e2e

Blobstream end to end integration test

This directory contains the Blobstream e2e integration tests. It serves as a way to fully test the Blobstream orchestrator and relayer in real network scenarios

Topology

as discussed under #398 The e2e network defined under blobstream_network.go has the following components:

• 4 Celestia-app nodes that can be validators
• 4 Orchestrator nodes that will each run aside of a celestia-app
• 1 Ethereum node. Probably Ganache as it is easier to set up
• 1 Relayer node that will listen to Celestia chain and relay attestations
• 1 Deployer node that can deploy a new Blobstream contract when needed.

For more information on the environment variables required to run these tests, please check the docker-compose.yml file and the shell scripts defined under celestia-app directory.

P2P network

In some test scenarios, we only care about running a single orchestrator node. The problem is that in order for the DHT to start putting values, it needs to have at least 1 connected peer. Thus, in our testing scenarios, we create an extra DHT that is used mainly used to create the P2PQuerier, but also to connect to that orchestrator DHT and allow it to run.

// create dht for querying
bootstrapper, err := helpers.ParseAddrInfos(network.Logger, BOOTSTRAPPERS)
HandleNetworkError(t, network, err, false)
_, _, dht := blobstreamtesting.NewTestDHT(ctx, bootstrapper)
defer dht.Close()

with bootstrapper being that orchestrator P2P ID and listening address.

How to run

Requirements

To run the e2e tests, a working installation of docker-compose is needed.

Makefile

A Makefile has been defined under this directory to run the tests, with a test target:

make test

Run a specific test

To run a single test, run the following:

BLOBSTREAM_INTEGRATION_TEST=true go test -mod=readonly -test.timeout 30m -v -run <test_name>

Run all the tests using go directly

BLOBSTREAM_INTEGRATION_TEST=true go test -mod=readonly -test.timeout 30m -v

Common issues

Currently, when the tests are run using the above ways, there are possible issues that might happen.

hanging docker containers after a sudden network stop

If the tests were stopped unexpectedly, for example, sending a SIGINT, ie, ctrl+c, the resources will not be released correctly (might be fixed in the future). This will result in seeing similar logs to the following :

ERROR: for core0  Cannot create container for service core0: Conflict. The container name "/core0" is already in use by container "4bdaf40e2cd26bf549738ea95f53ba49cb5407c3d892b50b5a75e72e08e
3e0a8". You have to remove (or rename) that container to be able to reuse that name.                                                                                                          
Host is already in use by another container                                                                                                                                                   
Creating 626fbf28-7c90-4842-be8e-3346f864b369_ganache_1 ... error                                                                                                                             
                                                                                                                                                                                              
ERROR: for 626fbf28-7c90-4842-be8e-3346f864b369_ganache_1  Cannot start service ganache: driver failed programming external connectivity on endpoint 626fbf28-7c90-4842-be8e-3346f864b369_gana
che_1 (23bf2faf8fbce45f4a112b59183739f294c0e2d4fb208fec89e4805f3d719381): Bind for 0.0.0.0:8545 failed: port is already allocated                                                             
                                                                                                                                                                                              
ERROR: for core0  Cannot create container for service core0: Conflict. The container name "/core0" is already in use by container "4bdaf40e2cd26bf549738ea95f53ba49cb5407c3d892b50b5a75e72e08e
3e0a8". You have to remove (or rename) that container to be able to reuse that name.                                                                                                          
                                                                                                                                                                                              
ERROR: for ganache  Cannot start service ganache: driver failed programming external connectivity on endpoint 626fbf28-7c90-4842-be8e-3346f864b369_ganache_1 (23bf2faf8fbce45f4a112b59183739f294c0e2d4fb208fec89e4805f3d719381): Bind for 0.0.0.0:8545 failed: port is already allocated
Encountered errors while bringing up the project.
Attaching to 626fbf28-7c90-4842-be8e-3346f864b369_ganache_1

To fix it, run the cleanup.sh script under scripts directory :

./scripts/cleanup.sh

NB: This will kill and remove hanging containers and networks related to the executed. But might also delete unrelated ones if they have the same name.