[Qemu-devel] Mounting qemu virtual disk images from host linux os

[Qemu-devel] Mounting qemu virtual disk images from host linux os

[Tom Sandholm]

[-- Attachment #1: Type: text/plain, Size: 683 bytes --]

I found the lomount program that is supposed to mount your qemu 
virtual-disk-images; but it isn't working for me.

I've installed debian "sarge" onto the virtual image, using ext3 
filesystems as default, with /dev/hda1 as the primary bootable linux 
partition (root).

When I execute the lomount program, here's what I get.
============================================================
nixsys:/qemu# ./lomount -t ext3 -diskimage debian02_sarge_hda.img /mnt
mount: wrong fs type, bad option, bad superblock on /dev/loop0,
       or too many mounted file systems
nixsys:/qemu#
===========================================================

What am I doing wrong?
Thanks!

Tom Sandholm


[-- Attachment #2: sandholm.vcf --]
[-- Type: text/x-vcard, Size: 101 bytes --]

begin:vcard
fn:Tom Sandholm
n:Sandholm;Tom
email;internet:sandholm@gsinet.net
version:2.1
end:vcard

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Magnus Damm]

On Sun, 16 Jan 2005 18:33:54 -0500, Tom Sandholm <sandholm@gsinet.net> wrote:
> I found the lomount program that is supposed to mount your qemu
> virtual-disk-images; but it isn't working for me.
> 
> I've installed debian "sarge" onto the virtual image, using ext3
> filesystems as default, with /dev/hda1 as the primary bootable linux
> partition (root).
> 
> When I execute the lomount program, here's what I get.
> ============================================================
> nixsys:/qemu# ./lomount -t ext3 -diskimage debian02_sarge_hda.img /mnt
> mount: wrong fs type, bad option, bad superblock on /dev/loop0,
>        or too many mounted file systems
> nixsys:/qemu#
> ===========================================================
> 
> What am I doing wrong?

I don't know about lomount, but are you sure your .img-file is a partition?
Maybe the file represents a virtual disk with partition table and partitions?

/ magnus

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Tom Sandholm]

[-- Attachment #1.1: Type: text/plain, Size: 1245 bytes --]

Oh yes, it is multi-partitioned.  Any suggestions on how I can select a 
specific partition?

Thanks
Tom


Magnus Damm wrote:

>On Sun, 16 Jan 2005 18:33:54 -0500, Tom Sandholm <sandholm@gsinet.net> wrote:
>  
>
>>I found the lomount program that is supposed to mount your qemu
>>virtual-disk-images; but it isn't working for me.
>>
>>I've installed debian "sarge" onto the virtual image, using ext3
>>filesystems as default, with /dev/hda1 as the primary bootable linux
>>partition (root).
>>
>>When I execute the lomount program, here's what I get.
>>============================================================
>>nixsys:/qemu# ./lomount -t ext3 -diskimage debian02_sarge_hda.img /mnt
>>mount: wrong fs type, bad option, bad superblock on /dev/loop0,
>>       or too many mounted file systems
>>nixsys:/qemu#
>>===========================================================
>>
>>What am I doing wrong?
>>    
>>
>
>I don't know about lomount, but are you sure your .img-file is a partition?
>Maybe the file represents a virtual disk with partition table and partitions?
>
>/ magnus
>
>
>_______________________________________________
>Qemu-devel mailing list
>Qemu-devel@nongnu.org
>http://lists.nongnu.org/mailman/listinfo/qemu-devel
>
>
>  
>

[-- Attachment #1.2: Type: text/html, Size: 1905 bytes --]

[-- Attachment #2: sandholm.vcf --]
[-- Type: text/x-vcard, Size: 101 bytes --]

begin:vcard
fn:Tom Sandholm
n:Sandholm;Tom
email;internet:sandholm@gsinet.net
version:2.1
end:vcard

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Mark Williamson]

I've not used lomount before so maybe you can pass flags to lomount directly.  
I can tell you how to do it using losetup / mount separately:

You can use losetup to bind a partition of the disk file to a /dev/loop device 
and then mount that.  Use the -o option to specify the offset into the file 
where the partition is to be found.  Use the -s option to specify the size of 
the partition.

