resolver, watcher: a domain's nameservers are only its own delegation (closes #222)

When a domain's parent zone's servers answer NXDOMAIN, LookupNS returns
ErrNXDomain. The watcher then saves the domain with no nameservers and
nxdomain set, shown on the dashboard and in /api/v1/status, asks for none
of its records and removes those saved, so its old nameservers go in one
NS Change. A domain with no delegation of its own gets an empty set and
its records are still asked at the zone it is in.
FindAuthoritativeNameservers moves to a parent name only on one of those
two answers; when the servers do not answer, it returns the error. After
an upgrade, a domain without its own delegation that was saved with its
parent zone's nameservers gets one NS Change; the README says so.

Model: opus-5-5

internal/resolver/errors.go

@@ -10,6 +10,10 @@ var (
 		"no authoritative nameservers found",
 	)
 
+	// ErrNXDomain is returned when the servers of the zone a domain
+	// is in answer NXDOMAIN: the domain does not exist.
+	ErrNXDomain = errors.New("domain does not exist")
+
 	// ErrNoNameserverAnswered is returned when every nameserver
 	// asked about a name timed out, failed or returned a referral,
 	// so whether the name has addresses is unknown.

internal/resolver/export_test.go

@@ -17,6 +17,43 @@ func NewWithFailingTCP(log *slog.Logger) *Resolver {
 	return r
 }
 
+// NewWithQueryTimeout returns a Resolver whose queries over UDP give up
+// after timeout, so a test that asks an address where nothing answers
+// does not wait out the usual timeout.
+func NewWithQueryTimeout(log *slog.Logger, timeout time.Duration) *Resolver {
+	r := NewFromLogger(log)
+	r.client = &udpClient{timeout: timeout}
+
+	return r
+}
+
+// FollowDelegation exports followDelegation for testing.
+func (r *Resolver) FollowDelegation(
+	ctx context.Context,
+	domain string,
+	servers []string,
+) ([]string, error) {
+	return r.followDelegation(ctx, domain, servers)
+}
+
+// FindAuthoritativeNameserversFrom exports findAuthoritativeNameservers
+// for testing.
+func (r *Resolver) FindAuthoritativeNameserversFrom(
+	ctx context.Context,
+	domain string,
+	servers []string,
+) ([]string, error) {
+	return r.findAuthoritativeNameservers(ctx, domain, servers)
+}
+
+// ResolveNSIterative exports resolveNSIterative for testing.
+func (r *Resolver) ResolveNSIterative(
+	ctx context.Context,
+	domain string,
+) ([]string, error) {
+	return r.resolveNSIterative(ctx, domain)
+}
+
 // ExtractRecordValue exports extractRecordValue for testing.
 func ExtractRecordValue(rr dns.RR) string {
 	return extractRecordValue(rr)

internal/resolver/iterative.go

@@ -204,6 +204,12 @@ func glueIPs(nsNames []string, glue map[string][]net.IP) []string {
 	return ips
 }
 
+// followDelegation follows referrals from servers, the root servers, to
+// domain and returns the NS set of domain's delegation. When the servers
+// of the zone domain is in answer that domain does not exist, the error
+// is ErrNXDomain. When they answer that it has no delegation of its own,
+// because it is not the zone's apex, the set is empty and there is no
+// error. Any other error means that no such answer came.
 func (r *Resolver) followDelegation(
 	ctx context.Context,
 	domain string,
@@ -234,10 +240,15 @@ func (r *Resolver) followDelegation(
 		// An authoritative reply comes from the servers of the zone
 		// domain is in; it is not a referral, even when its authority
 		// section lists that zone's NS records. Without NS records in
-		// the answer, domain is not the zone's apex and has no
-		// nameservers of its own.
+		// the answer, domain has no nameservers of its own: it does
+		// not exist, when the reply is NXDOMAIN, or else it is not the
+		// zone's apex.
+		if resp.Authoritative && resp.Rcode == dns.RcodeNameError {
+			return nil, ErrNXDomain
+		}
+
 		if resp.Authoritative {
-			return nil, ErrNoNameservers
+			return []string{}, nil
 		}
 
 		authNS := extractNSSet(resp.Ns)
@@ -486,7 +497,8 @@ func (r *Resolver) resolveNSIPs(
 
