Question

21

Is my command to clone a drive optimal?

rated 0 times [ 21] [ 0] / answers: 1 / hits: 39558 / 1 Year ago, sat, december 24, 2022, 6:06:27

I have done half an hour reading around to prepare to clone my hard drive. It has multiple partitions, including a Windows partition. I am going to buy a very large external hard drive for the backup. I would like to be able to use this clone to restore the whole drive in case something goes wrong (I'm about to do some OS re-shuffling). I want do learn how to do this using dd, as I like low-level tools that don't require installing anything.

I found the following useful code from the ubuntu forums (entered from a root shell using a live CD):

dd if=/dev/hda of=/dev/hdb & pid=$!

while kill -USR1 $pid; do sleep 1; done

(I know that I will have to edit the input and output locations.) However I have two questions. The first one is very noobie: this command is split across two lines. Surely when I press enter after the exclamation mark it will start the process?

Two, on other sites it recommended entering block size. Like this:

# dd if=/dev/hda conv=sync,noerror bs=64K of=/mnt/sda1/hda.img

I don't know anything about block sizes. Is 64K right? It looks like my block size is 512 bytes from the following, the output of sudo fdisk -ul:

Disk /dev/sda: 750.2 GB, 750156374016 bytes

255 heads, 63 sectors/track, 91201 cylinders, total 1465149168 sectors

Units = sectors of 1 * 512 = 512 bytes

Sector size (logical/physical): 512 bytes / 4096 bytes

I/O size (minimum/optimal): 4096 bytes / 4096 bytes

Disk identifier: 0xc3ffc3ff



   Device Boot      Start         End      Blocks   Id  System

/dev/sda1   *          63   143364059    71681998+   7  HPFS/NTFS/exFAT

Partition 1 does not start on physical sector boundary.

/dev/sda2       976867328  1465147391   244140032    7  HPFS/NTFS/exFAT

/dev/sda3       143364094   976867327   416751617    5  Extended

Partition 3 does not start on physical sector boundary.

/dev/sda5       143364096   162895871     9765888   82  Linux swap / Solaris

/dev/sda6       162897920   205864959    21483520   83  Linux

/dev/sda7       205867008   976867327   385500160   83  Linux



Partition table entries are not in disk order



Disk /dev/mapper/cryptswap1: 10.0 GB, 10000269312 bytes

255 heads, 63 sectors/track, 1215 cylinders, total 19531776 sectors

Units = sectors of 1 * 512 = 512 bytes

Sector size (logical/physical): 512 bytes / 4096 bytes

I/O size (minimum/optimal): 4096 bytes / 4096 bytes

Disk identifier: 0x433bb3a7



Disk /dev/mapper/cryptswap1 doesn't contain a valid partition table

Thank you.

Answers

Only authorized users can answer the question. Please sign in first, or register a free account.

fittecanap

Add To Favorites

Follow

Total Points: 322

Total Questions: 100

Total Answers: 105

Location: Israel

Member since Tue, Nov 17, 2020

4 Years ago

fittecanap questions

1 Hibernate entry in the menu on 22.04

Tue, Nov 23, 21, 22:02, 3 Years ago

1 How do I upgrade from 20.04 to 22.04 now? Are most guides simply wrong?

Wed, Feb 1, 23, 18:08, 1 Year ago

1 I am having this issue while running 'sudo apt-get update', "Problem with MergeList" or "status file could not be parsed or opened."

Fri, May 12, 23, 17:09, 1 Year ago

1 Foreground task scheduling in Linux similar to Windows's Task Scheduler

Wed, May 10, 23, 02:52, 1 Year ago

1 Updates Warning

Tue, Aug 9, 22, 17:36, 2 Years ago

View All

answered 1 Year ago azaaburge · Accepted Answer

Progress

The command you listed

dd if=/dev/hda of=/dev/hdb & pid=$!

while kill -USR1 $pid; do sleep 1; done

is a nice two-liner to get the the progress of dd on a regular basis. I use a very similar one too. Looks good. Found it here perhaps?

Blocksizes with `dd`: alignment and performance

You can add a block size in which the operations take place. It does not matter what the block size of the underlying block device is to get the operation done equally well, but for performance reasons you might want to pick one that suits your needs.

First of all, there's the alignment thing. In case your block device operates as 512KiB (like flash drives do), it would be very unfortunate to run dd with bs=512 (bytes) as this will cause 1024 writes (!) for each block from the device perspective. In practise it won't be this bad as writes are buffered and taken into one go, but during syncs it can still amplify the amount of writes a lot.

Then also consider the plain CPU usage overhead when dealing with a very large amount of small operations. It's just more efficient to take megabytes at once when copying over large amounts of data.

My best practice is to start with 1MB as that is a nice multiple of most set ups, including RAID stripe sizes, LVM extent sizes, etc. On my laptop with SSD I tend to see a slight improvement using 10MB as block size, whereas I don't see it anymore on my physical hard drive.

Last block

Don't worry about the drive/volume size not being a multiple of the block size. The last block dd will copy will be adjusted to match the last bit of data on it. You can see whether the last block had a different size by looking at the output.

18335302+0 records out

The +0 means that it was an exact match, a +1 means that it wasn't. No big deal.