You can probably fdisk the file (or, if that doesn't work, a loop device bound 
to the whole file) to figure out where the partition boundaries are.

HTH,
Mark

On Monday 17 January 2005 01:51, Tom Sandholm wrote:
> Oh yes, it is multi-partitioned.  Any suggestions on how I can select a
> specific partition?
>
> Thanks
> Tom
>
> Magnus Damm wrote:
> >On Sun, 16 Jan 2005 18:33:54 -0500, Tom Sandholm <sandholm@gsinet.net> 
wrote:
> >>I found the lomount program that is supposed to mount your qemu
> >>virtual-disk-images; but it isn't working for me.
> >>
> >>I've installed debian "sarge" onto the virtual image, using ext3
> >>filesystems as default, with /dev/hda1 as the primary bootable linux
> >>partition (root).
> >>
> >>When I execute the lomount program, here's what I get.
> >>============================================================
> >>nixsys:/qemu# ./lomount -t ext3 -diskimage debian02_sarge_hda.img /mnt
> >>mount: wrong fs type, bad option, bad superblock on /dev/loop0,
> >>       or too many mounted file systems
> >>nixsys:/qemu#
> >>===========================================================
> >>
> >>What am I doing wrong?
> >
> >I don't know about lomount, but are you sure your .img-file is a
> > partition? Maybe the file represents a virtual disk with partition table
> > and partitions?
> >
> >/ magnus
> >
> >
> >_______________________________________________
> >Qemu-devel mailing list
> >Qemu-devel@nongnu.org
> >http://lists.nongnu.org/mailman/listinfo/qemu-devel

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Bartosz Fabianowski]

> Oh yes, it is multi-partitioned.  Any suggestions on how I can select a 
> specific partition?

According to the source, lomount has a "-partition" option. So you might 
want to try:

./lomount -t ext3 -diskimage debian02_sarge_hda.img -partition 1 /mnt

- Bartosz

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Jim C. Brown]

On Mon, Jan 17, 2005 at 01:42:19AM +0100, Magnus Damm wrote:
> I don't know about lomount, but are you sure your .img-file is a partition?
> Maybe the file represents a virtual disk with partition table and partitions?
> 
> / magnus
> 

Shouldn't matter. lomount defaults to partition 1. You can specify a different
one  via the -partition option.

-- 
Infinite complexity begets infinite beauty.
Infinite precision begets infinite perfection.

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Jim C. Brown]

On Mon, Jan 17, 2005 at 05:01:07AM +0100, Bartosz Fabianowski wrote:
> >Oh yes, it is multi-partitioned.  Any suggestions on how I can select a 
> >specific partition?
> 
> According to the source, lomount has a "-partition" option. So you might 
> want to try:
> 
> ./lomount -t ext3 -diskimage debian02_sarge_hda.img -partition 1 /mnt
> 
> - Bartosz
> 

Well, lomount defaults to partition 1 if not given that option. He might want
a different partition tho. (Brute force method: try adding "-partition 2",
"-partition 3", and so on until it mounts the one you want it to.)

-- 
Infinite complexity begets infinite beauty.
Infinite precision begets infinite perfection.

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Jim C. Brown]

On Sun, Jan 16, 2005 at 06:33:54PM -0500, Tom Sandholm wrote:
> I found the lomount program that is supposed to mount your qemu 
> virtual-disk-images; but it isn't working for me.
> 
> I've installed debian "sarge" onto the virtual image, using ext3 
> filesystems as default, with /dev/hda1 as the primary bootable linux 
> partition (root).
> 
> When I execute the lomount program, here's what I get.
> ============================================================
> nixsys:/qemu# ./lomount -t ext3 -diskimage debian02_sarge_hda.img /mnt
> mount: wrong fs type, bad option, bad superblock on /dev/loop0,
>       or too many mounted file systems
> nixsys:/qemu#
> ===========================================================
> 
> What am I doing wrong?
> Thanks!
> 
> Tom Sandholm
> 

What partition do you want to mount? Are you sure it is the very first parition?
What is the output of "sdisk -l debian02_sarge_hda.img" ?

