The last paragraph of the SliTaz 3.0 Release Notes says the following about the liveCD:

The ISO image now uses a 'hybrid' system: it can also be copied onto an USB stick without formating it (using dd).

Does anyone know how to do this?

migrated from stackoverflow.com Oct 29 '11 at 11:11

This question came from our site for professional and enthusiast programmers.

  • 3
    DEAR GOD BE CAREFUL WITH dd! It's nicknamed disk destroyer for a reason. – Rob Oct 29 '11 at 20:49
  • @Rob Can you list the dos and donts please? I would hate to waste my drive trying to make it bootable. – Gui Imamura Dec 12 '15 at 12:42
up vote 20 down vote accepted

dd if=/path/to/your/isofile of=/your/usb/disk try this.

  • should I include ~/ in path ? – weis26 Oct 29 '11 at 10:14
  • 2
    @weis26: You can, ~ is just replace with the current users home directory path. Use it if either the if or of paths are in your home directory sure... – Matt Joiner Oct 29 '11 at 10:20
  • 2
    @weis26 Depends on whether you want to use a relative path or absolute path. By the way, using dd would make your usb device read-only and to make it a 'normal' usb device you shall re-partition the device. – starrify Oct 29 '11 at 10:21
  • 7
    bs=8M or so would speed this up quite a bit... also, should mention using mount to see what disk is mounted at /media/83... – Kimvais Oct 29 '11 at 10:26
  • @starrify - I can't figure out what you mean by saying dd "would make your usb device read-only". dd does not change capabilities of devices, it just shoves data from input to output. – Florenz Kley Sep 22 '17 at 13:43

If you want to be able to view the progress or get an ETA, you can add Pipe View (pv) into the mix, e.g.:

dd if=<path to input file> | pv -s <size e.g. 1377M> | dd of=<path to target device>

This will give output like:

850MiB 0:05:18 [6.44MiB/s] [===================> ] 61% ETA 0:03:16

  • If you have pv then pv /path/to/input/file > /path/to/target/device is already enough – Dzamo Norton Apr 11 at 7:32
  • One of the options in dd is 'status=progress' to get periodic transfer statistics. – user65913 Apr 25 at 14:29

Usually you want to write to the device. Use df to find the entry:

/dev/sdd1 31933168 27690992 4242176 87% /media/Fat

Then with dd, use the of=/dev/sdd, not /media/Fat or /dev/sdd1. If you give it a block size it's usually quicker, so I usually give it something like -bs=1M.

  • 3
    df will only show mounted filesystems. Bad Idea(TM) to dd stuff to devices that contain mounted filesystems.. – Florenz Kley Oct 16 '14 at 21:18
  • @FlorenzKley So instead I should use...? – Olle Härstedt Sep 19 '17 at 18:02
  • Use df if you can but only a device with a file system can be mounted. I sometimes use 'gparted', which will give you it's assigned device id for use in DD even sans an fs. DD is alluded to the DD card in IBM JCL and as a data dump utility. Triple check where you are writing. I use large block size along with the 'conv=fdatasync,' which forces a physical write to the device before finish. – user65913 Sep 20 '17 at 23:52
  • @OlleHärstedt before using a tool that writes directly to the blockdevice, remove upper layers that use it at the moment (and mostly expect they are the only ones writing to it). In this example: umount /media/Fat or umount /dev/sdd1. To get information about block device usage on Linux, suggest you use lsblk – Florenz Kley Sep 22 '17 at 13:37

Note: This is a generic, universal approach that should work on almost any *nix system. If you're having trouble with other recommendations, then give this one a try. So you want to run dd (disk destroyer)? Before attaching your block device, cat /proc/partitions; then attach your device and again cat /proc/partitions. The difference will indicate your target device. If there's no difference then you lack the correct driver to recognize it or have a hardware fault.

