +int inspector = 0; +const char *libvirt_uri = NULL; + +static const char *filesystem = NULL; +static const char *vg = NULL, *lv = NULL; +static const char *partition = "mbr"; +static int wipe = 0; + +static void parse_vg_lv (const char *lvm); +static int do_format (void); +static int do_rescan (char **devices); + +static inline char * +bad_cast (char const *s) +{ + return (char *) s; +} + +static void __attribute__((noreturn)) +usage (int status) +{ + char *warning = + _("IMPORTANT NOTE: This program ERASES ALL DATA on disks."); + + if (status != EXIT_SUCCESS) + fprintf (stderr, _("Try `%s --help' for more information.\n%s\n"), + program_name, warning); + else { + fprintf (stdout, + _("%s: erase and make a blank disk\n" + "Copyright (C) 2012 Red Hat Inc.\n" + "\n" + "%s\n" + "\n" + "Usage:\n" + " %s [--options] -a disk.img [-a disk.img ...]\n" + "Options:\n" + " -a|--add image Add image\n" + " --filesystem=.. Create empty filesystem\n" + " --format[=raw|..] Force disk format for -a option\n" + " --help Display brief help\n" + " --lvm=.. Create Linux LVM2 logical volume\n" + " --partition=.. Create / set partition type\n" + " -v|--verbose Verbose messages\n" + " -V|--version Display version and exit\n" + " --wipe Write zeroes over whole disk\n" + " -x Trace libguestfs API calls\n" + "For more information, see the manpage %s(1).\n" + "\n" + "%s\n\n"), + program_name, warning, program_name, program_name, + warning); + } + exit (status); +} + +int +main (int argc, char *argv[]) +{ + /* Set global program name that is not polluted with libtool artifacts. */ + set_program_name (argv[0]); + + setlocale (LC_ALL, ""); + bindtextdomain (PACKAGE, LOCALEBASEDIR); + textdomain (PACKAGE); + + enum { HELP_OPTION = CHAR_MAX + 1 }; + + static const char *options = "a:c:d:qvVx"; + static const struct option long_options[] = { + { "add", 1, 0, 'a' }, + { "filesystem", 1, 0, 0 }, + { "format", 2, 0, 0 }, + { "help", 0, 0, HELP_OPTION }, + { "lvm", 2, 0, 0 }, + { "partition", 2, 0, 0 }, + { "verbose", 0, 0, 'v' }, + { "version", 0, 0, 'V' }, + { "wipe", 0, 0, 0 }, + { 0, 0, 0, 0 } + }; + struct drv *drvs = NULL; + struct drv *drv; + const char *format = NULL; + int c; + int option_index; + int retry, retries; + + g = guestfs_create (); + if (g == NULL) { + fprintf (stderr, _("guestfs_create: failed to create handle\n")); + exit (EXIT_FAILURE); + } + + argv[0] = bad_cast (program_name); + + for (;;) { + c = getopt_long (argc, argv, options, long_options, &option_index); + if (c == -1) break; + + switch (c) { + case 0: /* options which are long only */ + if (STREQ (long_options[option_index].name, "format")) { + if (!optarg || STREQ (optarg, "")) + format = NULL; + else + format = optarg; + } else if (STREQ (long_options[option_index].name, "filesystem")) { + if (STREQ (optarg, "none")) + filesystem = NULL; + else if (optarg[0] == '-') { /* eg: --filesystem --lvm */ + fprintf (stderr, _("%s: no filesystem was specified\n"), + program_name); + exit (EXIT_FAILURE); + } else + filesystem = optarg; + } else if (STREQ (long_options[option_index].name, "lvm")) { + if (vg || lv) { + fprintf (stderr, + _("%s: --lvm option cannot be given multiple times\n"), + program_name); + exit (EXIT_FAILURE); + } + if (optarg == NULL) { + vg = strdup ("VG"); + lv = strdup ("LV"); + if (!vg || !lv) { perror ("strdup"); exit (EXIT_FAILURE); } + } + else if (STREQ (optarg, "none")) + vg = lv = NULL; + else + parse_vg_lv (optarg); + } else if (STREQ (long_options[option_index].name, "partition")) { + if (optarg == NULL) + partition = "mbr"; + else if (STREQ (optarg, "none")) + partition = NULL; + else + partition = optarg; + } else if (STREQ (long_options[option_index].name, "wipe")) { + wipe = 1; + } else { + fprintf (stderr, _("%s: unknown long option: %s (%d)\n"), + program_name, long_options[option_index].name, option_index); + exit (EXIT_FAILURE); + } + break; + + case 'a': + OPTION_a; + break; + + case 'v': + OPTION_v; + break; + + case 