Introduction 
 Being able to make one's own version of SystemRescueCd was a frequently requested feature. 
It's now possible to make your own version of SystemRescueCd. 
 
 This means you will be able to add files to the sysrcd.dat compressed image file, 
which will be the filesystem of SystemRescueCd. For example, you can add script files that make a 
full backup of a server, or a binary program that you need. It's also possible to add data files, 
in order to have everything on the CDRom (data and system files which allow you to restore data). 
For now, it's not possible to change the kernel. If you need a kernel option that is disabled, please contact us. 
 
 The goal is to produce a new ISO image file, with an updated sysrcd.dat. In order to be able to 
make your own version of SystemRescueCd, you will need a recent official version of this CDRom, 
and a partition with read-write access from within linux, with at least 500 MB. If you make a very l
arge version of the CDRom, you may need a lot more space. 
 
 Be aware that this operation requires a large amount of memory (both RAM and space on the hard disk). 
If you don't have 128 MB of RAM, and at least 300 MB free on your hard disk, it will fail. 
You can create a swap partition in order to avoid problems with the cloop step in the script.

you don't have enough ram: add more swap.
maybe a
dd if=/dev/zero of=somewhere/SWAPFILE bs=1M count=size
mkswap somewhere/SWAPFILE
swapon somewhere/SWAPFILE

All the personalization is processed from the CDRom system with a script. This means you have 
to burn the official version, boot your computer from this official SystemRescueCd version, 
and then you are able to make your own version. The script needs to be run from the CDRom system, 
because it's easier to use this way. I am sure all programs are installed. If this script was 
designed to run from any linux installed system, you would have to install several programs, 
such as a module for the kernel, and it would be a lot more complex. 
 Here are the instructions to follow in order to make your own ISO image. 
We assume that the /dev/hda1 partition is formatted for linux (ext2, ext3, reiserfs, reiser4, xfs, jfs, ...), 
and that it has at least 500 MB free.It must really be a linux diskspace, because other filesystem, 
such as FAT, NTFS don't supports symbolic links and files permissions. If you only have a FAT 
partition available, you can create a virtual linux filesystem image inside the FAT partition: 
# mount /dev/hda1 /mnt/fat
# cd /mnt/fat
# dd if=/dev/zero of=fsimage bs=1M count=500
# mke2fs -F -q -N 50000 fsimage
# mount -t ext2 -o loop fsimage /mnt/custom


 If you use the "cdcache" boot option, the original SystemRescueCd disc must be mounted when 
you run the first step (extract) of the script. 

procedure 
 This procedure will make a directory "customcd" at the root of this partition. 
All the files located in other directories won't be affected by this script. 
If you have an old version of this personalized CDRom in this directory, please make a backup. 
All files of "/dev/hda1/customcd/" may be lost! Note that you must keep the name "/mnt/custom" that 
we used in this example. 
You can't simply use "/mnt/another-name". 
Step-01: Mount the working partition: 
# mount /dev/hda1 /mnt/custom

 If you don't want the main temporary directory to be written to the root of /mnt/custom, 
you can choose another directory. 
For example, if the free space is an NFS share, you can try this: 
# mount -o loop /mnt/nfs/usr/dir1/dir2 /mnt/custom

Step-02: Extract the current files from the cloop image: 
# /usr/sbin/sysresccd-custom extract

 If there is no error, all the files of the cloop image should have been copied in /mnt/custom/customcd/files/ 
Step-03: Customize files (optional) 
 You have to customize the files which you have just extracted. 
You can do this from the SystemRescueCd, but it's also possible to restart the computer and do it from 
any other installed linux system. If you choose to restart, you will have to restart from the SystemRescueCd, 
and mount files again (follow step-01 again). All the changes have to be made in the following directory: 
(This is the root of the new filesystem) 
/mnt/custom/customcd/files/ 
Step-04: Create the new cloop image 
 In this step, all files from /mnt/custom/customcd/files/ are read, and put inside the new cloop image. 
The new cloop image is copied in 
"/mnt/custom/customcd/isoroot/sysrcd.dat", but you don't have to concern yourself with this. 
# /usr/sbin/sysresccd-custom cloop 300 20000

