Access all your corporate stuff and web stuff at the same time without fuss
You want this if you're being worn out by:
- Seeing the 'talk to the hand' page from the corporate web proxy/filter
- Tunnel/proxy setup in too many places and in too many ways
- Tunnels dropping silently
- Forgetting to manually bring up tunnels after logging onto vpn
- Re-entering ssh credentials over and over (key based auth isn't allowed everywhere)
no-YOU-talk-to-the-hand solves all these issues by providing a straight-forward combination of sshuttle for the heavy network lifting, supervisord to keep everything up and manageable, and yaml to keep it simple and organized
Works with Linux and MacOS but not MS Windows due to sshuttle though there is a workaround for windows described here
- Sets up your tunnels automatically when your VPN connects
- Takes down your tunnels automatically when your VPN disconnects
- Keeps your tunnels up when they should be up
- Organizes your tunnels with a single, simple, YAML configuration
- Enters passwords for you as needed (sshpass required for now)
- Supports multiple VPNs (roots). Have different vpns that require separate tunnels? Define them in one place and only tunnels dependent on the vpn that is up are established
- Supports any number of simultaneous tunnels (thanks to sshuttle )
- Supportes nested dependencies. For example: (qa_db, prod_db) -- depends --> (corp_private) -- depends --> (corp_vpn)
Config.yml replaces all your tunnel scripts/aliases, ssh setup inside db tools, application specific web proxy setup, etc:
HOST_PERSONAL_PROXY: &HOST_PERSONAL_PROXY 192.168.1.X
PROXY_USER: &PROXY_USER proxy.username
CORP_USER: &CORP_USER company.username
CORP_PASS: &CORP_PASS pA$$wuuuurd
# Define the corporate subnets. SUBNETS_CORP_ALL encompasses addresses that will already be sent
# through your default network interace to the compnay network. This var is defined for exclusion
# from other tunnels which will override your system defaults. SUBNETS_CORP_RESTRICTED is used
# to forward a subset of corporate traffic through a jump server in order to reach hosts that are
# not reachable directly on the VPN.
SUBNETS_CORP_ALL: &SUBNETS_CORP_ALL
- "10.0.0.0/8"
SUBNETS_CORP_RESTRICTED: &SUBNETS_CORP_RESTRICTED
- "10.0.1.0/24"
- "10.0.2.0/24"
# Define several special destinations on the corporate network. HOST_CORP_JUMP defines the host
# through which all protected subnets must be accessed. HOST_CORP_PRIVILEGED_APP and HOST_CORP_SEURE_DB
# define an application server and database where the database can only be reached from the
# application server. Reaching the database will require a nested tunnel
HOST_CORP_JUMP: &HOST_CORP_JUMP 10.0.0.1
HOST_CORP_PRIVILEGED_APP: &HOST_CORP_PRIVILEGED_APP 10.0.1.1
HOST_CORP_SECURE_DB: &HOST_CORP_SECURE_DB 10.0.2.1
# Global config options
log_level: DEBUG # Python log level. Default is DEBUG
monitor_poll_seconds: 5 # Monitor thread wakeup (may be exceeded by a long tunnel check). Default is 20
tunnels:
# Watch for connection to corporate VPN. This is the 'root', external tunnel
# In this configuraiton, if the corporate jump server is available, then the vpn is up
vpn:
check:
host: *HOST_CORP_JUMP
port: 22
# Bypass corporate network policies for web browsing, skype, streaming music, etc.
# You must have a proxy server available that is outside the corporate network. If
# you don't have one, this project is still useful for accessing restricted
# resources within the corporate network.
personal:
depends: vpn
proxy:
host: *HOST_PERSONAL_PROXY
user: *PROXY_USER
pass:
check:
# instead of an ip and port, a check target can be a url for an http check
url: https://twitter.com/
forwards:
# includes and excludes. items can be ips, subnets, or lists of ip/subnets.
include:
# By default, forward everything through the personal proxy
- 0/0
exclude:
# exclude home network and anything corporate
- 192.168.0.0/16
- *SUBNETS_CORP_ALL
# Forward traffic destined for restricted subnets through a corporate jump server.
corp_sec:
depends: vpn
proxy:
host: *HOST_CORP_JUMP
user: *CORP_USER
pass: *CORP_PASS
# verify by checking ssh access to the privileged app server
check:
# If the application server is reachable, this tunnel is up
host: *HOST_CORP_PRIVILEGED_APP
port: 22
forwards:
# Include anything destined for a secured corporate subnet
include:
- *SUBNETS_CORP_RESTRICTED
# Tunnel to access a secure db server from a privileged app server. This tunnel depends
# on corp_restricted being established. For traffic destined for the DB, this rule will
# fire first and the traffic will be forwarded through the APP server, however traffic
# destined for the APP server is forwarded through the JUMP server.
prod_db:
depends: corp_sec
proxy:
host: *HOST_CORP_PRIVILEGED_APP
user: *CORP_USER
pass: *CORP_PASS
check:
driver: mysql+pymysql
db: testdb
user: testuser
pass: testpass
host: 10.0.2.1
port: '3306'
forwards:
# includes and excludes. items can be ips, subnets, or lists of ip/subnets.
include:
- *HOST_CORP_SECURE_DB
$ pip install no_you_talk_to_the_hand
If pip install results in a error like 'TLSV1_ALERT_PROTOCOL_VERSION' you may first need to upgrade pip:
$ curl https://bootstrap.pypa.io/get-pip.py | python
If you configure a password for any remote server then sshpass is required.
