Difference between revisions of "Polipo"

From ArchWiki
Jump to: navigation, search
m (Run polipo as different user (not nobody))
(Add warning (polipo is no longer maintained))
 
(14 intermediate revisions by 12 users not shown)
Line 1: Line 1:
 
[[Category:Proxy servers]]
 
[[Category:Proxy servers]]
 +
[[ja:Polipo]]
 
[[zh-CN:Polipo]]
 
[[zh-CN:Polipo]]
{{Article summary start}}
+
{{Warning|Polipo is no longer maintained, It's recommended to use another solution (squid, privoxy...) instead.}}
{{Article summary text|Polipo is a light weight http proxy server for your small network. Polipo caches web data.}}
+
{{Related articles start}}
{{Article summary heading|Related}}
+
{{Related|Squid}}
{{Article summary wiki|Squid}}
+
{{Related articles end}}
{{Article summary end}}
+
  
 
From [http://www.pps.jussieu.fr/~jch/software/polipo/ Polipo's site]:
 
From [http://www.pps.jussieu.fr/~jch/software/polipo/ Polipo's site]:
Line 11: Line 11:
 
:"''Polipo is a small and fast caching web proxy (a web cache, an HTTP proxy, a proxy server). While Polipo was designed to be used by one person or a small group of people, there is nothing that prevents it from being used by a larger group.''"
 
:"''Polipo is a small and fast caching web proxy (a web cache, an HTTP proxy, a proxy server). While Polipo was designed to be used by one person or a small group of people, there is nothing that prevents it from being used by a larger group.''"
  
Unlike [[Squid]], Polipo is very light on resources and simple to configure. This makes it ideal for single user systems and other uncomplicated setups. Do keep in mind; however, that this versatility comes at a cost; Polipo will increase its space usage without restriction as it is not aware of how big its disk cache grows. This perceived fault is by design, since omitting these sanity checks drastically reduces Polipo's memory usage and overall toll on the system. A practical way of restricting disk usage is by making Polipo run as its own user and employing [[disk quota]].
+
Unlike [[Squid]], Polipo is very light on resources and simple to configure. This makes it ideal for single user systems and other uncomplicated setups. Do keep in mind, however, that this versatility comes at a cost: Polipo will increase its space usage without restriction as it is not aware of how big its disk cache grows. This perceived fault is by design, since omitting these sanity checks drastically reduces Polipo's memory usage and overall toll on the system. A practical way of restricting disk usage is by making Polipo run as its own user and employing [[disk quota]].
  
 
The following covers installing and setting up Polipo.
 
The following covers installing and setting up Polipo.
Line 17: Line 17:
 
== Installation ==
 
== Installation ==
  
Install {{Pkg|polipo}}, available in the [[Official Repositories]].
+
[[Install]] the {{Pkg|polipo}} package.
 
+
Alternatively, install the newer development branch {{aur|polipo-git}} from the [[AUR]] instead.
+
 
+
=== Run polipo as different user (not nobody) ===
+
 
+
Polipo should run as an unprivileged user. Such a user can either be created or reused. Lets create the polipo cache folder first
+
 
+
# mkdir /var/cache/polipo
+
 
+
Now we add the new polipo group
+
 
+
# groupadd -r polipo
+
 
+
and  we add the new polipo user afterwards to this group
+
 
+
# useradd -d /var/cache/polipo -g polipo -r -s /bin/false polipo
+
 
+
then we need a new polipo logfile which is assigned to  the upper user and group
+
 
+
# touch /var/log/polipo.log
+
# chown -R polipo:polipo /var/log/polipo.log /var/cache/polipo
+
 
+
To change user running polipo daemon, we have to create another {{ic|polipo.service}} file
+
 
+
# touch /etc/systemd/system/polipo.service
+
 
+
Copy the following code section to the new {{ic|/etc/systemd/system/polipo.service}} file
+
 
+
.include /usr/lib/systemd/system/polipo.service
+
[Service]
+
User=polipo
+
 
+
To make sure all files and folders are created before you start Polipo as a designated user. Restart Polipo and check its logfile.
+
 
+
# systemctl restart polipo
+
# tail /var/log/polipo.log
+
 
+
If everything went well you should read something like this
+
 
+
  Established listening socket on port 8123.
+
 
+
Also check, if polipo is running as polipo user:
+
 
+
# ps aux | grep polipo
+
 
+
Stop the polipo daemon
+
 
+
# systemctl stop polipo
+
  
 
== Starting the daemon ==
 
== Starting the daemon ==
Line 114: Line 66:
 
=== Browser ===
 
=== Browser ===
  
Set the browser so that it uses {{Ic|localhost:8123}} for proxying. Be sure to disable the browser's disk cache to avoid redundant IO operations and bad performance.
+
Set the browser so that it uses {{Ic|localhost:8123}} for proxying. Be sure to disable the browser's disk cache to avoid redundant IO operations and bad performance. If {{Ic|localhost:8123}} does not work, try matching {{Ic|/etc/hosts}} like {{Ic|''hostname''.localdomain:8123}}.
  