 This creates a cloop for a filesystem that uses 300 MB, and with less than 20000 files. 
If the command complains about space missing, try to increase these two parameters. 
For example, if your filesystem contains 50000 files and uses 10 MB, 
the previous command will complain (no space left on device). 
Step-05: Add files to the ISO image (optional) 
 By default, you only need the big cloop image file and the isolinux directory. 
The CDRom can work with only these two items on the disc. In this step, 
you can add other files to the ISO image, outside of the cloop image. 
For example, you can put big backup files inside the CDRom, and outside the cloop image, 
in order to avoid overloading the cloop. If you are interested in adding such files, 
you just have to copy the files along with the sysrcd.dat, in this directory: "/mnt/custom/customcd/isoroot" 
# cp -a my-files /mnt/custom/customcd/isoroot

 You may add an autorun script to the root of the CDRom, outside of the cloop image. 
If you put an autorun script there, it can be run at startup. It allows to automake many tasks, 
such as doing a backup. See the autorun section for more details. 
# cp autorun /mnt/custom/customcd/isoroot
# chmod 755 /mnt/custom/customcd/isoroot/autorun

 Of course, you need to make sure that the size of the isoroot directory is not too 
big for your CDR (usually 700 MB). 
You can use the "du /mnt/custom/customcd/" command for that. 
Step-06: Select your keymap (optional) 
 In the official version, SystemRescueCd asks you to select a keymap at startup. 
If you want to avoid seeing this menu and define a permanent keymap, you can run this command: 
# /usr/sbin/sysresccd-custom setkmap <your-keymap>

 The second parameter is the name or the number of your keymap in the menu. 
For example, french users will use this: 
# /usr/sbin/sysresccd-custom setkmap fr

Step-07: Create the new ISO image 
 You just have to type this command. The second parameter to provide is the volume name for the CDRom. 
# /usr/sbin/sysresccd-custom isogen my_srcd

Step-08: Synchronize your disks 
 Before leaving the system and rebooting, you should make sure all your disks are cleanly unmounted. 
The best way to be sure not to lose data is to umount all mounted partitions, 
especially the temporary one (/dev/hda1 in this example), and then type "sync" 
# cd / ; umount /mnt/custom ; sync

Step-09: Test and burn the ISO image 
 The new final ISO image is located in 
"/mnt/custom/customcd/isofile/sysresccd-new.iso". You can burn it with cdrecord, Nero, 
or any other burning software. If you don't want to burn a CDR, you can test it with VMWare. 
You have to tell this software that the CDRom drive to use is the virtual ISO-image file drive. 
Result of your changes 
 When you start your computer with your own SystemRescueCd version, 
you can see all the files in /mnt/cloop. If your new files belong to a standard directory, 
you can access it without a problem. If you made a new directory in the root of the filesystem, 
it will appear in /mnt/cloop, but you won't see that in the root of the system. 
In fact, the root of the system contains symbolic links to /mnt/cloop. You may have to make the link yourself.

 Summary of rules 
 At startup, a list of locations are checked against the presence of autorun files. 
They are, successively: 
if the ar_source= parameter was passed at startup, the root directory of the given mountable device. 
This is primarily intended for floppy disks : 
fb800 ar_source=/dev/fd0

 but may be used with hard disk partitions too : 
fb800 ar_source=/dev/hda2

 to look for autorun scripts in 2nd partition of 1st IDE hard disk 
fb800 ar_source=/dev/sdb6

 to look in 6th partition of 2nd SCSI hard disk 
 and with network shares. The syntax is : 
fb800 ar_source=nfs-server:/nfs/exported/directory

 for NFS shares 
fb800 ar_source=//samba-server/share

 for SMB shares (recommended without user/password) 
 Source devices are mounted to /mnt/autorun, allowing writing data into them if the device is not write-protected. 
the root directory of the CDrom 
the superuser home directory (/root) 
the /usr/share/sys.autorun directory 
 If autorun files are found in some location, they are run and the process ends. 
Last two locations require rebuilding of SystemRescueCd and are to be used for advanced or test purposes. 
 
 In each source location, there are two possible modes of operation : 
simple one : if a shell script named "autorun" is found, it is run 
more flexible : if "autorun#" scripts are found (# is a digit from 0 to 9) and either 
autoruns= boot parameter was NOT specified, or 
autoruns= boot parameter value contains # 
 (ex: "fb800 autoruns=0,1,4" - In this example, only autorun0, autorun1 or autorun4 scripts may be run if present. 
Other scripts, such as autorun2, autorun3 and autorun5, will be ignored) 
 then authorized scripts are run in alphabetic order. Whenever a script returns a non-zero code, 
the processing stops and next scripts are not run. 
You can prevent any "autorun#" execution with an autorun= boot parameter without any digit, 
ex: "autoruns=no". This has no effect on "autorun" script. 
 Although it is possible to put both "autorun" and "autorun#" scripts, this should probably be avoided. 
 
