Easily tunnel traffic to and from your LAN and ZeroTier Virtual Network using an SOCKS5 proxy (includes an optional TCP relay for use behind difficult NATs)
Currently you must build it and distribute it to your server manually. The build process will pull a submodule (which itself pulls submodules), build libzt there, and then link pylon against the resultant libzt.a
static library. Requires clang
, and cmake
to build.
make
We have docker images available if you prefer.
docker run --init -p 9443:443 -p 19993:9993/udp zerotier/pylon:latest \
reflect
See below for configuring zerotier-one to use your reflect
as a tcp-relay.
docker run --init -e ZT_PYLON_SECRET_KEY=$(cat identity.secret) -e ZT_PYLON_WHITELISTED_PORT=4545 \
--net=host --cap-add NET_ADMIN \
zerotier/pylon refract 6ab565387a111111 --listen-addr 0.0.0.0 --listen-port 1080
See [Usage] for more info on the commands.
Pylon can be run as one of two personalities that can work alone or together depending on your needs:
Name | What do | Is this a ZeroTier Node? |
---|---|---|
pylon refract |
This bridges traffic to and from your LAN | Yes |
pylon reflect |
This relays traffic over TCP/443 |
No |
In many cases a single refract
instance is enough to bridge devices onto your ZeroTier network. However, if you're behind some tricky NAT you might need to set up a reflect
instance on a machine with a static IP for your refract
instance to use.
Set environment variable to hold your ZeroTier secret identity:
export ZT_PYLON_SECRET_KEY=$(sudo cat identity.secret)
By default Pylon will chose a random port to send ZeroTier traffic over, but If you need it to only send traffic over a whitelisted port you can specify one like so:
export ZT_PYLON_WHITELISTED_PORT=4545
Run proxy service to listen for app traffic locally on 127.0.0.1:1080
:
./pylon refract b84ac5c40a2339c9 --listen-addr 0.0.0.0 --listen-port 1080
You can also listen on 0.0.0.0
.
If you have a tricky NAT situation and can allow TCP/443
, you can specify a relay on a machine with a static IP. To reduce latency, the tcp-relay should be as close as possible to the nodes it is serving. A datacenter in the same city or the LAN would be ideal.
./pylon reflect
Then tell your pylon instances to use that to proxy traffic:
./pylon refract b84ac5c40a2339c9 --listen-addr 0.0.0.0 --listen-port 1080 --relay-addr 0.0.0.0 --relay-port 443
Note: a reflect
instance is just a tcp-proxy and can be used by a regular ZeroTier client just the same. Expand the following instructions for more info:
Instructions for use with regular ZeroTier client
The default tcp relay is at 204.80.128.1/443
-an anycast address.
See Service docs for more info on local.conf
{ "settings": { "tcpFallbackRelay": "1.2.3.4/443", "forceTcpRelay": true } }
In this example, forceTcpRelay
is enabled. This is helpful for testing or if you know you'll need tcp relay. It takes a few minutes for zerotier-one to realize it needs to relay otherwise.
If you are the admin of the network that is blocking ZeroTier UDP, you can transparently redirect 204.80.128.1 to one of your IP addresses. Users won't need to edit their local client configuration.
Configuring this in your Enterprise Firewall is left as an exercise to the reader.
Here is an iptables example for illustrative purposes:
-A PREROUTING -p tcp -d 204.80.128.1 --dport 443 -j DNAT --to-destination 1.2.3.4
-A POSTROUTING -p tcp -d 1.2.3.4 --dport 443 -j SNAT --to-source 204.80.128.1
Set up remote resource:
mkdir serveme && cd serveme && echo "served data" > served.txt && \
python -m http.server -b 0.0.0.0 8000
Attempt a proxied HTTP GET:
curl --verbose --output output.txt 172.28.128.86:8000 --proxy socks5://127.0.0.1:1080
make debug|clean
You'll get:
pylon-debug
While a single Pylon instance will work for multiple networks and multiple applications simultaneously it will perform better if a new instance is started for each proxied network. The underlying libzt isn't multithreaded so it is recommended that you also split your proxied traffic across multiple instances if you notice performance bottlecks. Finally, Pylon only supports IPv4 TCP but IPv6 and UDP support can be added if there is sufficient interest.
Releasing to Docker Hub is done from our internal CI.
- create a tag on
main
:git tag v0.1.7
- push the tag:
git push --tags
This triggers the build and push to dockerhub. - create a Github Release through the Github ui. Select the tag you just pushed.