8

I have a puppy linux img that is 8gb but I need it to fit on a 6gb drive. How do I resize the ext2 partition on the img?

Benjily3
  • 81
  • 1
  • 1
  • 2

2 Answers2

8

NOTE Make a backup, anything can go wrong ... I ran all those as root :

  1. bind a loop device to the image: losetup /dev/loop10 $image_file
  2. refresh partitions: partprobe /dev/loop10
  3. adjust partition size: gparted /dev/loop10
  4. undo the loop: losetup -d /dev/loop10
  5. remove unwanted space: truncate -s -${SIZE}G $image_file (you might want to calculate the exact size based on values from fdisk/cfdisk * sector size)

NOTE You might need to double check the partitions' UUIDs, mine seem to have changed and I had to update fstab

abo
  • 81
  • 1
  • 4
  • Isn't the truncate command eg. like this: `truncate -s 6GB /yourimagefile.img` or `truncate -s 6442450944 /yourimagefile.img`? But even though I took the end sector and multiplied by the sector size it must somehow have been wrong, because afterwards `parted` told me "Can't have partition outside the disk" and the image doesn't work... – TheStoryCoder Jul 22 '19 at 10:24
  • Also, the `resizepart` command in `parted` was very quick (I used `parted` instead of `gparted`). Does it just cut of the partition at the given size? What if files are "physically" located outside the partition boundary - will they just be discarded? – TheStoryCoder Jul 22 '19 at 10:27
  • for fdisk, taking into account only the end sector is not exact. The right calculation is: ( start_sector + partition_sector_count) * sector_size. You use this amount A (in bytes) for truncate: `truncate --size={A} $image_file ` – sugarman Jul 09 '20 at 19:22
  • This answer works great, but the truncate line kept throwing an error. So I added the approximate sizes of the partitions plus a 10M safety margin (eg. truncate -s 250M my.img) and that worked great and left a little unallocated space at the end of the last partition. – Ken H May 04 '23 at 20:07
3

You'll want to shrink the filesystem - any kind of logical volume management or similar containers - then the image. I'm going to assume you're talking about a RAW disk image.

Steps at a glance:

  • Ensure the image is not being accessed (ex: lsof)
  • Shrink the filesystem (ex: resize2fs)
  • Perform a filesystem check (ex: fsck)
  • Shrink any LVM or other kind of containers (if needed)
  • Shrink the disk image (ex: dd to a new image with skip or use qemu-img)
  • Fsck again, test that it works!

Alternative:

  • Create a new image and copy the data / MBR / etc. over.
  • Clone the image with something like partclone

Similar questions:

Reference:

Community
  • 1
Criveti Mihai
  • 1,116
  • 7
  • 10
  • no need to assume as they stated "img file" not image file, but despite that, i am not able to follow your steps, can you elaborate on how to make my 2TB .img file to 1TB ? – oemb1905 Oct 31 '21 at 00:19