 The script has to be any valid shell script. Perl, python, ... etc. scripts are not supported at this time. 
Windows end-of-line terminators are translated to allow running shell scripts written with a MS editor.

Examples of autorun scripts 
 #!/bin/bash
 (
   ifconfig eth0 | head -n 3
   fdisk -l /dev/hda
 ) | tee -a /mnt/autorun/report
 sync
 exit 0

 This script will accumulate infos about eth0 network interface and disk partitions into 
a file named "report" on first floppy drive. BIOS boot sequence must specify CD first, 
ar_source= boot parameter has to be correctly set, diskette must be write-enabled, 
CD and diskette must be inserted at startup time. 
 
 If network is automatically configured at boot (DHCP), you may send infos through a NFS or 
samba share for example. 
 
 Don't forget to give autorun location at startup. Say that you share through NFS a writable 
directory named /nfs/backup on host myserver, you have to specify : 
fb800 ar_source=myserver:/nfs/backup 
 at SystemRescueCd boot prompt and to put in the shared directory an "autorun" script like this one : 
 #!/bin/bash
 date=$(date +%Y%m%d)
 cat /dev/hda | gzip > /mnt/autorun/$date.hda.gz && exit 0
 echo $date hda backup error
 exit 1

 If sufficient space is provided in NFS share, this will copy there a compressed image 
of your first IDE disk. 
 
 You must ensure that name resolution works if you will access NFS server by its name 
instead of its IP address.

Requirements 
 Of course, you need a DVD burner, a disc. You also need an installed Linux system, with dvd+rw-tools. 
If you don't have this software, you can download on the following homepage: 
  http://fy.chalmers.se/~appro/linux/DVD+RW/
You need to know the device name of your DVD burner. If you don't know it, you may find it with dmesg. 
You should see something like this: 
  # dmesg | grep DVD 
  hdd: PIONEER DVD-RW DVR-106D, ATAPI CD/DVD-ROM drive
  hdd: ATAPI 32X DVD-ROM DVD-R CD-R/RW drive, 2000kB Cache, UDMA(33)

 In this example, the two output lines allow to know the device name. It's "/dev/hdd". 
 This was tested under Linux-2.4.26, with dvd+rw-tools-5.17, on a DVD+RW, and the device was "Pioneer DVR-106D". 
It is supposed to work with all kind of DVD discs, and all DVD burners.

First step: burn the official SysRescCd ISO image 
 In this first step, we will just burn the official 100 MB image on the DVD disc. 
 growisofs -Z /dev/hdd=systemrescuecd-x86-0.2.14.iso

 Of course, you have to replace "hdd" with the device name of your burner, and update the path to the ISO file.

Second step: add your own files 
 Now, we will add files to the disc. All the files you want to add have to be stored in the same directory. 
In this example, we suppose files are stored in /home/user/mydvd-data-files/: 
 growisofs -M /dev/hdd -J -R /home/user/mydvd-data-files/

 It should be possible to add files several times by running this command several times.

How to install SystemRescueCd on harddisk
Introduction 
 The idea is to install SystemRescueCd tools onto a Linux partition to easily recover your system 
in case something goes wrong. You will have an installed system that works like other linux systems 
installed by other distributions. If you just want to be able to run SystemRescueCd from your hard disk, 
and have an entry in your bootmanager (Lilo or Grub), there is an easier method, 
that is explainted in the next chapter. 
 
 WARNING: Be aware that this operation may give you an useless system if you don't strictly 
follow instructions below. 
Usage 
 All the job is processed from the official SystemRescueCd CDRom. 
This means you have to burn the official version and boot your computer from it. 

procedure 
step-01: Create a new partition 
 You have to resize an existing partition and then create an empty ext2fs partition. 
Note: only ext2fs is supported at the moment. 
 # run_qtparted