'V': + OPTION_V; + break; + + case 'x': + OPTION_x; + break; + + case HELP_OPTION: + usage (EXIT_SUCCESS); + + default: + usage (EXIT_FAILURE); + } + } + + /* These are really constants, but they have to be variables for the + * options parsing code. Assert here that they have known-good + * values. + */ + assert (read_only == 0); + assert (inspector == 0); + assert (live == 0); + + /* Must be no extra arguments on the command line. */ + if (optind != argc) + usage (EXIT_FAILURE); + + /* The user didn't specify any drives to format. */ + if (drvs == NULL) + usage (EXIT_FAILURE); + + /* Because the libguestfs kernel can get stuck rereading the + * partition table after things have been erased, we sometimes need + * to completely restart the guest. Hence this complex retry logic. + */ + for (retries = 0; retries <= 1; ++retries) { + /* Add domains/drives from the command line (for a single guest). */ + add_drives (drvs, 'a'); + + if (guestfs_launch (g) == -1) + exit (EXIT_FAILURE); + + /* Perform the format. */ + retry = do_format (); + if (!retry) + break; + + if (retries == 0) { + /* We're going to silently retry, after reopening the connection. */ + guestfs_h *g2; + + g2 = guestfs_create (); + guestfs_set_verbose (g2, guestfs_get_verbose (g)); + guestfs_set_trace (g2, guestfs_get_trace (g)); + + guestfs_close (g); + g = g2; + } + else { + /* Failed. */ + fprintf (stderr, + _("%s: failed to rescan the disks after two attempts. This\n" + "may mean there is some sort of partition table or disk\n" + "data which we are unable to remove. If you think this\n" + "is a bug, please file a bug at http://libguestfs.org/\n"), + program_name); + exit (EXIT_FAILURE); + } + } + + /* Free up data structures. */ + free_drives (drvs); + + guestfs_close (g); + + exit (EXIT_SUCCESS); +} + +/* Parse lvm string of the form "/dev/VG/LV" or "VG/LV". + * This sets the global variables 'vg' and 'lv', or exits on failure. + */ +static void +parse_vg_lv (const char *lvm) +{ + size_t i; + + if (STRPREFIX (lvm, "/dev/")) + lvm += 5; + + i = strcspn (lvm, "/"); + if (lvm[i]) { + vg = strndup (lvm, i); + lv = strdup (lvm + i + 1); + + if (!vg || !lv) { + perror ("strdup"); + exit (EXIT_FAILURE); + } + } else { + cannot_parse: + fprintf (stderr, _("%s: cannot parse --lvm option (%s)\n"), + program_name, lvm); + exit (EXIT_FAILURE); + } + + if (strchr (vg, '/') || strchr (lv, '/')) + goto cannot_parse; +} + +/* Returns 0 on success, 1 if we need to retry. */ +static int +do_format (void) +{ + char **devices; + size_t i, pass; + + devices = guestfs_list_devices (g); + if (devices == NULL) + exit (EXIT_FAILURE); + + /* Erase the disks. */ + if (!wipe) { + char **parts; + + /* No wipe, but get rid of LVM metadata by erasing each partition. */ + parts = guestfs_list_partitions (g); + if (parts == NULL) + exit (EXIT_FAILURE); + + for (i = 0; parts[i] != NULL; ++i) { + if (guestfs_zero (g, parts[i]) == -1) + exit (EXIT_FAILURE); + } + + /* Then erase the partition table on each device. */ + for (i = 0; devices[i] != NULL; ++i) { + if (guestfs_zero (g, devices[i]) == -1) + exit (EXIT_FAILURE); + } + } + else /* wipe */ { + for (i = 0; devices[i] != NULL; ++i) { + if (guestfs_zero_device (g, devices[i]) == -1) + exit (EXIT_FAILURE); + } + } + + if (do_rescan (devices)) + return 1; /* which means, reopen the handle and retry */ + + /* Format each disk. */ + for (i = 0; devices[i] != NULL; ++i) { + char *dev = devices[i]; + int free_dev = 0; + + if (partition) { + if (guestfs_part_disk (g, devices[i], partition) == -1) + exit (EXIT_FAILURE); + if (asprintf (&dev, "%s1", devices[i]) == -1) { + perror ("asprintf"); + exit (EXIT_FAILURE); + } + free_dev = 1; + } + + if (vg && lv) { + char *devs[2] = { dev, NULL }; + + if (guestfs_pvcreate (g, dev) == -1) + exit (EXIT_FAILURE); + + if (guestfs_vgcreate (g, vg, devs) == -1) + exit (EXIT_FAILURE); + + if (guestfs_lvcreate (g, lv, vg, 32) == -1) /* 32 MB is smallest LV */ + exit (EXIT_FAILURE); + + if (free_dev) + free (dev); + dev = NULL; + if (asprintf (&dev, "/dev/%s/%s", vg, lv) == -1) { + perror ("asprintf"); + exit (EXIT_FAILURE); + } + free_dev = 1; + + if (guestfs_lvresize_free (g, dev, 100) == -1) + exit (EXIT_FAILURE); + } + + if (filesystem) { + if (guestfs_mkfs_opts (g, filesystem, dev, -1) == -1) + exit (EXIT_FAILURE); + } + + if (free_dev) + free (dev); + } + + if (guestfs_sync (g) == -1) + exit (EXIT_FAILURE); + + /* Free device list. */ + for (i = 0; devices[i] != NULL; ++i) + free (devices[i]); + free (devices); + + return 0; +} + +/* Rescan everything so the kernel knows that there are no partition + * tables, VGs etc. Returns 0 on success, 1 if we need to retry. + */ +static int +do_rescan (char **devices) +{ + size_t i; + int errors = 0; + guestfs_error_handler_cb old_error_cb; + void *old_error_data; + + old_error_cb = guestfs_get_error_handler (g, &old_error_data); + guestfs_set_error_handler (g, NULL, NULL); + + for (i = 0; devices[i] != NULL; ++i) { + if (guestfs_blockdev_rereadpt (g, devices[i]) == -1) + errors++; + } + + if (guestfs_vgscan (g) == -1) + errors++; + + guestfs_set_error_handler (g, old_error_cb, old_error_data); + + return errors ? 1 : 0; +} diff --git a/format/virt-format.pod b/format/virt-format.pod new file mode 100755 index 0000000..d00dfa6 --- /dev/null +++ b/format/virt-format.pod @@ -0,0 +1,203 @@ +=encoding utf8 + +=head1 NAME + +virt-format - Erase and make a blank disk + +=head1 SYNOPSIS + + virt-format [--options] -a disk.img [-a disk.img ...] + +=head1 DESCRIPTION + +Virt-format takes an existing disk file (or it can be a host +partition, LV etc), B<erases all data on it>, and formats it as a +blank disk. It can optionally create partition tables, empty +filesystems, logical volumes and more. + +To create a disk containing data, you may be better to use +L<virt-make-fs(1)>. If you are creating a blank disk to use in +L<guestfish(1)>, you should instead use the guestfish I<-N> option. + +Normal usage would be something like this: + + virt-format -a disk.qcow + +or this: + + virt-format -a /dev/VG/LV + +C<disk.qcow> or C</dev/VG/LV> must exist already. B<Any data on these +disks will be erased by these commands>. These commands will create a +single empty MBR partition covering the whole disk, with no filesystem +inside it. + +Additional parameters can be used to control the creation of +partitions, filesystems, etc. The most commonly used options are: + +=over 4 + +=item I<--filesystem=[ext3|ntfs|vfat|...]> + +Create an empty filesystem (C<ext3>, C<ntfs> etc) inside the partition. + +=item I<--lvm[=/dev/VG/LV]> + +Create a Linux LVM2 logical volume on the disk. When used with +I<--filesystem>, the filesystem is created inside the LV. + +=back + +For more information about these and other options, see +L</OPTIONS> below. + +The format of the disk is normally auto-detected, but you can also +force it by using the I<--format> option (q.v.). In situations where +you do not trust the existing content of the disk, then it is +advisable to use this option to avoid possible exploits. + +=head1 OPTIONS + +=over 4 + +=item B<--help> + +Display brief help. + +=item B<-a> file + +=item B<--add> file + +Add I<file>, a disk image, host partition, LV, external USB disk, etc. + +The format of the disk image is auto-detected. To override this and +force a particular format use the I<--format=..> option. + +B<Any existing data on the disk is erased.> + +=item B<--filesystem=ext3|ntfs|vfat|...> + +Create an empty filesystem of the specified type. Many filesystem +types are supported by libguestfs. + +=item B<--filesystem=none> + +Create no filesystem. This is the default. + +=item B<--format=raw|qcow2|..