mirror of
https://github.com/mulgadc/spinifex.git
synced 2026-10-19 19:01:42 +00:00
Josh Siviour
92fa467c90
* refactor(e2e/multinode): flatten phase 11 + parallelize bucket #1 + drop phase numbers Split monolithic phase11_PlacementAndNATGateway (~430 lines, 8 inline sub-steps) into 6 top-level Tests, each owning its own VPC graph via t.Cleanup and layered setup helpers (spreadSetupVPC, spreadSetupBastion, spreadSetupPrivateTrio, spreadSetupNATGW). The cleanup-ordering test flips *bool latches as each explicit teardown step succeeds so the LIFO unwind is a no-op on the success path. Add t.Parallel() to read-only / independent Tests: ClusterHealth, InstanceDistribution, GuestSSH, VPCNetworking. Deviates from the mulga-siv-127 bead's bucket-#1 listing by keeping CrossNodeGateway sequential (asserts a snapshot equality across gateways that the parallel VPC test would break by churning instance count mid-assert) and CrossNodeOps sequential (stops/starts trio[0], would race GuestSSH). Both deviations documented in tests_test.go + plan doc. Drop the bash phase numbering (phaseN_X funcs + "Phase N" log strings) from the rest of the suite — 10 funcs renamed to runX. Phase numbers were a port-time crutch; the Test names now describe each scenario and phase numbers ageing into noise as bash is retired in Stage H. * refactor(e2e/multinode): collapse spread Tests + control execution order Two changes wedged together since they touch the same files: 1. Collapse the 6 top-level TestMultinodeSpread* Tests into a single TestMultinodeSpread with 6 t.Run sub-tests sharing one VPC + bastion + private trio + NAT GW setup chain. The 6-test split paid setup cost 6× — bastion launch + private trio launch + NAT GW provision per Test pushed phase-11 wall-clock near 60min vs ~15min monolithic. Shared- setup layout keeps JUnit granularity (one row per sub-test) while restoring the monolithic budget. Setup helpers register t.Cleanup against the outer test so teardown runs once at the end (LIFO: NAT GW → trio → bastion → VPC). 2. Move the TestMultinodeSpread wrapper into tests_test.go so file-source order controls execution position. Previously the func lived in placement_nat_test.go which sorts before preflight_test.go and tests_test.go alphabetically — Spread ran FIRST in the sequential bucket, defeating Preflight's fail-fast on a sick cluster. Wrapper now runs after NodeRecovery so the cluster has been fully degrade-tested before Spread's 4-VM + NAT GW + custom-VPC graph spins up.
53 lines
2.3 KiB
Go
53 lines
2.3 KiB
Go
//go:build e2e
|
|
|
|
package multinode
|
|
|
|
import (
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/mulgadc/spinifex/tests/e2e/harness"
|
|
"github.com/stretchr/testify/require"
|
|
)
|
|
|
|
// runNodeRecovery is the Go port of node-recovery validation
|
|
// (run-multinode-e2e.sh:961-1037). Starts spinifex.target on node2, waits for
|
|
// NATS to reform to 2 peers, gateway to answer, spx get nodes to show 3
|
|
// Ready, and node2's gateway to answer DescribeInstanceTypes.
|
|
//
|
|
// Idempotent: if runNodeFailure didn't actually take node2 down (e.g.
|
|
// `go test -run TestMultinodeNodeRecovery` in isolation), the StartNode call
|
|
// no-ops and the assertions still hold.
|
|
func runNodeRecovery(t *testing.T, fix *Fixture) {
|
|
harness.Phase(t, "Multinode — Node Recovery")
|
|
|
|
require.GreaterOrEqualf(t, len(fix.Cluster.Nodes), 3, "node recovery requires a 3-node cluster, have %d", len(fix.Cluster.Nodes))
|
|
node2 := fix.Cluster.Nodes[1]
|
|
|
|
harness.Step(t, "start spinifex.target on %s (%s)", node2.Name, node2.Addr)
|
|
harness.StartNode(t, node2)
|
|
|
|
harness.Step(t, "wait NATS to reform to 2 peers")
|
|
fix.Cluster.WaitNATSPeers(t, 2, harness.WithTimeout(2*time.Minute), harness.WithPoll(2*time.Second))
|
|
|
|
harness.Step(t, "wait %s gateway to answer HTTPS", node2.Name)
|
|
harness.WaitNodeServiceReady(t, node2, harness.WithTimeout(60*time.Second), harness.WithPoll(2*time.Second))
|
|
|
|
// Bash (run-multinode-e2e.sh:1018) appends `|| echo ""` to swallow a
|
|
// non-zero spx exit and never checks Ready count strictly — recovery is
|
|
// gated on the gateway DescribeInstanceTypes call below, which actually
|
|
// proves end-to-end NATS routing through node2. Downgrade spx to WARN.
|
|
harness.Step(t, "spx get nodes shows %d Ready after recovery (best-effort)", len(fix.Cluster.Nodes))
|
|
out := harness.SpxGetNodesAcrossCluster(t)
|
|
harness.Detail(t, "spx_get_nodes", out)
|
|
if ready := readyNodeCount(out); ready < len(fix.Cluster.Nodes) {
|
|
t.Logf("WARN: spx get nodes shows %d Ready after recovery (want >= %d) — proceeding (bash treats this as non-fatal)\n%s",
|
|
ready, len(fix.Cluster.Nodes), out)
|
|
}
|
|
|
|
harness.Step(t, "DescribeInstanceTypes answers via %s after recovery", node2.Name)
|
|
cli := harness.AWSClientForGateway(t, fix.Env, node2)
|
|
_, err := cli.EC2.DescribeInstanceTypes(nil)
|
|
require.NoErrorf(t, err, "%s describe-instance-types after recovery", node2.Name)
|
|
}
|