 I will suppose you now have an NTFS partition /dev/hda1 and an ext2 partition /dev/hda2. 
You have to update this to your setup. 
step-02: Mount the new partition 
 # mount /dev/hda2 /mnt/custom

step-03: Extract the current files from the cloop image 
 # sysresccd-custom extract-nosizecheck
 # rm -rf /mnt/custom/customcd/isoroot
 # mv /mnt/custom/customcd/files/* /mnt/custom
 # rm -rf /mnt/custom/customcd

step-04: Copy the SystemRescueCD kernel 
 # cp /mnt/cdrom/isolinux/vmlinuz1 /mnt/custom/boot

step-05: Customize setup to your hard disk 
 Add this line to /mnt/custom/etc/fstab file 
/dev/hda2  /  ext2  errors=remount-ro  0  1

 Copy this file from /usr/share/sysresccd/hdinstall/lilo.conf.in to /mnt/custom/etc/lilo.conf 
  lba32
  boot = /dev/hda
  map = /boot/.map
  prompt
  install = /boot/boot-menu.b
  delay = 50
  vga = normal
  default=win
  image = /boot/vmlinuz1
        root = /dev/hda2
        label = sysrcd
        read-only
  other = /dev/hda1
        label = win
        table = /dev/hda

 WARNING: do not forget to replace /dev/hda1 and /dev/hda2 with your values if you want your system to boot. 
step-06: Run lilo 
  # mount -t devfs none /mnt/custom/dev
  # chroot /mnt/custom lilo
  # umount /mnt/custom/dev

step-07: Umount the new partition 
  # umount /mnt/custom

Changes as boot time 
 At boot time, you should now have a prompt like this one: 
LILO 22.5.1

 This means that LILO is installed and running. If you let the timer going down (about 5 secondes), 
your default target will be launched. With previous lilo.conf file, it will be Windows. 
 If you wish to test your new SystemRescue partition, press left shift key. You will get 
LILO 22.5.1 boot:

 Then enter "sysrcd". 
 If you do not remember your target list, hit tabulation key and LILO will show you 
LILO 22.5.1 boot:
win     sysrcd

Easy install SystemRescueCd on harddisk 

Subsections 
Introduction 
Requirements 
First step: copy important files 
Second step: update the bootmanager
Introduction 
 Since version 0.2.19, there is an easy way to boot SystemRescueCd from the harddisk. 
This method allows to use SystemRescueCd as you do with the CDRom version. 
But this method does not require to insert the disc in the drive each time you need it. 
It also allows to boot SystemRescueCd if you don't have an optic drive in your computer. 
You can use another operating system to process the installation of SystemRescueCd, and enable it. 
 Thus, this method is recommanded if you often use SystemRescueCd and you want to avoid having to insert the disc in the drive each time.
Requirements 
 There are two steps in this installation: first you will have to copy main SysRescCd files on a 
partition of your disk. Then, you will have to configure your bootmanager. 
The installation process requires a partition with about 130 MB free space, 
and you must have a working linux bootmanager installed, such as Lilo or Grub. 
(maybe it can work with other boot managers). 
 The partition where you will copy SystemRescueCd files does not have to be formatted for Linux. 
It can have any file system that Linux-2.6 is able to mount in read-only, 
and with a file system your boot manager is able to use. 
It mean you can use a FAT16 or FAT32 partition or a Linux native partition (ext2, ext3, reiserfs, 
reiser4, xfs, jfs). 
You can store sysrcd.dat on an NTFS partition, but this file system cannot store vmlinuz1 and initrd1, 
since there is a problem with grub and lilo when loading a file from an NTFS disk at boot time. 
This partition may be already used for something else (Linux installed on), or it can be a partition you made only for SystemRescueCd. 
 In the second step, you will need the Linux device name of this partition 
(a name such as /dev/hda1, /dev/sda1, /dev/hdb5, ...)
First step: copy important files 
 This step is very trivial: you just have to copy the three SysRescCd main files to the partition. 
If that's a FAT16/FAT32 partition, you can perform the copy with any operating system. Of course, 
if you choose a Linux native partition, you will have to copy from Linux 
(you can copy from SystemRescueCd itself). 
 You must create a directory /sysrcd/ that's in the root of the partition. 
The name and the position of this directory cannot be changed. 
 Now, take the disc with the latest SystemRescueCd version (you may prefer to mount the ISO 
image with mount under Linux or you can use a software such as DaemonTools under Windows 
if you did not burn the disc), in order to have the three main files. 
 Finally, just copy sysrcd.dat, initrd1 and vmlinuz1 from the CDRom to the directory you made (/sysrcd/). 
Two of these files are located in isolinux inside of the CDRom. 
Second step: update the bootmanager 
 Now, you must update your bootmanager (Lilo or Grub). 
 You need the Linux device name of the partition you have chosen 
(a name such as /dev/hda1, /dev/sda1, /dev/hdb5, ...). 
If you don't know this name, you should boot SystemRescueCd from the CDRom drive, and list 
all devices seen by the linux kernel. Just type "cat /proc/partitions" in a console. 
You may also run QtParted to identify the partition. 
 We will have to add several lines to the configuration file of the bootmanager (usually /etc/lilo.conf for Lilo, 
and /boot/grub/menu.lst or /boot/grub/grub.conf for grub). You have to personnalize the configuration given there. 
You must replace /dev/xxx with the device name of your partition. With grub, 
you will have to give the grub device name for the partition you have choosen. 
It's usually a name like (hd0,0) instead of hda1 (hd0,1) instead of hda2, ... Then, 
you may update several options related to your keyboard (setkmap=xxx), and your video settings (vga=xxx) 
You may have a look at isolinux.cfg that's in the CDRom, if you don't know the values you can use for these options. 
 First, here is an example of the Lilo configuration (edit /etc/lilo.conf) The path /mount-point/ 
must be remplaced with the mount point for the partition at the time you run lilo, and you must 
remplace /dev/xxx with the linux device name for the partition files were copied on. 
image=/mount-point/sysrcd/vmlinuz1
root=/dev/ram0
initrd=/mount-point/sysrcd/initrd1
label=SysRescCd
append="bootfrom=/dev/xxx init=/linuxrc setkmap=us"
vga=5