sshuttle requires root/admin privilege to change forward rules. If your user is prompted for sudo password, then you may encounter and error like sudo no tty present and no askpass program specified. A quick solution is to set the no password flag in the sudoers file. The following works currently on Macs:
$ sudo visudo
.. then add 'NOPASSWD' to the admin group like this:
$ %admin ALL=(ALL) NOPASSWD: ALL
If you check a tunnel via a sqlalchemy connection (see prod_db tunnel in sample config above) then sqlalchemy and the appropriate driver must be installed separately
Start daemon to begin managing the configured tunnels (in ~/.nyttth/config.yml)
$ nyttth start
Stop daemon along with any tunnels that are running
$ nyttth stop
Help:
$ nyttth status --help
Usage: nyttth status [OPTIONS]
View status of all configured tunnels
Options:
-t, --tunnel [qadb|riskdb|itun|dbtun|etun|vpn|rfindb]
specify a specific tunnel
-s, --skip skip tunnel health checks
--help Show this message and exit.
Example with VPN down:
$ nyttth status
Process Depends Proc State Conn Check
----------------------------------------------------------
vpn N/A down
itun vpn STOPPED Not started skipped
dbtun itun STOPPED Not started skipped
etun vpn STOPPED Not started skipped
qadb vpn STOPPED Not started skipped
Example with VPN up:
$ nyttth status
Process Depends Proc State Conn Check
--------------------------------------------------------------------
vpn N/A up
itun vpn RUNNING pid 1595, uptime 0:09:28 up
dbtun itun RUNNING pid 1603, uptime 0:09:23 up
etun vpn RUNNING pid 1565, uptime 0:09:33 up
qadb vpn RUNNING pid 2692, uptime 0:00:04 up
Help:
$ nyttth tail --help
Usage: nyttth tail [OPTIONS]
Use system tail command to display logs. If a specific tunnel is not specified
then all logs will be tailed including the supervisord main log and the vpnmon
tunnel monitor process.
Options:
-t, --tunnel [qadb|itun|dbtun|etun|vpn]
specify a specific tunnel to tail. If not
specified all tunnels and the tunnel monitor
(monitor) will be tailed
-f, --wait wait for additional data
-n, --lines INTEGER number of lines to display
--help Show this message and exit.
Tail output for a single (example) tunnel:
$ nyttth tail -f -t itun
server: warning: closed channel 158 got cmd=TCP_STOP_SENDING len=0
server: warning: closed channel 159 got cmd=TCP_STOP_SENDING len=0
server: warning: closed channel 160 got cmd=TCP_STOP_SENDING len=0
server: warning: closed channel 148 got cmd=TCP_STOP_SENDING len=0
server: warning: closed channel 162 got cmd=TCP_STOP_SENDING len=0
server: warning: closed channel 164 got cmd=TCP_STOP_SENDING len=0
When VPN Connects:
$ nyttth tail -f | grep nyttth
2017-05-17 11:52:53,357 DEBUG nyttth: checking tunnels
2017-05-17 11:52:53,497 INFO nyttth: qadb is down. starting
2017-05-17 11:52:53,498 INFO nyttth: dbtun is down. starting
2017-05-17 11:52:53,907 INFO nyttth: etun is down. starting
2017-05-17 11:52:55,493 INFO nyttth: itun is down. starting
2017-05-17 11:53:06,527 DEBUG nyttth: checking tunnels
2017-05-17 11:53:06,814 INFO nyttth: rfindb is down. starting
2017-05-17 11:53:17,826 DEBUG nyttth: checking tunnels
2017-05-17 11:53:28,129 DEBUG nyttth: checking tunnels
When VPN Disconnects:
$ nyttth tail -f | grep nyttth
2017-05-17 11:51:44,701 DEBUG nyttth: checking tunnels
2017-05-17 11:51:55,000 DEBUG nyttth: checking tunnels
2017-05-17 11:52:05,265 DEBUG nyttth: checking tunnels
2017-05-17 11:52:07,269 DEBUG nyttth: vpn is down
2017-05-17 11:52:07,274 INFO nyttth: qadb depends on vpn which is down. stopping
2017-05-17 11:52:07,281 INFO nyttth: itun depends on vpn which is down. stopping
2017-05-17 11:52:07,286 INFO nyttth: rfindb depends on itun which is down. stopping
2017-05-17 11:52:07,292 INFO nyttth: dbtun depends on vpn which is down. stopping
2017-05-17 11:52:07,299 INFO nyttth: etun depends on vpn which is down. stopping
2017-05-17 11:52:17,306 DEBUG nyttth: checking tunnels
2017-05-17 11:52:19,310 DEBUG nyttth: vpn is down
2017-05-17 11:52:29,324 DEBUG nyttth: checking tunnels
2017-05-17 11:52:31,329 DEBUG nyttth: vpn is down
2017-05-17 11:52:41,340 DEBUG nyttth: checking tunnels
2017-05-17 11:52:43,345 DEBUG nyttth: vpn is down
Run supervisorctl console
$ nyttth ctl
This project uses sshuttle version 0.78.1. Subsequent versions define PF (Packet Filter) exclusions in a way that breaks when there are exclusions in multiple instances of sshuttle.
Python 3 is not supported because supervisord does not
This docs ignores whatever technical differences there are between tunnels and forwards and just uses the word 'tunnels'.
Remote ssh servers through which trafffic is forwarded, are referred to as proxies.
The term 'VPN' refers to a 'root' tunnel in the configuration that specifies no proxy setup or forwards. It exsits to check an external condition (reachable network endpoint)and does not really have to be a true VPN