Very simple SOCKS proxy container. Allows you to proxy directly to a docker network or a Kubernetes cluster connecting a curl command, your Chrome or Firefox browser or even other applications like DBeaver to access a database.

Main features:

• IPv6 support
• JSON logs output support
• SOCKS 4, 5 and 5 with DNS resolution
• Filter the network addresses allowed (e.g. 192.168.1.0/16,fc00::/7, etc...)
• Filter the network domains allowed (e.g. default.svc,default.svc.cluster.local, etc...)
• Detailed connection log support
• Scale capacity by increasing deploy replicas
• Optional extra HTTP proxy for clients that don't support SOCKS proxy

Security scans to latest and build container images:

docker run -d --rm -p 1080:1080 --network mynet --name socksd nmaguiar/socksd

And then configure your SOCKS proxy to localhost:1080

kubectl create deployment socksd --image=nmaguiar/socksd --port=1080
kubectl port-forward deploy/socksd 1080:1080
# OR kubectl create service nodeport socksd --tcp=31080:1080 --node-port=31080

And then configure your SOCKS proxy to localhost:1080

You can avoid the port-forward by deploying a K8S NodePort service (e.g port 31080) to the socksd deployment. Please be aware that Istio will impact the connectinos to the socksd pod so it should be deployed in a non-Istio namespace.

Build command:

docker build -t socksd .

You can control the behaviour with environment variables:

You can add these variables with the option '-e' on the docker run -d --rm -p 1080:1080 -e LOGS=true -e ONLY_LOCAL=true -e OJOB_JSONLOG=true --network mynet --name socksd nmaguiar/socksd command or with kubectl set env deploy socksd LOGS=true ONLY_LOCAL=true OJOB_JSONLOG=true in Kubernetes

Creating a docker network, launching the socksd container and a nginx container and then "curl"ing directly the nginx container as if doing from another container:

docker network create test
docker run -d --rm -p 1080:1080 --network test --name socksd nmaguiar/socksd
docker run -d --rm -p 8888:80 --network test --name nginx nginx

# Curling inside the docker network 'test'
curl http://nginx --proxy socks5h://127.0.0.1:1080

Launching a socksd deployment, a nginx deployment and exposing the nginx deployment. Then "curl"ing directly the nginx service as if it was running inside the Kubernetes cluster:

kubectl create deployment socksd --image nmaguiar/socksd --port=1080
kubectl create deployment nginx --image nginx

kubectl expose deploy nginx --port=80
kubectl port-forward deploy/socksd 1080:1080 &
# OR JUST:  kubectl create service nodeport socksd --tcp=31080:1080 --node-port=31080

# Curling inside the Kubernetes cluster
curl http://nginx.default.svc --proxy socks5h://127.0.0.1:1080

Using the Kubernetes NodePort solution is faster than port-forward but requires that you have access to port 31080 on each node.

• AWS EKS example with AWS Route 53

Chrome

You can run a Chrome browser in your desktop and enter URLs as if the browser was running inside the Kubernetes cluster:

To launch a separate clean Chrome browser configured to use a SOCKS proxy with proxy DNS you can use the following scripts on each OS:

Use port 31080 instead of 1080 if you use the faster Kubernetes NodePort alternative

Firefox

Alternatively to Chrome based browsers you can configure Firefox by changing the connection settings for SOCKS proxy on a Firefox browser while also proxy DNS:

Use port 31080 instead of 1080 if you use the faster Kubernetes NodePort alternative

Launching a socksd deployment, a postgresql database deployment and then using DBeaver to access it:

kubectl create deployment socksd --image nmaguiar/socksd --port=1080
kubectl port-forward deploy/socksd 1080:1080 &
# OR kubectl create service nodeport socksd --tcp=31080:1080 --node-port=31080

helm install postgresql bitnami/postgresql
echo PASSWORD=$(kubectl get secret --namespace default postgresql -o jsonpath="{.data.postgres-password}" | base64 -d)

Then, using DBeaver, create a PostgreSQL connection like this:

setting the proxy SOCKS like this:

Use port 31080 instead of 1080 if you use the faster Kubernetes NodePort alternative

and you will be able to use the database directly as if you were running DBeaver inside the Kubernetes cluster:

To make it work with Oracle drivers follow these steps:

1. Choose an Oracle connection and then edit the "Driver settings" to update the Oracle Driver to the latest (>= 23.2.0.0)
2. On the connection, choose "Driver properties" and set:

Test the connection and connect to the Oracle database.

This was also tested with an Oracle's connection string