 Here is an example of Grub configuration (edit menu.lst or grub.conf in /boot/grub/) In this example, 
files are located in /dev/hda5. The grub device name is (hd0,4) 
 Warning: There is no line break for the line beginning with kernel. 
title    SystemRescueCd 
kernel   (hd0,4)/sysrcd/vmlinuz1 root=/dev/ram0 vga=5 
         bootfrom=/dev/xxx init=/linuxrc setkmap=us
initrd   (hd0,4)/sysrcd/initrd1
boot

 If your bootmanager is Lilo, don't forget to run lilo after lilo.conf was edited. With Grub, 
this step is not required. 
 Now, the installation process is complete. SystemRescueCd should appear in your bootmanager 
menu the next time you start your computer.

How to install SystemRescueCd on an USB-stick 

Subsections 
Introduction 
Installation from Linux 
Installation from Windows 
Booting from the USB stick
Introduction 
 This chapter explains how to install SystemRescueCd on an USB-stick. It allows to use this system 
from an USB stick (mini hard disk) instead of using a CDRom. First, you need SystemResuceCd 0.2.18 or newer, 
and an USB stick with enough space. If you use the official version, you need 128 Mb or better. 
Of course, you need a computer that supports USB, with almost a free USB socket. 
You BIOS must be able to boot on USB hard disks, and it must be defined before other devices in 
the boot devices order. 
Installation from Linux 
 If you don't have Linux installed, you can burn SystemRescueCd on a CDRom, and perform the install 
from this system itself. 
step-01: Find the device name 
 You need the device of the USB stick. In most cases, Linux detects this device as an SCSI hard disk. 
In other words, it is detected as /dev/sda1 (first partition of first SCSI device), or /dev/sdb1 if you 
have another SCSI device. Since kernel 2.6.8 USB sticks may be seen as /dev/uba, /dev/ubb, ... 
 Here is how you can know the device name. First, start your computer under Linux with the USB stick unplugged. 
Then, load all USB modules (you may need to load other USB modules first). 
This command should not be required if the USB storage support is build-in the kernel. 
If modprobe can't find this module, it might be normal. 
 # modprobe usb-storage

 Now, you have to plug your USB stick, and have a look at the kernel messages. 
You may find the device name. 
 # dmesg | tail -n 50

 Here is a example of report from the kernel: 
Initializing USB Mass Storage driver...
usb.c: registered new driver usb-storage
USB Mass Storage support registered.
hub.c: new USB device 00:02.0-2, assigned address 3
scsi1 : SCSI emulation for USB Mass Storage devices
  Vendor: EXATEL    Model: i-BEAD100         Rev: 0001
  Type:   Direct-Access                      ANSI SCSI revision: 02
WARNING: USB Mass Storage data integrity not assured
USB Mass Storage device found at 3