> _______________________________________________
> Qemu-devel mailing list
> Qemu-devel@nongnu.org
> http://lists.nongnu.org/mailman/listinfo/qemu-devel


-- 
Infinite complexity begets infinite beauty.
Infinite precision begets infinite perfection.

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Jim C. Brown]

On Mon, Jan 17, 2005 at 03:31:22AM +0000, Mark Williamson wrote:
> I've not used lomount before so maybe you can pass flags to lomount directly.
> I can tell you how to do it using losetup / mount separately:
> 
> You can use losetup to bind a partition of the disk file to a /dev/loop device
> and then mount that.  Use the -o option to specify the offset into the file
> where the partition is to be found.  Use the -s option to specify the size of
> the partition.
> 
> You can probably fdisk the file (or, if that doesn't work, a loop device bound
> to the whole file) to figure out where the partition boundaries are.
> 
> HTH,
> Mark

lomount does all of this internally. You can't pass it an offset into the file,
only which partition you wish to mount. It calculates the offset by looking it
up in the partition table.

I'm wondering if he is trying to mount a COW image or some fancy-but-qemu-supported
disk image. lomount has only been tested with raw images and sparse images, and
is unlikely to work w/ anything else.

-- 
Infinite complexity begets infinite beauty.
Infinite precision begets infinite perfection.

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Tom Sandholm]

[-- Attachment #1.1: Type: text/plain, Size: 2255 bytes --]

Hi,
  Thanks for all the replies. 
The image was created with qemu-img, with a size of 4GB.
I am able to fdisk image-name and see the partitions. (as follows)
Just can't mount partition 1.
====================================
nixsys:/qemu# fdisk debian01_sarge_hda.img

The number of cylinders for this disk is set to 8322.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): p

Disk debian01_sarge_hda.img: 4294 MB, 4294967296 bytes
16 heads, 63 sectors/track, 8322 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes

                 Device Boot      Start         End      Blocks   Id  System
debian01_sarge_hda.img1   *           1        7904     3983584+  83  Linux
debian01_sarge_hda.img2            7905        8322      210672    f  
W95 Ext'd (LBA)
debian01_sarge_hda.img5            7905        8322      210640+  82  
Linux swap

Command (m for help):
========================================

Thanks!
Tom Sandholm.




Jim C. Brown wrote:

>On Mon, Jan 17, 2005 at 03:31:22AM +0000, Mark Williamson wrote:
>  
>
>>I've not used lomount before so maybe you can pass flags to lomount directly.
>>I can tell you how to do it using losetup / mount separately:
>>
>>You can use losetup to bind a partition of the disk file to a /dev/loop device
>>and then mount that.  Use the -o option to specify the offset into the file
>>where the partition is to be found.  Use the -s option to specify the size of
>>the partition.
>>
>>You can probably fdisk the file (or, if that doesn't work, a loop device bound
>>to the whole file) to figure out where the partition boundaries are.
>>
>>HTH,
>>Mark
>>    
>>
>
>lomount does all of this internally. You can't pass it an offset into the file,
>only which partition you wish to mount. It calculates the offset by looking it
>up in the partition table.
>
>I'm wondering if he is trying to mount a COW image or some fancy-but-qemu-supported
>disk image. lomount has only been tested with raw images and sparse images, and
>is unlikely to work w/ anything else.
>
>  
>

[-- Attachment #1.2: Type: text/html, Size: 3380 bytes --]

[-- Attachment #2: sandholm.vcf --]
[-- Type: text/x-vcard, Size: 101 bytes --]

begin:vcard
fn:Tom Sandholm
n:Sandholm;Tom
email;internet:sandholm@gsinet.net
version:2.1
end:vcard

Re: [Qemu-devel] Mounting qemu virtual disk images from host linux os

[Tom Sandholm]

[-- Attachment #1.1: Type: text/plain, Size: 2051 bytes --]

Output of "sfdisk -l  debian02_sarge_hda.img".

