An ad- and malware-blocking utility for POSIX systems
- Description: Features
- Installation: Dependencies, Arch Linux, Other POSIX
- Configuration: Edit
hostsblock.conf
, Enable Timer, Enable Postprocessing - Usage: Configuring
sudo
, Manual Usage, UrlCheck Usage (examples) - FAQ
- News & Bugs: Upgrading to 0.999.8
- License
Hostsblock is a POSIX-compatible script designed to take advantage of the /etc/hosts
file to provide system-wide blocking of internet advertisements, malicious domains, trackers, and other undesirable content.
To do so, it downloads a configurable set of blocklists and processes their entries into a single HOSTS file.
Hostsblock also provides a command-line utility that allows you to configure how individual websites and any other domains contained in that website are handled.
-
Enhanced security - Runs as an unprivileged user instead of root. New: Includes systemd service files that heavily sandbox the background process.
-
System-wide blocking - All non-proxied connections use the HOSTS file (Proxied connections can be modified to use the HOSTS file)
-
Compression-friendly - Can download and process zip- and 7zip-compressed files automatically. (Provided that
unzip
andp7zip
are installed) -
Non-interactive - Can be run as a periodic background job without needing user interaction.
-
Extensive configurability - Allows for custom deny & allow listing, redirection,
post-processing scripting(now provided via systemd configuration), etc. -
Bandwith-efficient - Only downloads blocklists that have been changed, using http compression when available.
-
Resource-efficient - Only processes blocklists when changes are registered.
-
High performance blocking - Only when using dns caching.
-
Redirection capability - Enchances security by combating DNS cache poisoning.
-
Extensive choice of blocklists included - Allowing the user to choose how much or how little is blocked/redirected.
- curl
- A POSIX environment (which should already be in place on most Linux, *BSD, and macOS environments, including the following commands:
sh
(e.g. bash or dash,chmod
,cksum
,cp
,cut
,file
,find
,grep
,id
,mkdir
,mv
,rm
,sed
,sort
,tee
,touch
,tr
,wc
, andxargs
.
- sudo to enable the user-friendly wrapper script (highly recommended)
Unarchivers to use archive blocklists instead of plain text:
A DNS caching daemon to help speed up DNS resolutions:
If you use 127.0.0.1 as your blocking redirect address (redirecturl
in hostsblock.conf
), a pseudo-server that serves blank pages to remove boilerplate page and speed up page resolution on blocked domains:
Note that the default configuration gets no benefit from having a pseudo-server
If you have yaourt installed: yaourt -S hostsblock
or yaourt -S hostsblock-git
Or use one of the AUR packages: hostsblock, hostsblock-git
Don't forget to enable and start the systemd timer by running this:
$ sudo systemctl enable --now hostsblock.timer
Please check with your distribution to see if a package is available. If there is not, ask for it or contribute your own!
If you are a package maintainer, let me know so that I can post the instructions here.
First download the archive here or with curl like so: curl -O "https://github.com/gaenserich/hostsblock/archive/master.zip"
Unzip the archive, e.g. unzip hostsblock-master.zip
Execute the install.sh
script as root, which will guide you through installation.
By default, the configuration files are included in the /var/lib/hostsblock/config.examples/
directory. Copy them over to /var/lib/hostsblock/
to customize your setup.
Most of the hostsblock configuration is done in the hostsblock.conf
. This file is commented really well, so please read through it before first use:
# CACHE DIRECTORY. Directory where blocklists will be downloaded and stored.
#cachedir="$HOME/cache" # DEFAULT
# WORK DIRECTORY. Temporary directory where interim files will be unzipped and
# # processed. This directory will be deleted after hostsblock completes.
#
# #tmpdir="/tmp/hostsblock" # DEFAULT
# FINAL HOSTSFILE. Final hosts file that combines together all downloaded blocklists.
#hostsfile="$HOME/hosts.block" # DEFAULT
# REDIRECT URL. IP address to which blocked hosts will be redirect, either 0.0.0.0 or
# 127.0.0.1. This replaces any entries to 0.0.0.0 and 127.0.0.1. If you run a
# pixelserver such as pixelserv or kwakd, it is advisable to use 127.0.0.1.
#redirecturl="0.0.0.0" # DEFAULT
# HEAD FILE. File containing hosts file entries which you want at the beginning
# of the resultant hosts file, e.g. for loopback devices and IPv6 entries. Use
# your original /etc/hosts file here if you are writing your final blocklist to
# /etc/hosts so as to preserve your loopback devices. Give hostshead="0" to
# disable this feature. For those targeting /etc/hosts, it is advisable to copy
# their old /etc/hosts file to this file so as to preserve existing entries.
#hostshead="0" # DEFAULT
# DENYLISTED SUBDOMAINS. File containing specific subdomains to denylist which
# may not be in the downloaded denylists. Be sure to provide not just the
# domain, e.g. "google.com", but also the specific subdomain a la
# "adwords.google.com" without quotations.
#denylist="$HOME/deny.list" # DEFAULT
# ALLOWLIST. File containing the specific subdomains to allow through that may
# be blocked by the downloaded blocklists. In this file, put a space in front of
# a string in order to let through that specific site (without quotations), e.g.
# " www.example.com" will unblock "http://www.example.com" but not
# "http://subdomain.example.com". Leave no space in front of the entry to
# unblock all subdomains that contain that string, e.g. ".dropbox.com" will let
# through "www.dropbox.com", "dl.www.dropbox.com", "foo.dropbox.com",
# "bar.dropbox.com", etc.
#allowlist="$HOME/allow.list"
# CONNECT_TIMEOUT. Parameter passed to curl. Determines how long to try to
# connect to each blocklist url before giving up.
#connect_timeout=60 # DEFAULT
# RETRY. Parameter passed to curl. Number of times to retry connecting to
# each blocklist url before giving up.
#retry=0 # DEFAULT
# MAX SIMULTANEOUS DOWNLOADS. Hostsblock can check and download files in parallel.
# By default, it will attempt to check and download four files at a time.
#max_simultaneous_downloads=4 # DEFAULT
# BLOCKLISTS FILE. File containing URLs of blocklists to be downloaded,
# each on a separate line. Downloaded files may be either
# plaintext, zip, or 7z files. Hostsblock will automatically
# identify the file type.
#blocklists="$HOME/block.urls"
# REDIRECTLISTS FILE. File containing URLs of redirectlists to be downloaded,
# each on a separate line. Downloaded files may be either
# plaintext, zip, or 7z files. Hostsblock will automatically
# identify the file type.
#redirectlists="" # DEFAULT, otherwise "$HOME/redirect.urls"
# If you have any additional lists, please post a bug report to
# https://github.com/gaenserich/hostsblock/issues
Don't forget to enable and start the systemd timer with:
$ sudo systemctl enable --now hostsblock.timer
Hostsblock does not write to /etc/hosts
or manipulate any DNS caching daemons anymore. Instead, it will just compile a hosts-formatted file to /var/lib/hostsblock/hosts.block
. To make this file actually do work, you have one of two options:
Using a DNS caching daemon like dnsmasq offers better performance.
To use hostsblock together with dnsmasq, configure dnsmasq as DNS caching daemon. Please refer to your distribution's manual. For ArchLinux read the following: Wiki section.
After that, add the following line to dnsmasq.conf
(usually under /etc/dnsmasq.conf
) so that dnsmasq will reference the file:
addn-hosts=/var/lib/hostsblock/hosts.block
Enable and start hostsblock-dnsmasq-restart.path
:
$ sudo systemctl enable --now hostsblock-dnsmasq-restart.path
This has systemd watch the target file /var/lib/hostsblock/hosts.block
for changes and then restart dnsmasq
whenever they are found.
It is possible to have systemd
overwrite /etc/hosts
with the generated file.