 Then, you should try to mount the device: 
 # mkdir /mnt/usbstick
 # mount -t vfat /dev/sda1 /mnt/usbstick

 If there is no space enough left on the device (about 120 MB), you will have to erase the 
current files (all data will be lost): 
 # rm -rf /mnt/usbstick/*

step-02: Copy files from the CDRom 
 Now, you have to copy the most important files. The SystemRescueCd disc must be mounted on /mnt/cdrom. 
You can mount the ISO file too, in order to avoid burning a CD-R. The stick is mounted on /mnt/usbstick. 
 # cp /mnt/cdrom/syslinux/syslinux.cfg /mnt/usbstick/
 # cp /mnt/cdrom/sysrcd.dat /mnt/usbstick/
 # cp /mnt/cdrom/isolinux/* /mnt/usbstick/
 # cp /mnt/cdrom/bootdisk/* /mnt/usbstick/

step-03: Make the disk bootable 
 Now, you have to unmount the USB stick, and make it bootable with syslinux. Of course, 
you have to replace /dev/sda1 with the device name of your stick: 
 # umount /mnt/usbstick
 # syslinux /dev/sda1
 # sync

Mounting an NTFS partition with full Read-Write support 

Subsections 
Introduction 
Getting the Windows drivers 
Mounting NTFS with full support 
Remarks
Introduction 
 NTFS (New Technology File System) is a very powerful file system provided with NT versions of Windows (Windows-NT, Windows-2000, Windows-XP). Even if Windows still supports FAT, the Windows system disk is often 
formatted in the NTFS format. 
 Linux offers a complet read support for NTFS partitions for a long time. But the NTFS write support is still partial. 
The write support was improved in linux-2.6.15 that is provided in SysRescCd-0.2.16. It's now possible to write inside existing files, even if it makes the file grow. The write operation may fails, 
but it cannot damage the NTFS partition's structure. It still not possible to create new file 
with the current state of the NTFS driver. The NTFS write support is safe but incomplete. 
You should use it if you just want to commit a change in a document, 
and it allows to clean infected files by a virus on a system. 
 If you just want a read support or a basic write support on NTFS disks, you just have to mount 
it with the following command (you have to replace xxx with the device name of your disk): 
 # mount -t ntfs /dev/xxx /mnt/part

 If you really need a complete NTFS Write support, you will have to use CaptiveNtfs. 
It is not so easy to use, and it is a lot slower. 
This projects uses the NTFS supports from the Windows kernel itself. 
These drivers have to be copied from an installed Windows XP system. 
You can't get these files from Windows NT4 or Windows 2000. But once you get the drivers from XP, 
you can mount any NTFS version. 
It Captive does not want the NTFS driver that comes with the system that is installed on your computer, 
you will have to download Service Pack 1 for an English version of Windows XP, 
and extract ntfs.sys and ntoskrnl.exe. 
 This chapter explains how to use the Captive-Ntfs support from SystemRescueCd.
Getting the Windows drivers 
 We cannot distribute the Windows drivers because these files are distributed under a Microsoft license. 
Then you have to get these files yourself from the Windows system partition first. 
You have to copy these files inside the linux system: 
step-01: Mount the Windows system partition 
 Here is an example about how to mount the Windows-XP partition, supposed to be /dev/hda1: 
 # mkdir /mnt/win
 # mount -o ro -t ntfs /dev/hda1 /mnt/win

step-02: Copy the drivers files 
 Windows is supposed to be installed in C:/WINDOWS. Sometimes, the user can choose another directory, 
such as C:/WINNT. 
 # cp /mnt/win/WINDOWS/system32/drivers/ntfs.sys /var/lib/captive
 # cp /mnt/win/WINDOWS/system32/ntoskrnl.exe /var/lib/captive

 If a file or a directory does not exists, try to change the Case ("System32" instead of "system32" for example). 
You can find these files on an Windows NT4 or Windows 2000 system, 
but only Windows XP files will work with Captive Ntfs. 
step-03: Umount the Windows partition 
 # cd / ; umount /mnt/win
Mounting NTFS with full support 
 Now the drivers are copied, you can mount any NTFS partition, even if it does not have windows installed on: 
 # chmod 770 /
 # mount.captive-ntfs -o --rw /dev/hda1 /mnt/win

 Now, you must be able to write file on the mounted partition: 
 # cd /mnt/win
 # touch creating-an-empty-file
 # cp -a /path/to/a/directory /mnt/win