 
=== Tunneling ===
 
=== Tunneling ===
Line 149: Line 101:
  
 
=== Tor ===
 
=== Tor ===
 +
 +
{{Warning|The Tor Project advises against transparently routing traffic through Tor [https://www.torproject.org/docs/faq.html.en#TBBOtherBrowser] [https://www.torproject.org/docs/faq.html.en#TBBSocksPort], and strongly recommends using only the Tor Browser [https://www.torproject.org/download/download.html.en#warning]. Consider instead using an [https://trac.torproject.org/projects/tor/wiki/doc/TorifyHOWTO/IsolatingProxy Isolating Proxy].}}
  
 
[[Tor]] is an anonymizing proxy network.
 
[[Tor]] is an anonymizing proxy network.
Line 168: Line 122:
 
=== DNS Error ===
 
=== DNS Error ===
  
If the network is started in background there could be a error like this (in the polipo log file):
+
If the network is started in background there could be a error like this in the Polipo log:
  
 
  Couldn't send DNS query: Connection refused
 
  Couldn't send DNS query: Connection refused
Line 179: Line 133:
 
* Do not start the net-profiles in background mode (probably not wanted).
 
* Do not start the net-profiles in background mode (probably not wanted).
 
* Set {{ic|dnsNameServer}} manually on the wanted DNS server.
 
* Set {{ic|dnsNameServer}} manually on the wanted DNS server.
* Or add {{ic|sleep 10}} (or more, it depends) near the beginning of the Polipo daemon script {{ic|/etc/rc.d/polipo}} in the start section. This will Polipo start after the network has initialised.
+
* Or add {{ic|sleep 10}} (or more, it depends) near the beginning of the Polipo daemon script {{ic|/etc/rc.d/polipo}} in the start section. This will make Polipo start after the network has initialised.
  
 
See [https://bbs.archlinux.org/viewtopic.php?id=86452 this thread] for more information on this topic.
 
See [https://bbs.archlinux.org/viewtopic.php?id=86452 this thread] for more information on this topic.
Line 187: Line 141:
 
*[http://www.pps.jussieu.fr/~jch/software/polipo/faq.html Polipo FAQ]
 
*[http://www.pps.jussieu.fr/~jch/software/polipo/faq.html Polipo FAQ]
 
*[http://www.pps.jussieu.fr/~jch/software/polipo/manual/index.html The Polipo Manual]
 
*[http://www.pps.jussieu.fr/~jch/software/polipo/manual/index.html The Polipo Manual]
 
 
== Deprecated (rc.d) ==
 
 
=== Run polipo as different user (not nobody) ===
 
 
Polipo should run as an unpriviledged user. Such a user can either be created or reused:
 
 
# mkdir /var/cache/polipo
 
# groupadd -r polipo
 
# useradd -d /var/cache/polipo -g polipo -r -s /bin/false polipo
 
 
To make sure all files and folders are created before you start polipo as an a designated user. Start and stop the Polipo.
 
 
# rc.d start polipo
 
# rc.d stop polipo
 
 
While other daemons start as root and drop privileges as soon as possible, polipo runs as the user that invoked it. If polipo is invoked from {{ic|/etc/rc.d/polipo}}, change the invokation line from
 
 
<nowiki>[[</nowiki> ! -d /var/run/$DAEMON <nowiki>]]</nowiki> && install -d $DAEMON /var/run/$DAEMON /usr/bin/$DAEMON $ARGS >/dev/null 2>&1
 
 
to
 
 
<nowiki>[[</nowiki> ! -d /var/run/$DAEMON <nowiki>]]</nowiki> && install -d $DAEMON --group=polipo --owner=polipo /var/run/$DAEMON
 
su -c "/usr/bin/$DAEMON $ARGS" -s /bin/sh polipo >/dev/null 2>&1
 
 
It is then also necessary to change ownership and/or permissions of several files and directories written by polipo:
 
 
# chown -R polipo:polipo /var/run/polipo /var/cache/polipo /var/log/polipo
 
 
Although a better choice for the log file is to create a directory {{ic|/var/log/polipo}} owned by the designated user and set polipo's log file to {{ic|/var/log/polipo/polipo.log}} via the logFile variable in the config file.
 
 
=== Starting the daemon ===
 
 
To start the Polipo daemon:
 
 
# rc.d start polipo
 
 
Add it to {{ic|/etc/rc.conf}} to start it automatically at boot:
 
 
DAEMONS=(syslog-ng network netfs '''polipo''' crond)
 

Latest revision as of 13:48, 10 November 2016

Warning: Polipo is no longer maintained, It's recommended to use another solution (squid, privoxy...) instead.

Related articles

From Polipo's site:

"Polipo is a small and fast caching web proxy (a web cache, an HTTP proxy, a proxy server). While Polipo was designed to be used by one person or a small group of people, there is nothing that prevents it from being used by a larger group."

