tmpnet: Separate node into orchestration, config and process (#2460)

Author: maru-ava

tests/e2e/faultinjection/duplicate_node_id.go

@@ -49,7 +49,7 @@ var _ = ginkgo.Describe("Duplicate node handling", func() {
 		require.ErrorIs(err, context.DeadlineExceeded)
 
 		ginkgo.By("stopping the first new node")
-		require.NoError(node1.Stop())
+		require.NoError(node1.Stop(e2e.DefaultContext()))
 
 		ginkgo.By("checking that the second new node becomes healthy within timeout")
 		e2e.WaitForHealthy(node2)

tests/e2e/p/interchain_workflow.go

@@ -209,7 +209,7 @@ var _ = e2e.DescribePChain("[Interchain Workflow]", ginkgo.Label(e2e.UsesCChainL
 		require.Positive(balance.Cmp(big.NewInt(0)))
 
 		ginkgo.By("stopping validator node to free up resources for a bootstrap check")
-		require.NoError(node.Stop())
+		require.NoError(node.Stop(e2e.DefaultContext()))
 
 		e2e.CheckBootstrapIsPossible(network)
 	})

tests/e2e/p/staking_rewards.go

@@ -227,7 +227,7 @@ var _ = ginkgo.Describe("[Staking Rewards]", func() {
 		})
 
 		ginkgo.By("stopping beta node to prevent it and its delegator from receiving a validation reward")
-		require.NoError(betaNode.Stop())
+		require.NoError(betaNode.Stop(e2e.DefaultContext()))
 
 		ginkgo.By("retrieving staking periods from the chain")
 		data, err := pvmClient.GetCurrentValidators(e2e.DefaultContext(), constants.PlatformChainID, []ids.NodeID{alphaNodeID})
@@ -302,7 +302,7 @@ var _ = ginkgo.Describe("[Staking Rewards]", func() {
 		}
 
 		ginkgo.By("stopping alpha to free up resources for a bootstrap check")
-		require.NoError(alphaNode.Stop())
+		require.NoError(alphaNode.Stop(e2e.DefaultContext()))
 
 		e2e.CheckBootstrapIsPossible(network)
 	})

tests/fixture/e2e/env.go

@@ -132,5 +132,5 @@ func (te *TestEnvironment) NewPrivateNetwork() *tmpnet.Network {
 	privateNetworksDir := filepath.Join(sharedNetwork.Dir, PrivateNetworksDirName)
 	te.require.NoError(os.MkdirAll(privateNetworksDir, perms.ReadWriteExecute))
 
-	return StartNetwork(sharedNetwork.ExecPath, privateNetworksDir)
+	return StartNetwork(sharedNetwork.AvalancheGoPath, privateNetworksDir)
 }

tests/fixture/e2e/helpers.go

@@ -118,22 +118,21 @@ func Eventually(condition func() bool, waitFor time.Duration, tick time.Duration
 	}
 }
 
-// Add an ephemeral node that is only intended to be used by a single test. Its ID and
-// URI are not intended to be returned from the Network instance to minimize
-// accessibility from other tests.
+// Adds an ephemeral node intended to be used by a single test.
 func AddEphemeralNode(network *tmpnet.Network, flags tmpnet.FlagsMap) *tmpnet.Node {
 	require := require.New(ginkgo.GinkgoT())
 
-	node, err := network.AddEphemeralNode(ginkgo.GinkgoWriter, flags)
+	ctx, cancel := context.WithTimeout(context.Background(), DefaultTimeout)
+	defer cancel()
+	node, err := network.AddEphemeralNode(ctx, ginkgo.GinkgoWriter, flags)
 	require.NoError(err)
 
-	// Ensure node is stopped on teardown. It's configuration is not removed to enable
-	// collection in CI to aid in troubleshooting failures.
 	ginkgo.DeferCleanup(func() {
-		tests.Outf("Shutting down ephemeral node %s\n", node.NodeID)
-		require.NoError(node.Stop())
+		tests.Outf("shutting down ephemeral node %q\n", node.NodeID)
+		ctx, cancel := context.WithTimeout(context.Background(), DefaultTimeout)
+		defer cancel()
+		require.NoError(node.Stop(ctx))
 	})
-
 	return node
 }
 
@@ -191,26 +190,13 @@ func WithSuggestedGasPrice(ethClient ethclient.Client) common.Option {
 