 Of course, you will have to unmount this partition when finished to work on: 
 # cd / ; umount /mnt/win
Remarks 
 For the moment, DF is not supported with Captive-Ntfs. 
It mean the program that shows the disk space usage (DF=Disk Free) won't show any information 
about the mounted NTFS partition. If you really need this informations, 
you have to remount the partition read-only, with the official linux NTFS driver. 
 More informations about Captive-Ntfs: 
 http://www.jankratochvil.net/project/captive/ 

System boot floppy disks 

Subsections 
Introduction 
Overview of the boot floppy disks
Introduction 
 Of course, the most important feature of SystemRescueCd is the Linux system that is run from the CDRom, 
and that allows to do many administration tasks. Since version 0.2.9, 
SystemRescueCd allows to work from virtual bootable floppy disks. 
These virtual floppies offers the same feature you could have with a physical floppy disk 
you boot on when you start your computer. For example, several users need system DOS programs. 
They start from a DOS floppy disk, and they can run these tools. 
SystemRescueCd comes with several system floppy disks. 
The floppy list appears when you press F2 in the very first prompt. This page shows many option on the top, 
and floppy disks images at the bottom. In the next section, we will have an overview of each of them.
Overview of the boot floppy disks 
FreeDos 
 Just type this image name at prompt: 
freedos

 FreeDos is a free operating system compatible with MS-DOS. 
It allows to run DOS programs such as Partition Magic or Drive Image without any MS-DOs license. 
MemTest+ 
 Just type this image name at prompt: 
memtest

 Memtest+ 1.0 is an update of Memtest-x86 3.0. It's a memory tester. 
This program will tell you if your physical memory is damaged or not. 
GAG: Graphical-Boot-Manager 
 Just type this image name at prompt: 
gag

 GAG is a boot manager, such as Lilo, Grub, or Boot-Magic. It aims to be very nice and easy to use. 
Ranish Partition Manager 
 Just type this image name at prompt: 
ranish

 Ranish Partition Manager is a low level partition manager such as Fdisk, but it's more powerful. 
Aida 
 Just type this image name at prompt: 
aida

 Aida is an advanced diagnostic system program. It will detect many pieces of hardware of your computer, 
such as the Processor, chipset, mainboard, network card, ... 
It's very useful when you want to download a driver and you don't know which one to take. 
Dban (Darik's Boot and Nuke) 
 Just type this image name at prompt: 
dban

 Dban is a bootdisk that allows to destroy all data of a computer. 
Be aware it's dangerous to use. 
You can use dban if you want to be sure nobody can be able to read data of a computer. 
For example, if you want to give your computer to someone, and your computer had important data, 
you can use dban before giving the computer. It will wipe all data from all hard disks. 
Offline NT Password and Registry editor 
 
 Just type this image name at prompt: 
ntpass

 This program allows to change system informations written in the Windows settings. 
The most interesting one is the ability to change the password of a user account. 
Now, you won't have to reinstall Windows if you forget the password of the administrator account. 
You just have to edit the password with this utility. 
This program also allows to edit the Windows registry. It work even when Windows is installed on an NTFS disk.

Access a Windows computer with Samba 
 The samba package allows you to access a windows computer on the network. 
For example, if the windows server (192.168.10.3) has a share named "my-share", 
then you can mount it with these commands: 
 # mkdir /mnt/win
 # mount -t smbfs -o lfs //192.168.10.3/my-share/ /mnt/win/
 # cd /mnt/win
 # umount /mnt/win

 Samba is supported in the kernel, and samba tools are provided (smbcacls, smbcontrol,smbfilter, 
smbmount, smbspool, smbtar, smbclient, smbd, smbmnt, smbpasswd, smbstatus, smbumount). 
It's important not to forget option "lfs" (Large File Support), 
because it allows to handle files that are larger than 2 GB. 
Big file are often used when making a backup or an image file. 
Without this option, the copy would fail, with an error. 
Large files are supported since version 0.2.12 (linux kernel 2.4.25). 
 More help with samba: "http://www.samba.org/" 

Mount remote FTP/SSH shares as local file systems 
 If you want to access files located on an FTP server, there is a new very powerful way to do this. 
The "userLand FileSystem" allows you to mount the share, 
and work on the remote files just as you would work on any local files. 
With all these file systems, you can unmount the share with the standard umount command. 
 Here is an example of how to mount an FTP file system in /mnt/ftp as anonymous (read only) 
 # mkdir /mnt/ftp
 # lufsmount ftpfs://ftp.server.org /mnt/ftp
 # cd /mnt/ftp
 # umount /mnt/ftp


