dnswatcher/internal/resolver/livedns_harness_test.go

package resolver_test

import (
	"context"
	"sync"
	"testing"
	"time"

	"github.com/stretchr/testify/assert"

	"sneak.berlin/go/dnswatcher/internal/resolver"
)

// Tests for the live-DNS harness in livedns_test.go itself. These
// exercise pure logic and the retry/concurrency plumbing; they
// perform no DNS resolution of any kind, so they neither mock DNS
// nor depend on it.

// Names for the synthetic status maps below. Nothing is ever queried
// at them: they are map keys handed to the package's pure counting
// helpers, not a stand-in for a nameserver.
const (
	nsExample1 = "ns1.example."
	nsExample2 = "ns2.example."
	nsExample3 = "ns3.example."
	nsExample4 = "ns4.example."
)

func TestLiveQuorumIsStrictMajority(t *testing.T) {
	t.Parallel()

	cases := map[int]int{
		0:  1,
		1:  1,
		2:  2,
		3:  2,
		4:  3,
		5:  3,
		13: 7,
	}

	for total, want := range cases {
		assert.Equal(
			t, want, liveQuorum(total),
			"liveQuorum(%d)", total,
		)
	}
}

func TestStatusCountingIgnoresSilentNameservers(t *testing.T) {
	t.Parallel()

	results := map[string]*resolver.NameserverResponse{
		nsExample1: {
			Nameserver: nsExample1,
			Status:     resolver.StatusOK,
		},
		nsExample2: {
			Nameserver: nsExample2,
			Status:     resolver.StatusOK,
		},
		nsExample3: {
			Nameserver: nsExample3,
			Status:     resolver.StatusTimeout,
		},
		nsExample4: {
			Nameserver: nsExample4,
			Status:     resolver.StatusError,
		},
	}

	assert.Equal(
		t, 2, countStatus(results, resolver.StatusOK),
	)
	assert.Equal(
		t, 0, countStatus(results, resolver.StatusNXDomain),
	)

	// Two of four answered, which is short of the quorum of
	// three: this is the state that triggers a retry rather
	// than an assertion failure.
	assert.Equal(t, 2, answeredCount(results))
	assert.Less(t, answeredCount(results), liveQuorum(len(results)))

	assert.Equal(
		t,
		"ns1.example.=ok ns2.example.=ok "+
			"ns3.example.=timeout ns4.example.=error",
		describeStatuses(results),
	)
}

func TestRetryLiveRecoversFromTransientFailure(t *testing.T) {
	t.Parallel()

	const wantAttempts = 2

	attempts := 0

	retryLive(t, "transient", func(_ context.Context) error {
		attempts++

		if attempts < wantAttempts {
			return errLiveNoAnswer
		}

		return nil
	})

	assert.Equal(t, wantAttempts, attempts)
}

func TestRetryLiveGivesEachAttemptADeadline(t *testing.T) {
	t.Parallel()

	retryLive(t, "deadline", func(ctx context.Context) error {
		deadline, ok := ctx.Deadline()
		assert.True(t, ok, "attempt should carry a deadline")

		remaining := time.Until(deadline)

		assert.LessOrEqual(t, remaining, liveAttemptTimeout)

		// Lower bound too: without one this passes for a
		// deadline far shorter than intended, which would
		// silently turn every live attempt into an instant
		// timeout.
		assert.Greater(t, remaining, liveAttemptTimeout/2)

		return nil
	})
}

// TestUnsanctionedStatusesRejectsWrongAnswers is the regression test
// for the defect this allowlist exists to prevent: a minority of
// nameservers answering WRONGLY while quorum keeps the suite green.
// nodata is the case that motivated it — it is a wrong answer, not
// silence, and it was previously banned by neither test.
func TestUnsanctionedStatusesRejectsWrongAnswers(t *testing.T) {
	t.Parallel()

	// Four nameservers, three OK and one answering nodata: a
	// quorum of three is satisfied and no NXDOMAIN is present, so
	// the old blocklist assertions both passed on this input.
	results := map[string]*resolver.NameserverResponse{
		nsExample1: {
			Nameserver: nsExample1,
			Status:     resolver.StatusOK,
		},
		nsExample2: {
			Nameserver: nsExample2,
			Status:     resolver.StatusOK,
		},
		nsExample3: {
			Nameserver: nsExample3,
			Status:     resolver.StatusOK,
		},
		nsExample4: {
			Nameserver: nsExample4,
			Status:     resolver.StatusNoData,
		},
	}

	assert.GreaterOrEqual(
		t,
		countStatus(results, resolver.StatusOK),
		liveQuorum(len(results)),
	)
	assert.Zero(t, countStatus(results, resolver.StatusNXDomain))

	// nodata is an ANSWER, so it never triggers a retry: nothing
	// but the allowlist stands between it and a false green.
	assert.Equal(t, len(results), answeredCount(results))

	assert.Equal(
		t,
		[]string{nsExample4 + "=nodata"},
		unsanctionedStatuses(
			results,
			resolver.StatusOK,
			resolver.StatusTimeout,
			resolver.StatusError,
		),
		"nodata must be reported as an unsanctioned status",
	)
}

func TestUnsanctionedStatusesToleratesSilenceOnly(t *testing.T) {
	t.Parallel()

	results := map[string]*resolver.NameserverResponse{
		nsExample1: {
			Nameserver: nsExample1,
			Status:     resolver.StatusNXDomain,
		},
		nsExample2: {
			Nameserver: nsExample2,
			Status:     resolver.StatusTimeout,
		},
		nsExample3: {
			Nameserver: nsExample3,
			Status:     resolver.StatusError,
		},
	}

	allowed := []string{
		resolver.StatusNXDomain,
		resolver.StatusTimeout,
		resolver.StatusError,
	}

	assert.Empty(
		t,
		unsanctionedStatuses(results, allowed...),
		"timeout and error are non-answers and are tolerated",
	)

	// The same silent nameservers do not count towards a quorum.
	assert.Equal(t, 1, answeredCount(results))

	// An unknown status is treated as silence by answeredCount —
	// so it retries and fails loudly — and is unsanctioned by the
	// allowlist rather than quietly permitted.
	const laterStatus = "some-status-added-later"

	results[nsExample4] = &resolver.NameserverResponse{
		Nameserver: nsExample4,
		Status:     laterStatus,
	}

	assert.Equal(t, 1, answeredCount(results))
	assert.Equal(
		t,
		[]string{nsExample4 + "=" + laterStatus},
		unsanctionedStatuses(results, allowed...),
	)
}

func TestRunLiveBoundsConcurrency(t *testing.T) {
	t.Parallel()

	const workers = 24

	var (
		mu       sync.Mutex
		wg       sync.WaitGroup
		inFlight int
		maxSeen  int
	)

	wg.Add(workers)

	for range workers {
		go func() {
			defer wg.Done()

			_ = runLive(func(_ context.Context) error {
				mu.Lock()
				inFlight++

				if inFlight > maxSeen {
					maxSeen = inFlight
				}
				mu.Unlock()

				time.Sleep(time.Millisecond)

				mu.Lock()
				inFlight--
				mu.Unlock()

				return nil
			})
		}()
	}

	wg.Wait()

	assert.Positive(t, maxSeen)
	assert.LessOrEqual(
		t, maxSeen, liveConcurrency,
		"live queries must stay under the package-wide gate",
	)
}