diff options
author | Richard W.M. Jones <rjones@redhat.com> | 2011-10-21 22:48:10 +0100 |
---|---|---|
committer | Richard W.M. Jones <rjones@redhat.com> | 2011-10-22 07:29:39 +0100 |
commit | e192035bbd6f80262f60565e0a91588848f2a4e3 (patch) | |
tree | 42730b78601c685efce2f0819daf34efa6c06d90 /tools | |
parent | c214cdc54e955ff1a01c00b99b0d6bd20e8b6bcd (diff) | |
download | libguestfs-e192035bbd6f80262f60565e0a91588848f2a4e3.tar.gz libguestfs-e192035bbd6f80262f60565e0a91588848f2a4e3.tar.xz libguestfs-e192035bbd6f80262f60565e0a91588848f2a4e3.zip |
virt-make-fs: Refresh man page.
Diffstat (limited to 'tools')
-rwxr-xr-x | tools/virt-make-fs | 74 |
1 files changed, 38 insertions, 36 deletions
diff --git a/tools/virt-make-fs b/tools/virt-make-fs index 28b4abad..d97d2629 100755 --- a/tools/virt-make-fs +++ b/tools/virt-make-fs @@ -1,6 +1,6 @@ #!/usr/bin/perl -w # virt-make-fs -# Copyright (C) 2010 Red Hat Inc. +# Copyright (C) 2010-2011 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 @@ -56,39 +56,13 @@ data to a VM). Basic usage is: - virt-make-fs input output + virt-make-fs input output.img where C<input> is either a directory containing files that you want to add, or a tar archive (either uncompressed tar or gzip-compressed -tar); and C<output> is a disk image. The input type is detected -automatically. The output disk image defaults to a raw ext2 image -unless you specify extra flags (see L</OPTIONS> below). - -=head2 EXTRA SPACE - -Unlike formats such as tar and squashfs, a filesystem does not "just -fit" the files that it contains, but might have extra space. -Depending on how you are going to use the output, you might think this -extra space is wasted and want to minimize it, or you might want to -leave space so that more files can be added later. Virt-make-fs -defaults to minimizing the extra space, but you can use the I<--size> -flag to leave space in the filesystem if you want it. - -An alternative way to leave extra space but not make the output image -any bigger is to use an alternative disk image format (instead of the -default "raw" format). Using I<--format=qcow2> will use the native -QEmu/KVM qcow2 image format (check your hypervisor supports this -before using it). This allows you to choose a large I<--size> but the -extra space won't actually be allocated in the image until you try to -store something in it. - -Don't forget that you can also use local commands including -L<resize2fs(8)> and L<virt-resize(1)> to resize existing filesystems, -or rerun virt-make-fs to build another image from scratch. - -=head3 EXAMPLE - - virt-make-fs --format=qcow2 --size=+200M input output.img +tar); and C<output.img> is a disk image. The input type is detected +automatically. The output disk image defaults to a raw ext2 sparse +image unless you specify extra flags (see L</OPTIONS> below). =head2 FILESYSTEM TYPE @@ -150,6 +124,32 @@ better to use an EFI/GPT-compatible partition table: virt-make-fs --partition=gpt --size=+4T --format=qcow2 input disk.img +=head2 EXTRA SPACE + +Unlike formats such as tar and squashfs, a filesystem does not "just +fit" the files that it contains, but might have extra space. +Depending on how you are going to use the output, you might think this +extra space is wasted and want to minimize it, or you might want to +leave space so that more files can be added later. Virt-make-fs +defaults to minimizing the extra space, but you can use the I<--size> +flag to leave space in the filesystem if you want it. + +An alternative way to leave extra space but not make the output image +any bigger is to use an alternative disk image format (instead of the +default "raw" format). Using I<--format=qcow2> will use the native +QEmu/KVM qcow2 image format (check your hypervisor supports this +before using it). This allows you to choose a large I<--size> but the +extra space won't actually be allocated in the image until you try to +store something in it. + +Don't forget that you can also use local commands including +L<resize2fs(8)> and L<virt-resize(1)> to resize existing filesystems, +or rerun virt-make-fs to build another image from scratch. + +=head3 EXAMPLE + + virt-make-fs --format=qcow2 --size=+200M input output.img + =head1 OPTIONS =over 4 @@ -222,7 +222,7 @@ my $format = "raw"; Choose the output disk image format. -The default is C<raw> (raw disk image). +The default is C<raw> (raw sparse disk image). For other choices, see the L<qemu-img(1)> manpage. The only other choice that would really make sense here is C<qcow2>. @@ -260,13 +260,15 @@ Note that if you just use a lonesome I<--partition>, the Perl option parser might consider the next parameter to be the partition type. For example: - virt-make-fs --partition input.tar ... + virt-make-fs --partition input.tar output.img would cause virt-make-fs to think you wanted to use a partition type of C<input.tar> which is completely wrong. To avoid this, use I<--> -(a double dash) between options and the input file argument: +(a double dash) between options and the input and output arguments: + + virt-make-fs --partition -- input.tar output.img - virt-make-fs --partition -- input.tar ... +For MBR, virt-make-fs sets the partition type byte automatically. =back @@ -583,7 +585,7 @@ Richard W.M. Jones L<http://people.redhat.com/~rjones/> =head1 COPYRIGHT -Copyright (C) 2010 Red Hat Inc. +Copyright (C) 2010-2011 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 |