search

w3f / polkadot-lab

Testing framework for Polkadot networks
11 stars 10 forks source link

Capability to model latency between nodes #85

Open kirillt opened 2 years ago

kirillt commented 2 years ago

One of declared goals of this project is possibility to inject latency or network failures:

Define the network conditions, including potential injected failures: loss of connectivity, latency, etc.

However, it still seems to be not implemented.

For the start, it would be enough to handle latency entry in config (see config/main.sample.yaml) with numeric value in milliseconds. If this entry is present, then network latency between any 2 of spawned nodes must be equal to specified value. I suspect this might be easily achievable by tweaking Helm/Kubernetes/Docker configs.

The solution must also contain basic test case demonstrating the feature, or instruction how to verify the solution.

kirillt commented 2 years ago

Maybe it's possible to inject docker-netem into all Docker containers and configure it from top-level Polkadot-Lab config.

kirillt commented 2 years ago

This is how to display running nodes:

sudo docker exec -it e2e-test-control-plane /bin/bash
kubectl --kubeconfig=/etc/kubernetes/scheduler.conf get pods | grep polkadot-.-0
gitcoinbot commented 2 years ago

Issue Status: 1. Open 2. Started 3. Submitted 4. Done

This issue now has a funding of 0.007 BTC (310.62 USD @ $44374.21/BTC) attached to it.

gitcoinbot commented 2 years ago

Issue Status: 1. Open 2. Started 3. Submitted 4. Done

The funding of 0.007 BTC (303.79 USD @ $30379.0/BTC) attached to this issue has been cancelled by the bounty submitter