Unlike Squid, Polipo is very light on resources and simple to configure. This makes it ideal for single user systems and other uncomplicated setups. Do keep in mind, however, that this versatility comes at a cost: Polipo will increase its space usage without restriction as it is not aware of how big its disk cache grows. This perceived fault is by design, since omitting these sanity checks drastically reduces Polipo's memory usage and overall toll on the system. A practical way of restricting disk usage is by making Polipo run as its own user and employing disk quota.

The following covers installing and setting up Polipo.

Installation

Install the polipo package.

Starting the daemon

To start the polipo daemon:

# systemctl start polipo

To start it automatically at boot:

# systemctl enable polipo

Multiple instances

Polipo can also run without super user privileges. To do so, first copy /etc/polipo/config.sample to a suitable directory:

$ cp /etc/polipo/config.sample ~/.poliporc

Edit it so that it points at a writable location, instead of /var/cache/polipo:

# Uncomment this if you want to put the on-disk cache in a
# non-standard location:
diskCacheRoot = "~/.polipo-cache/"

Create the cache directory:

$ mkdir ~/.polipo-cache

Finally, launch Polipo with the new configuration:

$ polipo -c ~/.poliporc

Configuration

Management is mostly performed in /etc/polipo/config. Most users can opt for using the sample configuration file, which is sufficient for most situations and well documented.

# cd /etc/polipo; cp config.sample config

One element of configuration that warrants mentioning is polipo's default behavior of blocking outbound connections by port. There are two variables in polipo's config file that control allowed outbound ports. allowedPorts specifies ports for outbound HTTP connections. It defaults to 80-100 and 1024-65535. tunnelAllowedPorts specifies ports polipo will allow tunnel traffic to as well as HTTPS traffic. By default it is much more restricted: "It defaults to allowing ssh, HTTP, https, rsync, IMAP, imaps, POP, pops, Jabber, CVS and Git traffic."

If you see a "403 Forbidden Port" error message from polipo when attempting to browse to a host:port, you need to configure polipo to accept traffic to more ports for either HTTP or HTTPS. To set them wide open, add the following to /etc/polipo/config:

allowedPorts = 1-65535
tunnelAllowedPorts = 1-65535

Unlike other proxies, Polipo needs to be restarted after alterations.

Browser

Set the browser so that it uses localhost:8123 for proxying. Be sure to disable the browser's disk cache to avoid redundant IO operations and bad performance. If localhost:8123 does not work, try matching /etc/hosts like hostname.localdomain:8123.

Tunneling

Note: According to the Polipo FAQ on "intercepting proxy" this is not possible/supported!
Note: this requires to run Polipo as its own user.

Instead of manually configuring each browser or other utilities that might benefit from Polipo's caching, one can also use iptables to route traffic through polipo.

After installing iptables, add the appropiate rules to /etc/iptables/iptables.rules:

*nat
:PREROUTING ACCEPT [0:0]
:POSTROUTING ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
-A OUTPUT -p tcp --dport 80 -m owner ! --uid-owner polipo -j ACCEPT
-A OUTPUT -p tcp --dport 80 -j REDIRECT --to-ports 8123
COMMIT

This routes HTTP traffic through Polipo. Remove all proxy settings from browsers, if any, and restart iptables.

Privoxy

Privoxy is a proxy useful for intercepting advertisement and other undesirables.

According to Polipo's developer, in order to get the privacy enhancements of Privoxy and much (but not all) of the performance of Polipo, one should place Polipo upstream of Privoxy.

In other words:

  • point the browser at Privoxy: localhost:8118
  • and direct Privoxy traffic to Polipo: forward / localhost:8123 in the Privoxy configuration file.

Tor

Warning: The Tor Project advises against transparently routing traffic through Tor [1] [2], and strongly recommends using only the Tor Browser [3]. Consider instead using an Isolating Proxy.

Tor is an anonymizing proxy network.

To use Polipo with Tor, uncomment or include the following in /etc/polipo/config:

socksParentProxy = localhost:9050
socksProxyType = socks5

DansGuardian

DansGuardian is a web content filter. The only difference to using DansGuardian with Polipo (rather than squid or tinyproxy) is that in dansguardian.conf the proxyport needs to be set to polipo's 8123:

# the port DansGuardian connects to proxy on
proxyport = 8123

Troubleshooting

DNS Error

If the network is started in background there could be a error like this in the Polipo log:

Couldn't send DNS query: Connection refused
Falling back on gethostbyname.
Getaddrinfo failed: Temporary name server failure
Host ***.com lookup failed: Getaddrinfo failed: Temporary name server failure (131072).

This error occurs because in background mode the network hasn't initialised before Polipo wants to connect to the DNS server (especially using DHCP). Solving this error is possible on three ways:

  • Do not start the net-profiles in background mode (probably not wanted).
  • Set dnsNameServer manually on the wanted DNS server.
  • Or add sleep 10 (or more, it depends) near the beginning of the Polipo daemon script /etc/rc.d/polipo in the start section. This will make Polipo start after the network has initialised.

See this thread for more information on this topic.

See also