PiShrink

PiShrink is a bash script that automatically shrink a pi image that will then resize to the max size of the SD card on boot. This will make putting the image back onto the SD card faster and the shrunk images will compress better. In addition the shrinked image can be compressed with gzip and xz to create an even smaller image. Parallel compression of the image using multiple cores is supported.

Usage

Usage: $0 [-adhrspvzZ] imagefile.img [newimagefile.img]

  -s         Don't expand filesystem when image is booted the first time
  -v         Be verbose
  -r         Use advanced filesystem repair option if the normal one fails
  -z         Compress image after shrinking with gzip
  -Z         Compress image after shrinking with xz
  -a         Compress image in parallel using multiple cores
  -p         Remove logs, apt archives, dhcp leases and ssh hostkeys
  -d         Write debug messages in a debug log file

If you specify the newimagefile.img parameter, the script will make a copy of imagefile.img and work off that. You will need enough space to make a full copy of the image to use that option.

• -s prevents automatic filesystem expansion on the images next boot
• -v enables more verbose output
• -r will attempt to repair the filesystem using aditional options if the normal repair fails
• -z will compress the image after shrinking using gzip. .gz extension will be added to the filename.
• -Z will compress the image after shrinking using xz. .xz extension will be added to the filename.
• -a will use option -f9 for pigz and option -T0 for xz and compress in parallel.
• -d will create a logfile pishrink.log which may help for problem analysis.

Default options for compressors can be overwritten by defining PISHRINK_GZIP or PSHRINK_XZ environment variables for gzip and xz.

Prerequisites

If you are running PiShrink in VirtualBox you will likely encounter an error if you attempt to use VirtualBox's "Shared Folder" feature. You can copy the image you wish to shrink on to the VM from a Shared Folder, but shrinking directctly from the Shared Folder is know to cause issues.

If using Ubuntu, you will likely see an error about e2fsck being out of date and metadata_csum. The simplest fix for this is to use Ubuntu 16.10 and up, as it will save you a lot of hassle in the long run.

Installation

wget https://raw.githubusercontent.com/Drewsif/PiShrink/master/pishrink.sh
chmod +x pishrink.sh
sudo mv pishrink.sh /usr/local/bin

Example

[user@localhost PiShrink]$ sudo pishrink.sh pi.img
e2fsck 1.42.9 (28-Dec-2013)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/loop1: 88262/1929536 files (0.2% non-contiguous), 842728/7717632 blocks
resize2fs 1.42.9 (28-Dec-2013)
resize2fs 1.42.9 (28-Dec-2013)
Resizing the filesystem on /dev/loop1 to 773603 (4k) blocks.
Begin pass 2 (max = 100387)
Relocating blocks             XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Begin pass 3 (max = 236)
Scanning inode table          XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Begin pass 4 (max = 7348)
Updating inode references     XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
The filesystem on /dev/loop1 is now 773603 blocks long.

Shrunk pi.img from 30G to 3.1G

macOS Support

It is not currently possible to run PiShrink natively in macOS. Included in the repo is a Dockerfile and docker-compose.yml that should allow you to run PiShrink on a Docker host, including a macOS one.

Note 1: The cloned repo is mounted as a Docker host mount so any output file should be created under /pishrink to have it persisted on the host.

Note 2: Only *.img is present in the provided .dockerignore, if you don't use this extension the created container will include your image, potentially consuming a significant amount of disk space.

Example: docker-compose run pishrink /pishrink/pishrink.sh /pishrink/someimage.img

Contributing

If you find a bug please create an issue for it. If you would like a new feature added, you can create an issue for it but I can't promise that I will get to it.

Pull requests for new features and bug fixes are more than welcome!