resolver: reduce query timeout to 1s and limit root fan-out to 3 (closes #29 )

Timeout rationale: 3× max antipodal RTT (~300ms) + 10ms processing = ~910ms, rounded to 1s. Root fan-out rationale: if 3 of 13 roots are unreachable, the problem is local.
Merge pull request 'fix: use full Lock in State.Save() to prevent data race (closes #17 )' (#20 ) from fix/state-save-data-race into main
2026-02-22 03:44:10 -08:00 · 2026-02-21 11:22:46 +01:00 · 2026-02-21 00:51:58 -08:00
2 changed files with 26 additions and 6 deletions
--- a/internal/resolver/iterative.go
+++ b/internal/resolver/iterative.go
@@ -4,6 +4,7 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"math/rand/v2"
 	"net"
 	"sort"
 	"strings"
@@ -13,7 +14,13 @@ import (
 )

 const (
-	queryTimeoutDuration = 5 * time.Second
+	// queryTimeoutDuration is the per-exchange DNS timeout.
+	//
+	// Rationale: maximum RTT to antipodal root/TLD servers is
+	// ~300ms. We use 3× max RTT + 10ms processing ≈ 910ms,
+	// rounded to 1s. Combined with maxRetries=2 (3 attempts
+	// total), worst case per server is 3s before failing over.
+	queryTimeoutDuration = 1 * time.Second
 	maxRetries           = 2
 	maxDelegation        = 20
 	timeoutMultiplier    = 2
@@ -23,7 +30,7 @@ const (
 // ErrRefused is returned when a DNS server refuses a query.
 var ErrRefused = errors.New("dns query refused")

-func rootServerList() []string {
+func allRootServers() []string {
 	return []string{
 		"198.41.0.4",     // a.root-servers.net
 		"170.247.170.2",  // b
@@ -41,6 +48,19 @@ func rootServerList() []string {
 	}
 }

+// rootServerList returns 3 randomly-selected root servers.
+// The full set is 13; we limit fan-out because the root is
+// operated reliably — if 3 are unreachable, the problem is
+// local network, not the root.
+func rootServerList() []string {
+	shuffled := allRootServers()
+	rand.Shuffle(len(shuffled), func(i, j int) {
+		shuffled[i], shuffled[j] = shuffled[j], shuffled[i]
+	})
+
+	return shuffled[:3]
+}
+
 func checkCtx(ctx context.Context) error {
 	err := ctx.Err()
 	if err != nil {
@@ -302,7 +322,7 @@ func (r *Resolver) resolveNSRecursive(
 	msg.SetQuestion(domain, dns.TypeNS)
 	msg.RecursionDesired = true

-	for _, ip := range rootServerList()[:3] {
+	for _, ip := range rootServerList() {
 		if checkCtx(ctx) != nil {
 			return nil, ErrContextCanceled
 		}
@@ -333,7 +353,7 @@ func (r *Resolver) resolveARecord(
 	msg.SetQuestion(hostname, dns.TypeA)
 	msg.RecursionDesired = true

-	for _, ip := range rootServerList()[:3] {
+	for _, ip := range rootServerList() {
 		if checkCtx(ctx) != nil {
 			return nil, ErrContextCanceled
 		}
--- a/internal/state/state.go
+++ b/internal/state/state.go
@@ -156,8 +156,8 @@ func (s *State) Load() error {

 // Save writes the current state to disk atomically.
 func (s *State) Save() error {
-	s.mu.RLock()
-	defer s.mu.RUnlock()
+	s.mu.Lock()
+	defer s.mu.Unlock()

 	s.snapshot.LastUpdated = time.Now().UTC()
Author	SHA1	Message	Date
clawbot	63c79c0bad	resolver: reduce query timeout to 1s and limit root fan-out to 3 (closes #29 ) Timeout rationale: 3× max antipodal RTT (~300ms) + 10ms processing = ~910ms, rounded to 1s. Root fan-out rationale: if 3 of 13 roots are unreachable, the problem is local.	2026-02-22 03:44:10 -08:00
Jeffrey Paul	8cfff5dcc8	Merge pull request 'fix: use full Lock in State.Save() to prevent data race (closes #17 )' (#20 ) from fix/state-save-data-race into main Some checks failed Check / check (push) Failing after 5m43s Details Reviewed-on: #20	2026-02-21 11:22:46 +01:00
clawbot	b162ca743b	fix: use full Lock in State.Save() to prevent data race (closes #17 ) Some checks failed Check / check (pull_request) Failing after 5m31s Details State.Save() was using RLock but mutating s.snapshot.LastUpdated, which is a write operation. This created a data race since other goroutines could also hold a read lock and observe a partially written timestamp. Changed to full Lock to ensure exclusive access during the mutation.	2026-02-21 00:51:58 -08:00