 Here is an example of how to mount a private FTP account in /mnt/ftp2 
 # mkdir /mnt/ftp2
 # lufsmount ftpfs://login:password@ftp.server.org /mnt/ftp2
 # cd /mnt/ftp2
 # umount /mnt/ftp2

 Here is an example of how to mount an SSH file system in /mnt/ssh as anonymous (read only) 
 # mkdir /mnt/ssh
 # lufsmount sshfs://login@ssh.server.org /mnt/ssh
 # cd /mnt/ssh
 # umount /mnt/ssh

Graphical Partition Tools 
 This CDRom provides PartitionMagic clones, which aim to be very easy to use, and very powerful. 
Running these tools is very easy, but the FrameBuffer is required. 
Please refer to the boot process if you don't have the FrameBuffer. 
You can run QtParted by typing this command : "run_qtparted". 
You will be asked the mouse model. Try to give the right answer. 
If the system crashes, reboot and give another answer to this question. 
If you can't use your mouse, try to exit the software via the keyboard, 
and restart the software with another option for the mouse. 
Another graphical partition tool is available: PartGui. 
You can run PartGui by typing "run_partgui". PartGui is not developped anymore, then QtParted is a better choice. 

Backing up your partition table 
 sfdisk is a tiny program that allows you to save your partition table. 
If you delete a partition accidently, you can try to restore the table. 
The backup is done into a file. Of course, you must keep this file on a removeable media (such as a floppy disk), 
or on another computer. If your hard disk is damaged, you won't be able to read the backup from the disk itself ! 
 
 This command allows you to backup the full partition table of your disk HDA: "sfdisk -d /dev/hda > bak-hda" Now, 
this command allows you to restore from the file: "sfdisk /dev/hda < bak-hda". 
Of course, this command is very dangerous, because it overwrites the existing partition table !!! Before restoring, 
you should back up the current version, even if it seems to be damaged. 
This would allow you to undo the restoration. 

Backing up the contents of a partition 
 Partimage (client and server) are provided. 
Partimage allows you to save an existing partition into an image file. 
If something goes wrong, you can restore the partition from the image. 
Here is the official homepage of the project: http://www.partimage.org/. 
A semi-text interface is provided, and this should not be very hard to use. 
You need a partition with a lot of free space in order to store the image file. 
 
 Since version 0.2.14 of SysRescCd, two partimage versions are provided: 
The default version is compiled without SSL and without login support. 
The client is available by typing "partimage", and the server with "partimaged". 
You should use this client in most cases, and when you connect a partimaged server that does not support SSL. 
The secured version is compiled with SSL and login support. 
The client is available by typing "partimagessl", and the server with "partimagedssl". 
You should use this client when you connect a partimaged server that expects and SSL encrypted connexion with login.

Clam AntiVirus 
 Clam AntiVirus is a free anti-virus software provided since version 0.2.3 of SystemRescueCd. 
It provides several commands: 
"clamscan -r /path/to/dir". It scan all files of this directory an search for viruses. 
"freshclam". This command updates the virus definitions. Of course, 
you need to be connected to internet for updating. 
 Before scanning for virus, you have to update virus definitions. 
Because this is run from a CDRom, you may have read-only related problems. 
That's why you have to specify a read-write directory for definitions. 
Here is how to update definitions and make a scan: 
Step-01. Make empty directories 
        # mkdir /virdefs /mnt/testpart
        # chmod 777 /virdefs

Step-02. Copy the current virus definitions 
        # cp /usr/share/clamav/* /virdefs/

Step-03. Download the latest virus definitions 
        # freshclam --datadir /virdefs

Step-04. Mount the partition to be scanned 
        # mount /dev/hda1 /mnt/testpart

Step-05. Perform the scan on the files. 
        # clamscan -r -d /virdefs /mnt/testpart

 If you really don't have any internet connexion, you can scan for virus without updaing virus definitions. 
Of course, it's less safe, and only old virus can be detected. Here are the steps: 
Step-01. Make empty directories 
        # mkdir /mnt/testpart

Step-02. Mount the partition to be scanned 
        # mount /dev/hda1 /mnt/testpart

Step-03. Perform the scan on the files. 
        # clamscan -r /mnt/testpart