Configure hostshead=
in hostsblock.conf
to make sure you don't remove the default system loopback address(es), e.g.:
hostshead="/var/lib/hostsblock/hosts.head"
Then put your necessary loopback entries in /var/lib/hostsblock/hosts.head
. For example, you can copy over your existing /etc/hosts
to this file:
$ sudo cp /etc/hosts /var/lib/hostsblock/hosts.head
$ sudo chown hostsblock:hostsblock /var/lib/hostsblock/hosts.head
$ sudo chmod 600 /var/lib/hostsblock/hosts.head
Enable and start hostsblock-hosts-clobber.path
:
$ sudo systemctl enable --now hostsblock-hosts-clobber.path
This has systemd watch the target file /var/lib/hostsblock/hosts.block
for changes and then copy /var/lib/hostsblock/hosts.block
to /etc/hosts
.
In its normal systemd-job configuration, hostsblock requires no interaction from the user aside from the steps above. If, however, you want to manually run the process, or to use the UrlCheck tool (hostsblock -c URL
), you need to configure sudo
:
Because hostsblock executes as a heavily sandboxed unpriviledged user (instead of root), you must configure sudo
to allow other users to manually execute it.
To do so, edit sudoers
by typing sudo visudo
and add the following line to the end:
%hostsblock ALL = (hostsblock) NOPASSWD: /usr/lib/hostsblock.sh
Add any users you want to be able to manually execute or use the urlcheck mode to the hostsblock
group:
$ sudo gpasswd -a [MY USER NAME] hostsblock
The wrapper script installed in your PATH will automatically use sudo to execute the main script as the user hostsblock
.
Without the -c URL
option, hostsblock will check to see if its monitored blocklists have changed. If it detects changes (or if forced by the -u
flag), it will download the changed blocklist(s) and recompile the target HOSTS file.
Help Options:
-h Show help options
Options:
-f CONFIGFILE Specify an alternative configuration file
-q Show only fatal errors
-v Be verbose
-d Be very verbose/debug
-u Force hostsblock to update its target file
With the -c URL
flag option, hostsblock can check and manipulate how it handles specific domains.
Note: The hostsblock-urlcheck
symlink is now officially depreciated. Use hostsblock -c
instead.
In addition to the above options, the following commands and subcommands can be used with hostsblock -c URL
:
hostsblock -c URL (urlCheck) Commands:
-s [-r -k] State how hostblock modifies URL
-b [-o -r] Temporarily (un)block URL
-e [-o -r -b] Add/remove URL to/from denylist
-a [-o -r -b] Add/remove URL to/from allowlist
-i [-o -r -k] Interactively inspect URL
hostsblock -c URL Command Subcommands:
-r COMMAND recurses to all domains on URL's page
-k COMMAND recurses for all BLOCKED domains on page
-o Perform opposite of COMMAND (e.g UNblock)
-b With "-e", immediately block URL
With "-a", immediately unblock URL
Note that the -o
subcommand turns a command into its opposite, e.g.
hostsblock -c URL -b -o
unblocks URLhostsblock -c URL -e -o
removes URL from the denylisthostsblock -c URL -a -o
removes URL from the allowlist
Once you have configured sudo
, you can execute the following as any user in the hostsblock
group:
See if "http://github.com/gaenserich/hostsblock" is blocked, denylisted, allowlisted, or redirected by hostsblock:
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -s
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -s -r
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -s -k
Block the domain containing "http://github.com/gaenserich/hostsblock" (that is, "github.com"):
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -b
Note that "blocking" (and "unblocking", i.e. -b -o
) a domain only works until the next time hostsblock refreshes /var/lib/hostsfile/hosts.block
, unless you use a blocklist that does include it. To permanently block this domain, use the denylist (-e
) command.
Permanently block (denylist) the domain containing "http://github.com/gaenserich/hostsblock" (that is, "github.com"):
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -e
Note that "denylisting" on its own will not block the target domain until hostblock refreshes. You can combine both "blocking" and "denylisting" in one command, however:
Permanently and immediately block the domain containing "http://github.com/gaenserich/hostsblock" (that is, "github.com"):
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -e -b
Temporarily unblock all blocked domains on "http://github.com/gaenserich/hostsblock" (helpful if the page isn't working quite right):
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -b -o -k
Interactively scan through "http://github.com/gaenserich/hostsblock", prompting you if you want the domains referenced therein to be blocked, denylisted, or allowlisted
$ hostsblock -c "http://github.com/gaenserich/hostsblock" -i -r
-
Why isn't it working with Chrome/Chromium?