> + +=item B<--format> + +The default for the I<-a> option is to auto-detect the format of the +disk image. Using this forces the disk format for I<-a> options which +follow on the command line. Using I<--format> with no argument +switches back to auto-detection for subsequent I<-a> options. + +For example: + + virt-format --format=raw -a disk.img + +forces raw format (no auto-detection) for C<disk.img>. + + virt-format --format=raw -a disk.img --format -a another.img + +forces raw format (no auto-detection) for C<disk.img> and reverts to +auto-detection for C<another.img>. + +If you have untrusted raw-format guest disk images, you should use +this option to specify the disk format. This avoids a possible +security problem with malicious guests (CVE-2010-3851). + +=item B<--lvm=/dev/I<VG>/I<LV>> + +Create a Linux LVM2 logical volume called C</dev/I<VG>/I<LV>>. You +can change the name of the volume group and logical volume. + +=item B<--lvm> + +Create a Linux LVM2 logical volume with the default name +(C</dev/VG/LV>). + +=item B<--lvm=none> + +Create no logical volume. This is the default. + +=item B<--partition=gpt> + +Create a GPT partition instead of MBR. This is useful for disks which +are 2TB or larger in size. + +=item B<--partition> + +=item B<--partition=mbr> + +Create an MBR partition. This is the default. + +=item B<--partition=none> + +Create no partition table. Note that Windows may not be able to see +these disks. + +=item B<-v> + +=item B<--verbose> + +Enable verbose messages for debugging. + +=item B<-V> + +=item B<--version> + +Display version number and exit. + +=item B<--wipe> + +Normally virt-format does not wipe data from the disk (because that +takes a long time). Thus if there is data on the disk, it is only +hidden and partially overwritten by virt-format, and it might be +recovered by disk editing tools. + +If you use this option, virt-format writes zeroes over the whole disk +so that previous data is not recoverable. + +=item B<-x> + +Enable tracing of libguestfs API calls. + +=back + +=head1 EXIT STATUS + +This program returns C<0> on success, or C<1> on failure. + +=head1 SEE ALSO + +L<guestfs(3)>, +L<guestfish(1)>, +L<virt-filesystems(1)>, +L<virt-make-fs(1)>, +L<virt-rescue(1)>, +L<virt-resize(1)>, +L<http://libguestfs.org/>. + +=head1 AUTHOR + +Richard W.M. Jones L<http://people.redhat.com/~rjones/> + +=head1 COPYRIGHT + +Copyright (C) 2012 Red Hat Inc. + +This program is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2 of the License, or +(at your option) any later version. + +This program is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with this program; if not, write to the Free Software +Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. diff --git a/po/POTFILES.in b/po/POTFILES.in index 380a3c7..84972db 100644 --- a/po/POTFILES.in +++ b/po/POTFILES.in @@ -124,6 +124,7 @@ fish/supported.c fish/tilde.c fish/time.c fish/virt.c +format/format.c fuse/dircache.c fuse/guestmount.c inspector/virt-inspector.c diff --git a/src/guestfs.pod b/src/guestfs.pod index f7740b6..55f6e7f 100644 --- a/src/guestfs.pod +++ b/src/guestfs.pod @@ -2955,6 +2955,10 @@ L<guestfish(1)>, the command-line shell, and various shell scripts built on top such as L<virt-copy-in(1)>, L<virt-copy-out(1)>, L<virt-tar-in(1)>, L<virt-tar-out(1)>. +=item C<format> + +L<virt-format(1)> command and documentation. + =item C<fuse> L<guestmount(1)>, FUSE (userspace filesystem) built on top of libguestfs. @@ -3271,6 +3275,7 @@ L<virt-copy-out(1)>, L<virt-df(1)>, L<virt-edit(1)>, L<virt-filesystems(1)>, +L<virt-format(1)>, L<virt-inspector(1)>, L<virt-list-filesystems(1)>, L<virt-list-partitions(1)>, diff --git a/tests/extra/Makefile.am b/tests/extra/Makefile.am index 8f34958..83abc65 100644 --- a/tests/extra/Makefile.am +++ b/tests/extra/Makefile.am @@ -102,6 +102,7 @@ test-prerequisites: test-tools-null: $(RUN_VG) ../../fish/guestfish -N part exit $(RUN_VG) ../../align/virt-alignment-scan -a test1.img >/dev/null + $(RUN_VG) ../../format/virt-format -a test1.img >/dev/null rm test1.img $(RUN_VG) ../../cat/virt-filesystems -a /dev/null >/dev/null $(RUN_VG) ../../cat/virt-filesystems -a /dev/null --all --long -h --uuid >/dev/null diff --git a/tools/virt-make-fs b/tools/virt-make-fs index d4e231b..3afbb43 100755 --- a/tools/virt-make-fs +++ b/tools/virt-make-fs @@ -54,6 +54,8 @@ which can be useful if you want to attach these filesystems to existing virtual machines (eg. to import large amounts of read-only data to a VM). +To create blank disks, use L<virt-format(1)>. + Basic usage is: virt-make-fs input output.img @@ -557,6 +559,7 @@ manual page L<sh(1)> for details. =head1 SEE ALSO L<guestfish(1)>, +L<virt-format(1)>, L<virt-resize(1)>, L<virt-tar-in(1)>, L<mkisofs(1)>, -- 1.7.6
