btrfs-undelete

btrfs-undelete

#!/bin/bash


# btrfs-undelete
# Copyright (C) 2013 Jörg Walter <[email protected]>
# Copyright (C) 2018 Franz-Josef Faerber <[email protected]>
# This program is free software; you can redistribute it and/or modify it under
# the term of the GNU General Public License as published by the Free Software
# Foundation; either version 2 of the License, or any later version.


# Use the Unofficial Bash Strict Mode (Unless You Looove Debugging)
# http://redsymbol.net/articles/unofficial-bash-strict-mode/
set -euo pipefail


# Ensures device isn't mounted any more.
# If variable $auto_unmount is set, tries to unmount.
function ensure_not_mounted {
	local device="$1"
	local mountpoint="$(cat /proc/mounts | grep -Po '(?<=^'"$device"' )([^ ]+)')"
	
	if [[ -n "$mountpoint" ]]; then
		echo "Device $device is still mounted on $mountpoint ."
		if [[ -n "$auto_unmount" ]]; then
			echo "Unmounting."

			if ! umount "$device"; then
				echo "Umount failed. The following processes have open files on this mountpoint:"
				lsof_for_path "$mountpoint"
				echo "Aborting."
				exit 2
			fi
		else
			echo "Aborting."
			exit 3
		fi
	fi
}

# List open files for given path prefix
function lsof_for_path {
	local path_prefix="$1"

	lsof -Fpcfn \
       		| awk '
			BEGIN					{ print "COMMAND\tPID\tNAME" }
			/^p/					{ pid = substr($0, 2) }
			/^c/					{ command = substr($0, 2) }
			substr($0,1,5)=="n'$path_prefix'"	{ print command "\t" pid "\t" substr($0, 2) }'
}

function summary_and_exit {
	echo
	echo "The following files were restored in directory $out :"
	find . -ls
	exit 0
}


all_versions_flag=
keep_empty_files_flag=
auto_unmount=

while getopts "aeu" opt; do
	case $opt in
		a)
			all_versions_flag=1
			;;
		e)
			keep_empty_files_flag=1
			;;
		u)
			auto_unmount=1
			;;
		\?)
			echo "Invalid option: -$OPTARG" >&2
			exit 1
			;;
	esac
done
shift "$((OPTIND-1))"


if [[ $# -lt 3 ]]; then
	echo "Usage: $0 [OPTION]... <dev> <file/dir> <dest>" 1>&2
	echo
	echo "This program tries to recover the given files or directories."
	echo "Unless otherwise specified with the options, restoration will"
	echo "stop after a generation of matching non-empty files was restored."
	echo
	echo "<dev> must not be mounted, otherwise this program may appear"
	echo "to work but find nothing."
	echo
	echo "<file/dir> must be specified relative to the filesystem root,"
	echo "obviously. It may contain * and ? as wildcards."
	echo
	echo "<dest> must be a writable directory with enough free space"
	echo "to hold the files you're trying to restore."
	echo
	echo "Options:"
	echo "  -a  Restore all found versions of the matching files"
	echo "      (instead of only the ones with the highest generation"
	echo "      id). Files will be suffixed by the generation it first"
	echo "      occurred in."
	echo "  -e  Restore empty files."
	echo "  -u  Try to unmount automatically."
	exit 1
fi

# Canonicalize potentially relative path $1 because we will change dir
dev=$(readlink -f "$1")
file="$2"

file="${file#/}"
file="${file%/}"
regex="${file//\\/\\\\}"

# quote regex special characters
regex="${regex//./\.}"
regex="${regex//+/\+}"
regex="${regex//|/\|}"
regex="${regex//(/\(}"
regex="${regex//)/\)}"
regex="${regex//\[/\[}"
regex="${regex//]/\]}"
regex="${regex//\{/\{}"
regex="${regex//\}/\}}"