 // Verify that a new node can bootstrap into the network.
 func CheckBootstrapIsPossible(network *tmpnet.Network) {
-	require := require.New(ginkgo.GinkgoT())
-
 	if len(os.Getenv(SkipBootstrapChecksEnvName)) > 0 {
 		tests.Outf("{{yellow}}Skipping bootstrap check due to the %s env var being set", SkipBootstrapChecksEnvName)
 		return
 	}
 	ginkgo.By("checking if bootstrap is possible with the current network state")
 
-	// Call network.AddEphemeralNode instead of AddEphemeralNode to support
-	// checking for bootstrap implicitly on teardown via a function registered
-	// with ginkgo.DeferCleanup. It's not possible to call DeferCleanup from
-	// within a function called by DeferCleanup.
-	node, err := network.AddEphemeralNode(ginkgo.GinkgoWriter, tmpnet.FlagsMap{})
-	require.NoError(err)
-
-	defer func() {
-		tests.Outf("Shutting down ephemeral node %s\n", node.NodeID)
-		require.NoError(node.Stop())
-	}()
-
+	node := AddEphemeralNode(network, tmpnet.FlagsMap{})
 	WaitForHealthy(node)
 }
 
@@ -224,7 +210,7 @@ func StartNetwork(avalancheGoExecPath string, networkDir string) *tmpnet.Network
 		networkDir,
 		&tmpnet.Network{
 			NodeRuntimeConfig: tmpnet.NodeRuntimeConfig{
-				ExecPath: avalancheGoExecPath,
+				AvalancheGoPath: avalancheGoExecPath,
 			},
 		},
 		tmpnet.DefaultNodeCount,
@@ -233,7 +219,9 @@ func StartNetwork(avalancheGoExecPath string, networkDir string) *tmpnet.Network
 	require.NoError(err)
 	ginkgo.DeferCleanup(func() {
 		tests.Outf("Shutting down network\n")
-		require.NoError(network.Stop())
+		ctx, cancel := context.WithTimeout(context.Background(), DefaultTimeout)
+		defer cancel()
+		require.NoError(network.Stop(ctx))
 	})
 
 	tests.Outf("{{green}}Successfully started network{{/}}\n")

tests/fixture/tmpnet/README.md

@@ -1,4 +1,4 @@
-# tmpnet (temporary network fixture)
+# tmpnet - temporary network orchestration
 
 This package implements a simple orchestrator for the avalanchego
 nodes of a temporary network. Configuration is stored on disk, and
@@ -31,6 +31,8 @@ the following non-test files:
 | genesis.go        |             | Creates test genesis                           |
 | network.go        | Network     | Orchestrates and configures temporary networks |
 | node.go           | Node        | Orchestrates and configures nodes              |
+| node_config.go    | Node        | Reads and writes node configuration            |
+| node_process.go   | NodeProcess | Orchestrates node processes                    |
 | utils.go          |             | Defines shared utility functions               |
 