On 01/17/2012 07:48 PM, Richard W.M. Jones wrote:
Rather than me explaining over again what it does, I've included the man page below.
Rich.
----------------------------------------------------------------------
virt-format(1) Virtualization Support virt-format(1)
NAME virt-format - Erase and make a blank disk
SYNOPSIS virt-format [--options] -a disk.img [-a disk.img ...]
DESCRIPTION Virt-format takes an existing disk file (or it can be a host partition, LV etc), erases all data on it, and formats it as a blank disk. It can optionally create partition tables, empty filesystems, logical volumes and more.
To create a disk containing data, you may be better to use virt-make-fs(1). If you are creating a blank disk to use in guestfish(1), you should instead use the guestfish -N option.
Normal usage would be something like this:
virt-format -a disk.qcow
or this:
virt-format -a /dev/VG/LV
"disk.qcow" or "/dev/VG/LV" must exist already. Any data on these disks will be erased by these commands. These commands will create a single empty MBR partition covering the whole disk, with no filesystem inside it.
Some thoughts: - does it autodetect when disk is too large for MBR and GPT would be needed? - does it align partitions to SSD erase block size? (LVM is particularly tricky to get right manually) - can I specify a backing-file for qemu disks? (qcow, and qed I think) - can --wipe use TRIM on SSDs instead of writing zeroes? - can I setup LVM striping and mirror with it? If the answer is yes to all these then this tool could even be used as a general purpose format tool, not just for virt environments :) Best regards, --Edwin
On Tue, Jan 17, 2012 at 07:58:15PM +0200, Török Edwin wrote:
- does it autodetect when disk is too large for MBR and GPT would be needed?
No, but you're right that it should do. I'll make this change.
- does it align partitions to SSD erase block size? (LVM is particularly tricky to get right manually)
It aligns to 64K at the moment, for partitions. For LVM we just use the defaults, whatever they are. I don't know if it is possible to pass the SSD erase block size up through the virt stack. CC'd to Paolo Bonzini: Will virtio-scsi make this easier for us in future?
- can I specify a backing-file for qemu disks? (qcow, and qed I think)
Yes, in the sense that it just uses the qcow2 disk and doesn't attempt to create one, so whatever backing-file etc settings are already on the qcow2 disk are not modified in any way.
- can --wipe use TRIM on SSDs instead of writing zeroes?
No. This is a missing qemu feature which I've asked about a few times, but until qemu and the rest of the virt stack implements it, we won't have it in libguestfs (and conversely when it is implemented, we'll most likely get it for free). Paolo Bonzini: ^^ same question about virtio-scsi ^^
- can I setup LVM striping and mirror with it?
No. In general these are not exposed through libguestfs, although they should be, and could be easily enough if someone wrote the patch. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones New in Fedora 11: Fedora Windows cross-compiler. Compile Windows programs, test, and build Windows installers. Over 70 libraries supprt'd http://fedoraproject.org/wiki/MinGW http://www.annexia.org/fedora_mingw
On 01/17/2012 08:11 PM, Richard W.M. Jones wrote:
On Tue, Jan 17, 2012 at 07:58:15PM +0200, Török Edwin wrote:
- does it autodetect when disk is too large for MBR and GPT would be needed?