- Because they bypass the system's DNS settings and use their own.
To force them to use the system's DNS settings, refer to this superuser.com question.
-
Hostsblock's systemd job fails with error "FAILED TO COMPILE BLOCK/REDIRECT ENTRIES FROM [...]" and leaves an empty
hosts.block.new
file.- You may have a blank line with a single space in your allowlist. Hostsblock matches that line with the space in between the IP address and the domain name that every single line has, i.e. it matches every single would-be entry in your target file. Remove the empty line, and hostsblock will function as expected.
For existing hostsblock users, please note the following changes in version 0.999.8:
Due to the shift to POSIX-shell compatibility, the list of blocklists to be downloaded cannot be held in hostsblock.conf
via the blocklists=
parameter. Instead, this parameter contains the path to a file that contains the list of URLs, e.g. /var/lib/hostsblock/block.urls
.
The new block.urls
file is simply a newline separated list of URLs without quotations. Whitespace and text after # are ignored. An example block.urls
file could look like this:
http://hosts-file.net/download/hosts.zip # General blocking meta-list
http://winhelp2002.mvps.org/hosts.zip
http://hostsfile.mine.nu/Hosts.zip
See the example block.urls
in the /var/lib/hostsblock/config.examples
directory for details.
Due to enhanced security and sandboxing, hostsblock no longer handles postprocessing on its own. Instead, users should use other systemd capabilities to replace the postprocess() {}
functionality.
Hostsblock comes with systemd service files that replicate the most common scenarios. See the directions above for instructions on how to enable them.
sudo
is no longer as widely used as before. The main systemd service no longer requires it. You only need it if you want to use the hostsblock -c URL
(urlcheck) utility. See the above directions for details.
- The
hostsblock-urlcheck
symlink is depreciated. Please usehostsblock -c URL
instead. - In UrlCheck mode, large hosts files will generate large temporary cache files that will eat up a lot of temporary storage. If you have a machine with little RAM (<6GB) and want to block a lot of domains, consider changing your $tmpdir to an HDD- or SSD-backed filesystem instead of the default tmpfs under
/tmp
. - UrlCheck mode will not be able to provide information on which blocklist blocked which domains anymore (annotation feature removed)
- Hostsblock uses 0.0.0.0 as default redirection IP address instead of 127.0.0.1. 0.0.0.0 theoretically offers better performance without the need of a pseudo-server.
- Systemd service now heavily hardened and sandboxed for enhanced security
- Fixed simultaneous download feature so that it actually does what it is supposed to
- Added processing support for source blocklists that just list domain names to be blocked, e.g.
ads.google.com
instead of0.0.0.0 ads.google.com
- Added support to read directly from zip and 7z files containing a single file without decompressing to a cache
- Optimized filters used to process domains with improved throughput
- If run with dash instead of bash, hostsblock has significant performance improvements
- Removed annotation feature to reduce dependencies and overall processing demands
- Vastly expanded list of potential blocklists (see
block.urls
)
- Supports POSIX shells (dash, ash, zsh) instead of just bash
- Removed GNU-specific utilities, relies only on POSIX options
- Should now run on *BSD and macOS (and perhaps even Android and iOS!) if proper POSIX environments are installed. UNTESTED
- User-facing command now a wrapper script that handles
sudo
execution for the user, reducing configuration demands - Significant performance improvements by moving from incremental to mass handling of domain names
- Added noninteractive commands
-s
(status),-b
(block),-e
(denylist),-a
(allowlist),-b -o
(unblock),-e -o
(dedenylist),-a -o
(deallowlist) - Interactive and noninteractive commands can now recursively handle URLs contained in target page (with
-r
subcommand), and even target just blocked domains (with-k
subcommand) - To minimize repeated writes, changes to target hosts file now don't write to file until after the whole process completes
Hostsblock is licensed under GNU GPL