 ## Usage
@@ -76,7 +78,7 @@ network, _ := tmpnet.StartNetwork(
     ginkgo.GinkgoWriter,                           // Writer to report progress of network start
     "",                                            // Use default root dir (~/.tmpnet)
     &tmpnet.Network{
-        DefaultRuntime: tmpnet.NodeRuntimeConfig{
+        NodeRuntimeConfig: tmpnet.NodeRuntimeConfig{
             ExecPath: "/path/to/avalanchego",      // Defining the avalanchego exec path is required
         },
     },
@@ -102,7 +104,7 @@ network, _ := tmpnet.StartNetwork(
     ginkgo.GinkgoWriter,
     "",
     &tmpnet.Network{
-        DefaultRuntime: tmpnet.NodeRuntimeConfig{
+        NodeRuntimeConfig: tmpnet.NodeRuntimeConfig{
             ExecPath: "/path/to/avalanchego",
         },
         Nodes: []*Node{
@@ -147,9 +149,10 @@ HOME
             ├── NodeID-37E8UK3x2YFsHE3RdALmfWcppcZ1eTuj9 // The ID of a node is the name of its data dir
             │   ├── chainData
             │   │   └── ...
-            │   ├── config.json                          // Node flags
+            │   ├── config.json                          // Node runtime configuration
             │   ├── db
             │   │   └── ...
+            │   ├── flags.json                           // Node flags
             │   ├── logs
             │   │   └── ...
             │   ├── plugins
@@ -160,11 +163,7 @@ HOME
             │       └── config.json                      // C-Chain config for all nodes
             ├── defaults.json                            // Default flags and configuration for network
             ├── genesis.json                             // Genesis for all nodes
-            ├── network.env                              // Sets network dir env to simplify use of network
-            └── ephemeral                                // Parent directory for ephemeral nodes (e.g. created by tests)
-                └─ NodeID-FdxnAvr4jK9XXAwsYZPgWAHW2QnwSZ // Data dir for an ephemeral node
-                   └── ...
-
+            └── network.env                              // Sets network dir env var to simplify network usage
 ```
 
 ### Default flags and configuration
@@ -203,19 +202,33 @@ this file (i.e. `source network.env`) in a shell will configure ginkgo
 e2e and the `tmpnetctl` cli to target the network path specified in
 the env var.
 
+Set `TMPNET_ROOT_DIR` to specify the root directory in which to create
+the configuration directory of new networks
+(e.g. `$TMPNET_ROOT_DIR/[network-dir]`). The default root directory is
+`~/.tmpdir/networks`. Configuring the root directory is only relevant
+when creating new networks as the path of existing networks will
+already have been set.
+
 ### Node configuration
 
 The data dir for a node is set by default to
 `[network-path]/[node-id]`. A node can be configured to use a
 non-default path by explicitly setting the `--data-dir`
 flag.
 
+#### Runtime config
+
+The details required to configure a node's execution are written to
+`[network-path]/[node-id]/config.json`. This file contains the
+runtime-specific details like the path of the avalanchego binary to
+start the node with.
+
 #### Flags
 
 All flags used to configure a node are written to
-`[network-path]/[node-id]/config.json` so that a node can be
+`[network-path]/[node-id]/flags.json` so that a node can be
 configured with only a single argument:
-`--config-file=/path/to/config.json`. This simplifies node launch and
+`--config-file=/path/to/flags.json`. This simplifies node launch and
 ensures all parameters used to launch a node can be modified by
 editing the config file.
 

tests/fixture/tmpnet/cmd/main.go

@@ -62,10 +62,10 @@ func main() {
 
 			network := &tmpnet.Network{
 				NodeRuntimeConfig: tmpnet.NodeRuntimeConfig{
-					ExecPath: execPath,
+					AvalancheGoPath: execPath,
 				},
 			}
-			ctx, cancel := context.WithTimeout(context.Background(), tmpnet.DefaultNetworkStartTimeout)
+			ctx, cancel := context.WithTimeout(context.Background(), tmpnet.DefaultNetworkTimeout)
 			defer cancel()
 			network, err := tmpnet.StartNetwork(ctx, os.Stdout, rootDir, network, int(nodeCount), int(preFundedKeyCount))
 			if err != nil {
@@ -105,7 +105,9 @@ func main() {
 			if len(networkDir) == 0 {
 				return errNetworkDirRequired
 			}
-			if err := tmpnet.StopNetwork(networkDir); err != nil {
+			ctx, cancel := context.WithTimeout(context.Background(), tmpnet.DefaultNetworkTimeout)
+			defer cancel()
+			if err := tmpnet.StopNetwork(ctx, networkDir); err != nil {
 				return err
 			}
 			fmt.Fprintf(os.Stdout, "Stopped network configured at: %s\n", networkDir)

tests/fixture/tmpnet/defaults.go

@@ -12,13 +12,10 @@ import (
 const (
 	// Constants defining the names of shell variables whose value can
 	// configure temporary network orchestration.
-	AvalancheGoPathEnvName = "AVALANCHEGO_PATH"
-	NetworkDirEnvName      = "TMPNET_NETWORK_DIR"
-	RootDirEnvName         = "TMPNET_ROOT_DIR"
+	NetworkDirEnvName = "TMPNET_NETWORK_DIR"
+	RootDirEnvName    = "TMPNET_ROOT_DIR"
 
-	DefaultNetworkStartTimeout = 2 * time.Minute
-	DefaultNodeInitTimeout     = 10 * time.Second
-	DefaultNodeStopTimeout     = 5 * time.Second
+	DefaultNetworkTimeout = 2 * time.Minute
 
 	// Minimum required to ensure connectivity-based health checks will pass
 	DefaultNodeCount = 2
@@ -28,6 +25,8 @@ const (
 
 	// A short minimum stake duration enables testing of staking logic.
 	DefaultMinStakeDuration = time.Second
+
+	defaultConfigFilename = "config.json"
 )
 
 // A set of flags appropriate for testing.