No, but you're right that it should do. I'll make this change.
- does it align partitions to SSD erase block size? (LVM is particularly tricky to get right manually)
It aligns to 64K at the moment, for partitions. For LVM we just use the defaults, whatever they are.
I think you need to account for the LVM header size, there was a blog-post from Theodore Tso on this topic: http://web.archive.org/web/20101130200444/http://thunk.org/tytso/blog/2009/0...
I don't know if it is possible to pass the SSD erase block size up through the virt stack.
I was thinking more about a command-line flag to specify alignment. My SSD reports 512 bytes as both logical and physical sector size, and thats a physical drive. Only my HDD reports a 4096 physical sector size (512 logical).
CC'd to Paolo Bonzini: Will virtio-scsi make this easier for us in future?
- can I specify a backing-file for qemu disks? (qcow, and qed I think)
Yes, in the sense that it just uses the qcow2 disk and doesn't attempt to create one, so whatever backing-file etc settings are already on the qcow2 disk are not modified in any way.
- can --wipe use TRIM on SSDs instead of writing zeroes?
No. This is a missing qemu feature which I've asked about a few times, but until qemu and the rest of the virt stack implements it, we won't have it in libguestfs (and conversely when it is implemented, we'll most likely get it for free).
Paolo Bonzini: ^^ same question about virtio-scsi ^^
- can I setup LVM striping and mirror with it?
No. In general these are not exposed through libguestfs, although they should be, and could be easily enough if someone wrote the patch.
Rich.
On 01/17/2012 07:19 PM, Török Edwin wrote:
I don't know if it is possible to pass the SSD erase block size up through the virt stack.
I was thinking more about a command-line flag to specify alignment. My SSD reports 512 bytes as both logical and physical sector size, and thats a physical drive. Only my HDD reports a 4096 physical sector size (512 logical).
What do you have for the SSD in /sys/block/sda/discard_alignment? Paolo
On 01/18/2012 03:12 PM, Paolo Bonzini wrote:
On 01/17/2012 07:19 PM, Török Edwin wrote:
I don't know if it is possible to pass the SSD erase block size up through the virt stack.
I was thinking more about a command-line flag to specify alignment. My SSD reports 512 bytes as both logical and physical sector size, and thats a physical drive. Only my HDD reports a 4096 physical sector size (512 logical).
What do you have for the SSD in /sys/block/sda/discard_alignment?
# cat /sys/block/sdg/discard_alignment 0 # hdparm -I /dev/sdg /dev/sdg: ATA device, with non-removable media Model Number: OCZ-VERTEX2 Serial Number: OCZ-RWK816SY5C086YHS Firmware Revision: 1.35 Transport: Serial Standards: Used: unknown (minor revision code 0x0028) Supported: 8 7 6 5 Likely used: 8 Configuration: Logical max current cylinders 16383 16383 heads 16 16 sectors/track 63 63 -- CHS current addressable sectors: 16514064 LBA user addressable sectors: 175836528 LBA48 user addressable sectors: 175836528 Logical Sector size: 512 bytes Physical Sector size: 512 bytes Logical Sector-0 offset: 0 bytes device size with M = 1024*1024: 85857 MBytes device size with M = 1000*1000: 90028 MBytes (90 GB) cache/buffer size = unknown Nominal Media Rotation Rate: Solid State Device Best regards, --Edwin
On 01/17/2012 07:11 PM, Richard W.M. Jones wrote:
- does it align partitions to SSD erase block size? (LVM is particularly tricky to get right manually)
It aligns to 64K at the moment, for partitions. For LVM we just use the defaults, whatever they are. I don't know if it is possible to pass the SSD erase block size up through the virt stack.
CC'd to Paolo Bonzini: Will virtio-scsi make this easier for us in future?
SSD erase data will be passed up transparently if you pass the entire disk to the VM. Otherwise, it will make this a bit easier (there will be no changes required to the guest, only the host) but some code has to be written to get the info from sysfs.
- can --wipe use TRIM on SSDs instead of writing zeroes?
Paolo Bonzini: ^^ same question about virtio-scsi ^^
Same here: no changes required to the guest, only the host. You will also be able to punch holes in image files when TRIM commands are issued to the SCSI disk, and the image files will become sparse as long as the underlying file systems supports that. Paolo
participants (3)
-
Paolo Bonzini -
Richard W.M. Jones -
Török Edwin