Findcruft

From Gentoo Linux Wiki

[edit] What is Cruft?

   "Over time, any system builds up cruft."
         Redmond concerned anonymous

Cruft 

Files and directories that don't belong to any package.

"This script tries to list all the cruft on a system with as few as possible false positives to help you keep your system in good working order. Note that plenty of packages drop extra files all over the place; any extra help is appreciated."

The term "cruft" imbues worth (or lack of worth) to files that may well be important.

[edit] Install findcruft

First you should add /usr/local/portage in your PORTDIR_OVERLAY variable in /etc/make.conf

Then download the ebuild in your overlay and emerge it:

mkdir -p /usr/local/portage
cd /usr/local/portage
wget http://ifp.loeber1.de/findcruft-ebuild.tar.bz2
tar xjf findcruft-ebuild.tar.bz2
rm -f findcruft-ebuild.tar.bz2
ebuild /usr/local/portage/app-portage/findcruft/findcruft-1.0.4-r1.ebuild digest
ACCEPT_KEYWORDS="~x86" emerge findcruft

[edit] Run the Script

root# findcruft | tee cruft.out

This will output all the files and dirs that are considered cruft to your shell and a file called cruft.out. Each file is on a new line. This will enable you to use the cat command along with rm to delete all of the cruft. Before you do this, though, you must look through the cruft.out and make sure there are no files you want to keep.

[edit] Cleaning up Cruft.out

root# nano -w cruft.out

Remove each file's line of the files you wish to keep. All lines in this file will remove the corresponding file.

[edit] Move the cruft

If you are not sure if you can delete the cruft you can move it in another directory to restore it easily later. You can use this script:

#!/usr/bin/perl
scalar(@ARGV) == 2 or die "Usage: mvcruft cruftfile backupdir\n";
open FILELIST, "$ARGV[0]" or die "Cannot open cruft file: $ARGV[0]\n";
$CRUFTDIR="$ARGV[1]";

while(<FILELIST>)
{
        /(.*\/)(.*)/;
        system("mkdir -p -- \"$CRUFTDIR$1\" && `[ -e \"$1$2\" ] || [ -h \"$1$2\" ]`  && mv -- \"$1$2\" \"$CRUFTDIR$1$2\"");
}

root# chmod 755 mvcruft
root# mvcruft /path/to/cruft.out /home/backup_dir

[edit] Deleting the Cruft

root# xargs rm -rf < cruft.out

[edit] Re-run the Script

root# rm cruft.out

Re-run the script to make sure everything was removed that you want to be.

[edit] False Positives

if you think they are false positives and know to which package they belong you can create/edit a rule file in /usr/lib/findcruft/<category>/<package>.
Then, we can exchange these rule files and build new rule archives from time to time.
If it is a false positive due to a personal/local setting on your machine, you can add it into /usr/lib/findcruft/ignore; there it won't interfere with the exchanged rules.

[edit] False Positives rules needed

You can here add your false positives:

• To exclude by default:

echo /var/log
echo /var/lock
echo /var/run
echo /var/lib/init.d
echo /etc/udev/rules.d/10-local.rules
echo /etc/udev/scripts
echo /lib/udev/devices
echo /var/www
echo /usr/src

• app-admin/eselect

echo /etc/profile.csh
echo /etc/eselect/ld-mtimedb 

• app-admin/eselect-wxwidgets

cruftdir /var/lib/wxwidgets

• app-admin/metalog

echo /var/run/metalog.pid

• app-emulation/vmware-workstation

echo /usr/bin/vmware

• app-misc/beagle

cruftdir /var/lib/cache/beagle/indexes
cruftdir /var/lib/cache/beagle

• app-misc/uptimed

echo /var/spool/uptimed/bootid
echo /var/spool/uptimed/records

• app-office/abiword

echo /usr/bin/abiword

• app-portage/eix

echo /var/cache/eix

• app-portage/kuroo

cruftdir /var/cache/kuroo

• app-shells/bash

echo /usr/bin/bash

• app-text/docbook-xml-dtd

cruftdir /etc/sgml

• app-text/tetex

cruftdir /var/lib/texmf

• dev-db/mysql

echo /var/run/svc-started-mysqld

• dev-db/phpmyadmin

echo /var/db/webapps/phpmyadmin

• dev-libs/openssl

cruftdir /etc/ssl/certs

• media-video/nvidia-glx

echo /usr/lib/libGLcore.so
echo /usr/lib/xorg/modules/extensions/libglx.so

• net-fs/samba

echo /etc/samba/smb.conf

• net-misc/dhcp

cruftdir /var/lib/dhcp

• net-misc/ntp

echo /etc/ntp.conf
echo /var/run/ntpd.pid

• net-misc/rsync

echo /var/run/rsyncd.pid 

• net-misc/tor

echo /etc/tor/torrcecho 

• net-proxy/squid

cruftdir /var/cache/squid

• sci-misc/boinc

cruftdir /var/lib/boinc

• sys-apps/dbus

echo /var/lib/dbus/machine-id
echo /var/run/dbus/system_bus_socket

• sys-apps/hal

echo /var/run/hald.pid
echo /var/lib/cache/hald/fdi-cache

• sys-apps/sed

echo /usr/bin/sed

• sys-devel/distcc

echo /var/run/distccd/distccd.pid

• sys-kernel/module-rebuild

echo /var/lib/module-rebuild
echo /var/lib/module-rebuild/moduledb
echo /var/lib/module-rebuild/moduledbe

• sys-libs/glibc

echo /etc/locale.gen
for locale in `locale -a | grep -v "^C\$" | grep -v "^POSIX\$"`; do
       echo "/usr/lib/locale/$locale"
done

• sys-libs/gpm

echo /var/run/gpm.pid 

• sys-process/at

echo /var/run/atd.pid

• www-servers/jboss

cruftdir /var/lib/jboss
cruftdir /var/cache/jboss

[edit] Notes

• Original script is here: http://catmur.co.uk/~ed/main/my-bin/cruft
• More scripts by Ed Catmur: http://catmur.co.uk/~ed/main/my-bin/
• Original forum post: http://forums.gentoo.org/viewtopic.php?t=152618
• findcruft forum post: http://forums.gentoo.org/viewtopic.php?t=254197

[edit] Alternative, smaller version

If you don't want to clean your whole system but just check the /etc/ directory for orphaned files you might be interested in this slightly shorter script.

#!/usr/bin/python
# This script tries to find orphaned files in /etc
# based upon the information from portage.
# 
# by Phillip 'Firebird' Berndt
# 
import sys, glob, re, os, os.path
staticProtections = [
	"/etc/config-archive", "/etc/passwd", "/etc/shadow",
	"/etc/runlevels", "/etc/group", "/etc/gshadow",
	"/etc/make.conf", "/etc/make.profile", "/etc/magic",
	"/etc/ssl", "/etc/runlevels", "/etc/ssl", "/etc/gtk",
	"/etc/env.d", "/etc/portage", "/etc/gconf"
	]

sys.stderr.write("Creating list of protected files\n")
protectedFiles = ["/etc"]
for fileList in glob.glob("/var/db/pkg/*/*/CONTENTS"):
	for file in open(fileList).read().split("\n"):
		match = re.match("^(dir|obj) (/etc/[^ ]+)", file)
		if match: protectedFiles.append(match.group(2))
   
sys.stderr.write("Comparing against files in /etc/\n")
def checkFile(file):
	if file in protectedFiles: return False
	for other in staticProtections:
		if file.find(other) == 0: return False
	return True
for file in os.popen("find /etc").read().split("\n"):
	if checkFile(file): print file