====================================================
:nixsys:/qemu# sfdisk -l debian02_sarge_hda.img
Disk debian02_sarge_hda.img: cannot get geometry

Disk debian02_sarge_hda.img: 0 cylinders, 0 heads, 0 sectors/track
Warning: The partition table looks like it was made
  for C/H/S=*/16/63 (instead of 0/0/0).
For this listing I'll assume that geometry.
Units = cylinders of 516096 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls    #blocks   Id  System
debian02_sarge_hda.img1   *      0+   7903    7904-   3983584+  83  Linux
debian02_sarge_hda.img2       7904    8321     418     210672    f  W95 
Ext'd (LBA)
debian02_sarge_hda.img3          0       -       0          0    0  Empty
debian02_sarge_hda.img4          0       -       0          0    0  Empty
debian02_sarge_hda.img5       7904+   8321     418-    210640+  82  
Linux swap

>On Sun, Jan 16, 2005 at 06:33:54PM -0500, Tom Sandholm wrote:
>  
>
>>I found the lomount program that is supposed to mount your qemu 
>>virtual-disk-images; but it isn't working for me.
>>
>>I've installed debian "sarge" onto the virtual image, using ext3 
>>filesystems as default, with /dev/hda1 as the primary bootable linux 
>>partition (root).
>>
>>When I execute the lomount program, here's what I get.
>>============================================================
>>nixsys:/qemu# ./lomount -t ext3 -diskimage debian02_sarge_hda.img /mnt
>>mount: wrong fs type, bad option, bad superblock on /dev/loop0,
>>      or too many mounted file systems
>>nixsys:/qemu#
>>===========================================================
>>
>>What am I doing wrong?
>>Thanks!
>>
>>Tom Sandholm
>>
>>    
>>
>
>What partition do you want to mount? Are you sure it is the very first parition?
>What is the output of "sdisk -l debian02_sarge_hda.img" ?
>
>  
>
>>_______________________________________________
>>Qemu-devel mailing list
>>Qemu-devel@nongnu.org
>>http://lists.nongnu.org/mailman/listinfo/qemu-devel
>>    
>>
>
>
>  
>

[-- Attachment #1.2: Type: text/html, Size: 3450 bytes --]

[-- Attachment #2: sandholm.vcf --]
[-- Type: text/x-vcard, Size: 101 bytes --]

begin:vcard
fn:Tom Sandholm
n:Sandholm;Tom
email;internet:sandholm@gsinet.net
version:2.1
end:vcard

[Qemu-devel] Re: Mounting qemu virtual disk images from host linux os

[Heike C. Zimmerer]

Tom Sandholm <sandholm@gsinet.net> writes:

[ mounting disk partitions ]

You might want to try the following script I'm using for mounting
entire disk images.  I wouldn't say it's really finished, but it works
for me so far so I hope it might be of some use.

==========================================================================
~/bin/mount-hd-image
==========================================================================
#!/bin/sh
#
# Time-stamp: <2005-01-10 10:50:16 hcz>

pname=${0##*/}
usage() {
    iconv -f utf-8 <<EOF
Syntax: $pname [-t type] raw_image mountpoint
Function:  mount all partitions of a complete hd image
           optionally create a new one
opts:
        -c        create image if it doesn't exist
        -f        run fdisk on the image
        -m        make a filesystem on it
        -v        debug level (may be given multiple times)
        -x        Script debug (det -x)
EOF
}


if [ -z "$nopt" ]; then
    black=$(tput oc)$(tput sgr0) # Reset color + attribs
    bold=$(tput bold)
    red=$(tput setaf 1)
    green=$(tput setaf 2)
    blue=$(tput setaf 4)
    cyan=$(tput setaf 6)
fi 2>/dev/null

d_echo(){
    # $1: minimum debug level req'd for display
    # $2 ..: message
    if [ "$vopt" -ge "$1" ]; then
	shift
	echo "$@"
    fi
}

muss(){
    res=0
    d_echo 1 -e "$cyan$*$black"
    "$@" || abort "$*"
}

kann(){
    d_echo 1 -e "$cyan$*$black"
    "$@"
}

kann0(){
    d_echo 1 -e "$cyan$*$black"
    "$@" >/dev/null 2>&1
}

abort(){
    [ $# -gt 0 ] && echo -n "$pname: "
    for i in "$@"; do echo >&2 "$i"; done
    echo "$pname: aborted."
    exit 2
}