 // resolveNSIterative queries for NS records using iterative
 // resolution as a fallback when followDelegation finds no
-// authoritative answer in the delegation chain.
+// authoritative answer in the delegation chain. Its result means what
+// followDelegation's does.
 func (r *Resolver) resolveNSIterative(
 	ctx context.Context,
 	domain string,
@@ -516,6 +528,16 @@ func (r *Resolver) resolveNSIterative(
 			return nsNames, nil
 		}
 
+		// As in followDelegation: domain has no nameservers of its
+		// own.
+		if resp.Authoritative && resp.Rcode == dns.RcodeNameError {
+			return nil, ErrNXDomain
+		}
+
+		if resp.Authoritative {
+			return []string{}, nil
+		}
+
 		// Follow delegation.
 		authNS := extractNSSet(resp.Ns)
 		if len(authNS) == 0 {
@@ -601,12 +623,23 @@ func (r *Resolver) resolveARecord(
 // FindAuthoritativeNameservers traces the delegation chain from
 // root servers to discover all authoritative nameservers for the
 // given domain, as the delegation from its parent zone's servers lists
-// them. For a name that is not a zone apex it tries each
-// parent name in turn, so it returns the nameservers of the zone the
-// name is in.
+// them. When the servers asked answer that the name has no delegation
+// of its own, or does not exist, it tries each parent name in turn, so
+// it returns the nameservers of the zone the name is in. When they do
+// not answer, it returns the error and tries no parent name.
 func (r *Resolver) FindAuthoritativeNameservers(
 	ctx context.Context,
 	domain string,
+) ([]string, error) {
+	return r.findAuthoritativeNameservers(ctx, domain, rootServerList())
+}
+
+// findAuthoritativeNameservers is FindAuthoritativeNameservers with each
+// walk starting at servers, the root servers.
+func (r *Resolver) findAuthoritativeNameservers(
+	ctx context.Context,
+	domain string,
+	servers []string,
 ) ([]string, error) {
 	if checkCtx(ctx) != nil {
 		return nil, ErrContextCanceled
@@ -622,19 +655,16 @@ func (r *Resolver) FindAuthoritativeNameservers(
 
 		candidate := strings.Join(labels[i:], ".") + "."
 
-		nsNames, err := r.followDelegation(
-			ctx, candidate, rootServerList(),
-		)
-		if err == nil && len(nsNames) > 0 {
+		nsNames, err := r.followDelegation(ctx, candidate, servers)
+		if err != nil && !errors.Is(err, ErrNXDomain) {
+			return nil, err
+		}
+
+		if len(nsNames) > 0 {
 			sort.Strings(nsNames)
 
 			return nsNames, nil
 		}
-
-		// The root servers would refuse every parent name too.
-		if errors.Is(err, ErrIntercepted) {
-			return nil, err
-		}
 	}
 
 	return nil, ErrNoNameservers
@@ -852,9 +882,7 @@ func readReply(
 	// A reply with no answer that lists other nameservers, from a server
 	// that does not hold the name's zone, is a referral and says nothing
 	// about the name's records. A server named in the delegation that
-	// does not hold the zone may send one, as do a parent zone's servers
-	// when FindAuthoritativeNameservers found no delegation for the
-	// name's zone and moved on to a parent name.
+	// does not hold the zone may send one.
 	if !msg.Authoritative && len(msg.Answer) == 0 &&
 		len(extractNSSet(msg.Ns)) > 0 {
 		state.gotReferral = true
@@ -1022,12 +1050,22 @@ func (r *Resolver) queryEachNS(
 	return results, nil
 }
 
-// LookupNS returns the NS record set for a domain.
+// LookupNS returns the NS record set of a domain, as the delegation from
+// its parent zone's servers lists it, and never a parent name's. When
+// they answer that the domain does not exist, the error is ErrNXDomain.
+// When they answer that it has no delegation of its own, the set is
+// empty and there is no error.
 func (r *Resolver) LookupNS(
 	ctx context.Context,
 	domain string,
 ) ([]string, error) {
-	return r.FindAuthoritativeNameservers(ctx, domain)
+	if checkCtx(ctx) != nil {
+		return nil, ErrContextCanceled
+	}
+
+	return r.followDelegation(
+		ctx, dns.Fqdn(strings.ToLower(domain)), rootServerList(),
+	)
 }
 
