[LocalNet] Execution Height and Connection Errors

[gokutheengineer]

Objective

I am experiencing the following unexpected behaviors when running LocalNet:

Issue 1. When LocalNet is run on main branch with the current commands in the README, and running the fresh LocalNet after executing the first TriggerNextView command and running PrintNodeState the height is printed as 2, which is supposed to be 1.

Issue 2. Also, there is one other weird error I am receiving, but I am not sure if it is only on my setup. After doing few iterations of TriggerNextView, PrintNodeState, and then ResetToGenesis one or some of the nodes refuse connections.

Origin Document

Issue 1:

Issue 2:

Goals

• Identify the reason of the behavior and fix the problem (if exists)
• Enable smooth execution of LocalNet

Deliverable

• LocalNet runs as expected

Non-goals / Non-deliverables

• Change business logic

General issue deliverables

• Update the appropriate CHANGELOG(s)
• Update any relevant local/global README(s)
• Update relevant source code tree explanations
• Add or update any relevant or supporting mermaid diagrams

Testing Methodology

• LocalNet: verify a LocalNet is still functioning correctly by following the instructions at docs/development/README.md

---

Creator: @gokutheengineer
Co-Owners:

[Olshansk]

Issue 1. When LocalNet is run on main branch with the current commands in the README, and running the fresh LocalNet after executing the first TriggerNextView command and running PrintNodeState the height is printed as 2, which is supposed to be 1.

Check the state and you can also do make db_drop && make docker_wipe_nodes to clear the state and restart local net from scratch.

Inside of consensus/module.go we have this function:

// TODO: Populate the entire state from the persistence module: validator set, quorum cert, last block hash, etc...
func (m *consensusModule) loadPersistedState() error {
	persistenceContext, err := m.GetBus().GetPersistenceModule().NewReadContext(-1) // Unknown height
	if err != nil {
		return nil
	}
	defer persistenceContext.Close()

	latestHeight, err := persistenceContext.GetLatestBlockHeight()
	if err != nil || latestHeight == 0 {
		// TODO: Proper state sync not implemented yet
		return nil
	}

	m.height = uint64(latestHeight) + 1 // +1 because the height of the consensus module is where it is actively participating in consensus

	m.nodeLog(fmt.Sprintf("Starting consensus module at height %d", latestHeight))

	return nil
}

[Olshansk]

Issue 2. Also, there is one other weird error I am receiving, but I am not sure it is only in my setup. After doing few iterations of TriggerNextView, PrintNodeState, and then ResetToGenesis one or some of the node refuses connections.

Quoting @deblasis from discord:

• issue 2: it is known concurrency issue (my fault I guess) and there are 2 PRs that incorporate a fix (both pretty close to be merged: [Core] Consolidate all node configurations into a new configs package - (Issue #396) #403 and [SPIKE][Consensus] Remove validatorMap from the consensusModule struct - (Issue #331) #425). I suppose we should have created a hotfix PR for that. Lesson learnt. Sorry if this is causing you headaches

[gokutheengineer]

make db_drop && make docker_wipe_nodes

This command doesn't solve this issue, because I execute make docker_wipe before restarting the local net every time, which deletes the containers, images and volumes. Such that, as expected, running make db_drop returns

Error: No such container: pocket-db
make: *** [db_drop] Error 1

Therefore, I believe the persistence module should not be loading from an existing state. I will investigate further to see what might be going wrong. I also tried executing make db_drop && make docker_wipe_nodes first and make docker_wipe later but it didn't help, as expected.

Could you please confirm you are able to run Local Net on the latest main branch? If that is the case, then something is wrong with my local env.

[gokutheengineer]

@Olshansk I think this can be closed, since the following issue are dedicated to problems outlined in issue: #441 (for errors starting at height 2), and PRs #403 and #425 (connection issues) which are already merged. Wdyt?

[Olshansk]

Closed per @gokutheengineer's comment above 👌