###################################################

vopt=0
dopt=/home/hcz/qemu
while getopts xvh-:fmc: argv; do
    case $argv in
        c) eval ${argv}opt="\$OPTARG";;
        f|m) eval ${argv}opt=\$argv;;
        v) eval ': $(('$argv'opt++))';;
        x) set -x;;
        *) usage; exit 0;;
    esac
done
shift $((OPTIND-1))

if [ $# -lt 2 ]; then
    abort "Not enough arguments"
    exit 1
fi

src="$1"
dest="$2"
shift; shift
mkdir -p "${dest##*/}" || abort "Can't create $dest"
uid=$(id -u)
gid=$(id -g)
losetup=/sbin/losetup
#types="vfat captive-ntfs ntfs"

if [ -n "$copt" ]; then
    if [ -f "$src" ]; then
        read -p "$src already exists.  Overwrite (Y/N*): "
        echo $REPLY | egrep -i '^ *y' || abort
    fi
    muss qemu-img create -f raw "$src" $copt
fi

if [ -n "$copt$fopt" ]; then
    muss /sbin/fdisk -C $(( $(stat "$src" -c "%s") / 255 / 63 /512)) "$src"
fi

x=$(/sbin/fdisk -ul "$src" 2>/dev/null)
[ -z "$x" ] && abort "Can't 'fdisk -ul $src'"
 echo $x | grep -q "This doesn't look like a partition table" &&
  abort "You selected a partition, but there is no valid partition table"
echo "Mounting harddisk image '$src' ..."
# win-b1              63    20964824    10482381    7  HPFS/NTFS
#      ()1            ()2   (                       ){3}  ()3
echo "$x" | grep -v 'Extended$' |
 sed -ne "s#^${src}p\?\([0-9]\+\)[ *]*\([0-9]*\) \+\([0-9]\++\? \+\)\{3\}\(.\+\)#\1 \2 \3 \4#p" |
 while read partno start id system; do
    d_echo 2 "p:$partno start:$start id:$id sys:$system"
    case $system in
        Extended) continue;;
        FAT*) types="msdos vfat"; mkfs=(/sbin/mkfs.msdos -F32);;
        HPFS/NTFS) types="captive-ntfs ntfs";;
        *) types=auto;;
    esac
    echo -n "Patition $partno: trying "
    mkdir -p "$dest/$partno"
    for type in $types; do
        echo -n "$type "
        res=99
        loop_dev=$($losetup -f) 
        [ -z "$loop_dev" ] && abort "Can't find a loop device"
        muss $losetup -o $((start * 512)) $loop_dev "$src"
        if [ -n "$mopt" ]; then
            [ -n "$mkfs" ] ||
               echo "Don't know how to make a $system file system"
            muss "${mkfs[@]}" $loop_dev
            # muss $losetup -d $loop_dev
        fi
        if kann0 sudo mount -o uid=$uid,gid=$gid,users \
                  -t  $type "$loop_dev" "$dest/$partno"; then
            kann0 sudo chmod a+xrw "$2"
            echo "... $type: ok."
            res=0
            break
        fi
    done
    [ "$res" -eq 0 ] || abort "Error mounting partition $partno"
done

[Qemu-devel] Re: Mounting qemu virtual disk images from host linux os

[Heike C. Zimmerer]

"Heike C. Zimmerer" <usenet04q4@gmx.net> writes:

> Tom Sandholm <sandholm@gsinet.net> writes:
>
> [ mounting disk partitions ]
>
> You might want to try the following script I'm using for mounting
> entire disk images.  I wouldn't say it's really finished, but it works
> for me so far so I hope it might be of some use.
>
> [ ~/bin/mount-hd-image ]

Let me add that only the disk image mounting part of the script has
seen some testing.  The options to create entire partitioned disk
images are pretty much untested.


- Heike