 // LookupAllRecords performs iterative resolution to find all DNS

internal/resolver/livedns_test.go

@@ -187,8 +187,8 @@ func liveFindAuthoritative(
 	return out
 }
 
-// liveLookupNS is liveFindAuthoritative through the LookupNS entry
-// point, so that both entry points stay independently exercised.
+// liveLookupNS looks up the NS record set of domain, a domain that has
+// one, retrying until the delegation chain can be walked.
 func liveLookupNS(
 	t *testing.T,
 	r *resolver.Resolver,

internal/resolver/resolver_test.go

@@ -25,6 +25,13 @@ import (
 // Test helpers
 // ----------------------------------------------------------------
 
+// nonexistentDomain is a .com domain that does not exist.
+const nonexistentDomain = "dnswatcher-test-does-not-exist.com"
+
+// noAnswerAddress is 192.0.2.1, a documentation address: nothing
+// answers there.
+const noAnswerAddress = "192.0.2.1"
+
 func newTestResolver(t *testing.T) *resolver.Resolver {
 	t.Helper()
 
@@ -88,6 +95,47 @@ func TestFindAuthoritativeNameservers_Subdomain(
 	assert.Equal(t, fromZone, fromHost)
 }
 
+// TestFindAuthoritativeNameservers_DelegatedSubdomain looks up the
+// nameservers of www.cs.cmu.edu, a name in cs.cmu.edu, a zone that
+// cmu.edu delegates to other servers. The servers of cs.cmu.edu answer
+// that the name has no delegation of its own, so it gets their names,
+// not those of the cmu.edu servers. Every referral on the way gives the
+// nameservers' addresses, so the walk sends few queries.
+func TestFindAuthoritativeNameservers_DelegatedSubdomain(
+	t *testing.T,
+) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+	fromHost := liveFindAuthoritative(t, r, "www.cs.cmu.edu")
+	fromZone := liveLookupNS(t, r, "cs.cmu.edu")
+	fromParent := liveLookupNS(t, r, "cmu.edu")
+
+	assert.Equal(t, fromZone, fromHost)
+	assert.NotEqual(t, fromParent, fromHost)
+}
+
+// TestFindAuthoritativeNameservers_NoAnswer starts each walk for
+// www.google.com at 192.0.2.1, a documentation address where nothing
+// answers. A walk that got no answer does not say that the name has no
+// delegation of its own, so the lookup returns that walk's error, about
+// www.google.com, and tries no parent name: trying google.com and com
+// would end in ErrNoNameservers, or in the error of a walk for one of
+// them.
+func TestFindAuthoritativeNameservers_NoAnswer(t *testing.T) {
+	t.Parallel()
+
+	r := resolver.NewWithQueryTimeout(slog.Default(), 100*time.Millisecond)
+
+	nameservers, err := r.FindAuthoritativeNameserversFrom(
+		t.Context(), "www.google.com", []string{noAnswerAddress},
+	)
+	require.Error(t, err)
+	require.NotErrorIs(t, err, resolver.ErrNoNameservers)
+	assert.Contains(t, err.Error(), "query www.google.com. @"+noAnswerAddress)
+	assert.Empty(t, nameservers)
+}
+
 func TestFindAuthoritativeNameservers_ReturnsSorted(
 	t *testing.T,
 ) {
@@ -831,8 +879,7 @@ func TestLookupNS_MatchesFindAuthoritative(t *testing.T) {
 // nameservers of g.ntpns.org. The org servers delegate its parent zone,
 // ntpns.org, without the addresses of its nameservers, so the walk has
 // to look them up to ask them. If it did not, the walk for g.ntpns.org
-// would fail and LookupNS would return the nameservers of ntpns.org,
-// which a.ntpns.org is not one of.
+// would fail.
 func TestLookupNS_ParentZoneDelegatedWithoutAddresses(t *testing.T) {
 	t.Parallel()
 