Here's what mine looks like on my laptop, before and after, respectively:

jcholsap@T430s:~$ cat /proc/partitions 

major minor  #blocks  name
8        0  117220824 sda
8        1  117219328 sda1
11       0    1048575 sr0
253      0  109109248 dm-0
253      1    8077312 dm-1

jcholsap@T430s:~$ cat /proc/partitions 

major minor  #blocks  name
8        0  117220824 sda
8        1  117219328 sda1
11       0    1048575 sr0
253      0  109109248 dm-0
253      1    8077312 dm-1
8       16   30670848 sdb

Use the third column, size in kilobytes, to verify your block device. If you're nervous, and you should be, you can further verify your device with dmesg. Here it is on my laptop:

jcholsap@T430s:~$ dmesg | tail -30
[   10.197383] thinkpad_acpi: EC reports that Thermal Table has changed
[   34.599234] random: crng init done
[   35.133277] rfkill: input handler disabled
[ 9737.190954] usb 4-2: new SuperSpeed USB device number 2 using xhci_hcd
[ 9737.211756] usb 4-2: New USB device found, idVendor=0781, idProduct=5580
[ 9737.211762] usb 4-2: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 9737.211765] usb 4-2: Product: Extreme
[ 9737.211768] usb 4-2: Manufacturer: SanDisk
[ 9737.211770] usb 4-2: SerialNumber: AA010507161254310422
[ 9737.244492] usb-storage 4-2:1.0: USB Mass Storage device detected
[ 9737.245540] scsi host6: usb-storage 4-2:1.0
[ 9737.245746] usbcore: registered new interface driver usb-storage
[ 9737.247465] usbcore: registered new interface driver uas
[ 9738.259485] scsi 6:0:0:0: Direct-Access     SanDisk  Extreme          0001 PQ: 0 ANSI: 6
[ 9738.260340] sd 6:0:0:0: Attached scsi generic sg2 type 0
[ 9738.260698] sd 6:0:0:0: [sdb] 61341696 512-byte logical blocks: (31.4 GB/29.3 GiB)
[ 9738.260919] sd 6:0:0:0: [sdb] Write Protect is off
[ 9738.260925] sd 6:0:0:0: [sdb] Mode Sense: 53 00 00 08
[ 9738.261142] sd 6:0:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
[ 9738.267727] sd 6:0:0:0: [sdb] Attached SCSI removable disk
[ 9738.496445] UDF-fs: warning (device sdb): udf_load_vrs: No anchor found
[ 9738.496450] UDF-fs: Scanning with blocksize 512 failed
[ 9738.505586] UDF-fs: warning (device sdb): udf_load_vrs: No anchor found
[ 9738.505589] UDF-fs: Scanning with blocksize 1024 failed
[ 9738.519311] UDF-fs: warning (device sdb): udf_load_vrs: No anchor found
[ 9738.519324] UDF-fs: Scanning with blocksize 512 failed
[ 9738.528153] UDF-fs: warning (device sdb): udf_load_vrs: No anchor found
[ 9738.528157] UDF-fs: Scanning with blocksize 1024 failed
[ 9738.533493] UDF-fs: INFO Mounting volume 'UDF Volume', timestamp 2011/04/12 09:38 (1000)
[10100.854101] perf: interrupt took too long (2519 > 2500), lowering kernel.perf_event_max_sample_rate to 79250

There's all the details of my thumb drive. Ah, now I'm convinced I have the right device!

But my attached device may have auto-mounted. For dd, that's a problem. So to be certain, let's unmount it with umount (note the difference in spelling). I'll run the command twice because the error message let's me know it succeeded:

jcholsap@T430s:~$ sudo umount /dev/sdb
jcholsap@T430s:~$
jcholsap@T430s:~$ sudo umount /dev/sdb
umount: /dev/sdb: not mounted.

Now, because everything in Linux is treated as a file, dd calls the target block device, in my case /dev/sdb, an output file. So, the option-argument pair to use, here for my thumb drive, with the dd command is of=/dev/sdb. You want to get this right or else dd could mean disk destroyer for you. Here it is on my laptop:

jcholsap@T430s:~$ sudo dd of=/dev/sdb if=/home/jcholsap/centos7.iso bs=4M

There's no output from the running dd command. You'll just get a blinking cursor until it finishes, which can be a long time. If you interrupt the process then you'll have a partial image copy - probably not what you wanted.

SEE ALSO Full documentation at: http://www.gnu.org/software/coreutils/dd or available locally via: info '(coreutils) dd invocation'

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.