# treat shell wildcards specially
regex="${regex//\*/.*}"
regex="${regex//\?/.}"

# extract number of slashes in order to get correct number of closing parens
slashes="${regex//[^\/]/}"

# build final regex
# example: ^/(|home(|/username(|/Desktop(|/.*))))$
# see man btrfs restore, option --path-regex for details
regex="^/(|${regex//\//(|/}(|/.*${slashes//?/)}))\$"

roots="$(mktemp --tmpdir btrfs-undelete.roots.XXXXX)"
out="$(mktemp --tmpdir="$3" -d btrfs-undelete.XXXXX)"
cd "$out"

trap "rm $roots" EXIT
trap "rm -r $out &> /dev/null; exit 1" SIGINT

ensure_not_mounted "$dev"

echo -ne "Searching roots..."
btrfs-find-root "$dev" \
	| sed -n -r -e 's/^Well block ([0-9]+)\(gen: ([0-9]+).*/\2\t\1/p' \
	| sort -rn \
	> $roots || exit 1
echo

i=0
max="$(wc -l <$roots)"

while read gen bytenr; do
	((i+=1))
	# Add some spaces at the end to erase potential garbage from last echo:
	echo -ne "Trying root gen=$gen bytenr=$bytenr... ($i/$max)       \r"

	gendir="gen_$gen"
	genpath="$out/$gendir"
	mkdir "$genpath"

	# Restore a generation
	btrfs restore -t $bytenr --path-regex "$regex" "$dev" "$genpath" &>/dev/null
	if [ "$?" = 0 ]; then
		found=$(find "$genpath" ! -type d | wc -l)
		if [ $found -gt 0 ]; then
			echo -e "\nRestored $found file(s) into $genpath"
			
			# Remove already restored files in other generations:

			cd "$gendir"

			while IFS= read -r -d '' myfile; do
				# The asterisk itself must not be quoted 
				# in order to enable bash filename expansion
				find "../"*"/$myfile" \
					! -path "../$gendir/*" \
					-type f \
					-exec cmp --silent "$myfile" {} \; \
					-delete
			done < <(find "." -type f -print0)

			#

			if [[ -z "$keep_empty_files_flag" ]]; then
				find "." -type f -size 0c -exec echo "Restored {} but it's empty. Deleting it." \; -delete
			fi

			if [[ -z "$all_versions_flag" ]]; then
				echo "Found first match. Exiting."
				cd ..
				mv "$gendir"/* .
				rmdir "$gendir"
				summary_and_exit
			fi

			cd ..
		fi
	fi

	found=$(find "$gendir" ! -type d | wc -l)
	if [[ $found -eq 0 ]]; then
		# Just empty directories. Remove them.
		rm -r "$gendir"
	fi
done <$roots

# Add a newline to the last "Trying..." clause
echo

if [[ -z "$all_versions_flag" ]]; then
	rm -r $out
	echo "Didn't find '$file'"
	exit 1
else
	# Copy all remaining files together in one directory
	gens_mixed="gens_mixed"
	mkdir "$gens_mixed"

	while read gen bytenr; do
		gendir="gen_$gen"
		
		if [[ -d "$gendir" ]]; then
			cd "$gendir"

			# Suffix files by generation id
			find "." ! -type d -exec mv {} {}.$gendir \;
			
			# Poor man's move-and-merge-directories:
			cp -al "." "../$gens_mixed"
			cd ..
			rm -r "$gendir"
		fi
	done <$roots

	mv "$gens_mixed"/* .
	rmdir "$gens_mixed"
	summary_and_exit
fi

This is an attempt to enhance Changaco/btrfs-undelete .

• restore multiple versions
• when restoring multiple versions, only keep files that actually differ in file content
• fail on errors
• error message if device is still mounted
• ability to unmount or list open files if unmount fails
• print summary of restored files

Can you write some examples with different <file/dir> formats? wildcards, etc?

Wow I haven't looked at that code for years and can't remember a thing from scratch, sorry. But this fork shouldn't be too far away from Changaco/btrfs-undelete?