@@ -842,6 +889,131 @@ func TestLookupNS_ParentZoneDelegatedWithoutAddresses(t *testing.T) {
 	assert.Contains(t, nameservers, "a.ntpns.org.")
 }
 
+// TestLookupNS_DomainThatDoesNotExist looks up the nameservers of a .com
+// domain that does not exist. The .com servers answer NXDOMAIN, so the
+// error is ErrNXDomain, and the domain does not get their names.
+func TestLookupNS_DomainThatDoesNotExist(t *testing.T) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+
+	var (
+		nameservers []string
+		err         error
+	)
+
+	livednstest.Retry(
+		t,
+		"LookupNS("+nonexistentDomain+")",
+		func(ctx context.Context) error {
+			nameservers, err = r.LookupNS(ctx, nonexistentDomain)
+			if errors.Is(err, resolver.ErrNXDomain) {
+				return nil
+			}
+
+			return err
+		},
+	)
+
+	require.ErrorIs(t, err, resolver.ErrNXDomain)
+	assert.Empty(t, nameservers)
+}
+
+// TestLookupNS_NoDelegationOfItsOwn looks up the nameservers of
+// www.google.com, a name in the google.com zone with no delegation of
+// its own, as a domain such as octocat.github.io is. The google.com
+// servers answer with no NS records for it: the set is empty, and it is
+// not ErrNXDomain.
+func TestLookupNS_NoDelegationOfItsOwn(t *testing.T) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+
+	var nameservers []string
+
+	livednstest.Retry(
+		t,
+		"LookupNS(www.google.com)",
+		func(ctx context.Context) error {
+			var err error
+
+			nameservers, err = r.LookupNS(ctx, "www.google.com")
+
+			return err
+		},
+	)
+
+	assert.Empty(t, nameservers)
+}
+
+// TestFollowDelegation_NoAnswer starts the walk LookupNS uses, for
+// google.com, at 192.0.2.1, a documentation address where nothing
+// answers. A walk that got no answer is an error, not an empty set,
+// which the watcher would report as an NS Change with every nameserver
+// removed.
+func TestFollowDelegation_NoAnswer(t *testing.T) {
+	t.Parallel()
+
+	r := resolver.NewWithQueryTimeout(slog.Default(), 100*time.Millisecond)
+
+	nameservers, err := r.FollowDelegation(
+		t.Context(), "google.com.", []string{noAnswerAddress},
+	)
+	require.Error(t, err)
+	assert.Empty(t, nameservers)
+}
+
+// TestResolveNSIterative_NoDelegationOfItsOwn walks to the nameservers
+// of www.google.com as the fallback walk does. As in
+// TestLookupNS_NoDelegationOfItsOwn, the set is empty, with no error.
+func TestResolveNSIterative_NoDelegationOfItsOwn(t *testing.T) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+
+	var nameservers []string
+
+	livednstest.Retry(
+		t,
+		"ResolveNSIterative(www.google.com)",
+		func(ctx context.Context) error {
+			var err error
+
+			nameservers, err = r.ResolveNSIterative(ctx, "www.google.com")
+
+			return err
+		},
+	)
+
+	assert.Empty(t, nameservers)
+}
+
+// TestResolveNSIterative_DomainThatDoesNotExist walks to the nameservers
+// of a .com domain that does not exist as the fallback walk does. As in
+// TestLookupNS_DomainThatDoesNotExist, the error is ErrNXDomain.
+func TestResolveNSIterative_DomainThatDoesNotExist(t *testing.T) {
+	t.Parallel()
+
+	r := newTestResolver(t)
+
+	var err error
+
+	livednstest.Retry(
+		t,
+		"ResolveNSIterative("+nonexistentDomain+")",
+		func(ctx context.Context) error {
+			_, err = r.ResolveNSIterative(ctx, nonexistentDomain)
+			if errors.Is(err, resolver.ErrNXDomain) {
+				return nil
+			}
+
+			return err
+		},
+	)
+
+	require.ErrorIs(t, err, resolver.ErrNXDomain)
+}
+
 // ----------------------------------------------------------------
 // ResolveIPAddresses tests
 // ----------------------------------------------------------------