sneak/routewatch
1
0
Fork 0
Code Issues 2 Pull Requests 1 Actions 47 Packages Projects Releases Wiki Activity

Re-vendor the shared files from sneak/prompts at 0653579 (closes #52)
check / check (push) Canceled after 0s
Details

Browse Source 
Takes .golangci.yml, .dockerignore, .gitignore, .editorconfig, the
workflow and REPO_POLICIES.md from sneak/prompts 0653579, with this
repository's own entries after the shared content. Lint (golangci-lint
v2.14.0) and test are Dockerfile phases that script/lint and script/test
build alone with a new CACHEBUST value, keeping Go's and golangci-lint's
caches; the image build depends on both. make fmt and make fmt-check also
format the markdown with prettier, pinned in yarn.lock. The agent
guidance moves to AGENTS.md. The Go code is fixed for what the new
configuration finds, and the AS and prefix detail reads now return the
rows' error after their loops.

Model: opus-5-5
This commit was merged in pull request #85.
This commit is contained in:
clawbot
2026-10-19 12:10:11 +02:00
parent 4f515d718b
commit b89987e356
65 changed files with 5624 additions and 3509 deletions
Download Patch File Download Diff File Expand all files Collapse all files
internal/config/config.go
+27 -12
View File
@@ -2,6 +2,7 @@
package config
import (
"errors"
"fmt"
"os"
"path/filepath"
@@ -27,6 +28,12 @@ const (
maxPort = 65535
)
var (
errRelativeXDGDataHome = errors.New("XDG_DATA_HOME must be an absolute path")
errUnsupportedOS = errors.New("unsupported operating system")
errInvalidPort = errors.New("PORT must be a whole number")
)
// Config holds configuration for the entire application
type Config struct {
// StateDir is the directory for all application state (database, snapshots)
@@ -38,11 +45,14 @@ type Config struct {
// MaxRuntime is the maximum runtime (0 = run forever)
MaxRuntime time.Duration
// EnableBatchedDatabaseWrites enables batched database operations for better performance
// EnableBatchedDatabaseWrites enables batched database operations for
// better performance
EnableBatchedDatabaseWrites bool
// RouteExpirationTimeout is how long a route can go without being refreshed before expiring
// Default is 2 hours which is conservative for BGP (typical BGP hold time is 90-180 seconds)
// RouteExpirationTimeout is how long a route can go without being
// refreshed before expiring
// Default is 2 hours which is conservative for BGP (typical BGP hold
// time is 90-180 seconds)
RouteExpirationTimeout time.Duration
}
@@ -59,11 +69,14 @@ func New() (*Config, error) {
}
return &Config{
StateDir: stateDir,
Port: port,
MaxRuntime: 0, // Run forever by default
EnableBatchedDatabaseWrites: true, // Enable batching by default
RouteExpirationTimeout: defaultRouteExpirationMinutes * time.Minute, // For active route monitoring
StateDir: stateDir,
Port: port,
// Run forever by default
MaxRuntime: 0,
// Enable batching by default
EnableBatchedDatabaseWrites: true,
// For active route monitoring
RouteExpirationTimeout: defaultRouteExpirationMinutes * time.Minute,
}, nil
}
@@ -89,7 +102,7 @@ func getStateDirectory() (string, error) {
// the database somewhere unexpected.
xdgData := os.Getenv("XDG_DATA_HOME")
if xdgData != "" && !filepath.IsAbs(xdgData) {
return "", fmt.Errorf("XDG_DATA_HOME must be an absolute path, got %q", xdgData)
return "", fmt.Errorf("%w, got %q", errRelativeXDGDataHome, xdgData)
}
// Unix-like: /var/lib/berlin.sneak.app.routewatch if root, else XDG_DATA_HOME
@@ -111,7 +124,7 @@ func getStateDirectory() (string, error) {
return filepath.Join(home, ".local", "share", AppIdentifier), nil
default:
return "", fmt.Errorf("unsupported operating system: %s", runtime.GOOS)
return "", fmt.Errorf("%w: %s", errUnsupportedOS, runtime.GOOS)
}
}
@@ -125,7 +138,8 @@ func getPort() (int, error) {
// ParseUint, unlike Atoi, refuses a sign: the health check URL cannot use "+9090"
port, err := strconv.ParseUint(value, 10, 0)
if err != nil || port < 1 || port > maxPort {
return 0, fmt.Errorf("PORT must be a whole number from 1 to %d, got %q", maxPort, value)
return 0, fmt.Errorf("%w from 1 to %d, got %q",
errInvalidPort, maxPort, value)
}
return int(port), nil
@@ -134,7 +148,8 @@ func getPort() (int, error) {
// EnsureDirectories creates all necessary directories if they don't exist
func (c *Config) EnsureDirectories() error {
// Ensure state directory exists
if err := os.MkdirAll(c.StateDir, dirPermissions); err != nil {
err := os.MkdirAll(c.StateDir, dirPermissions)
if err != nil {
return fmt.Errorf("failed to create state directory: %w", err)
}
internal/config/config_test.go → internal/config/config_internal_test.go
+8 -3
View File
@@ -36,7 +36,8 @@ func TestNewRefusesInvalidPort(t *testing.T) {
t.Setenv("PORT", value)
t.Setenv("XDG_DATA_HOME", "")
if _, err := New(); err == nil {
_, err := New()
if err == nil {
t.Errorf("New() with PORT=%q returned no error", value)
}
})
@@ -51,12 +52,16 @@ func TestNewRefusesRelativeXDGDataHome(t *testing.T) {
t.Setenv("PORT", "")
t.Setenv("XDG_DATA_HOME", "relative/path")
if _, err := New(); err == nil {
_, err := New()
if err == nil {
t.Error("New() with a relative XDG_DATA_HOME returned no error")
}
t.Setenv("XDG_DATA_HOME", "/var/lib")
if _, err := New(); err != nil {
_, err = New()
if err != nil {
t.Errorf("New() with XDG_DATA_HOME=/var/lib: %v", err)
}
}
internal/database/counts.go
+21 -6
View File
@@ -33,8 +33,10 @@ type liveCounts struct {
// seed sets every count to the value read from the tables at startup. It runs
// before any writer, so it needs no coordination with the adjust methods.
func (c *liveCounts) seed(asns, prefixesV4, prefixesV6, peerings, peers, routesV4, routesV6 int,
distributionV4, distributionV6 []PrefixDistribution) {
func (c *liveCounts) seed(
asns, prefixesV4, prefixesV6, peerings, peers, routesV4, routesV6 int,
distributionV4, distributionV6 []PrefixDistribution,
) {
c.mu.Lock()
defer c.mu.Unlock()
@@ -45,9 +47,11 @@ func (c *liveCounts) seed(asns, prefixesV4, prefixesV6, peerings, peers, routesV
c.peers = peers
c.routesV4 = routesV4
c.routesV6 = routesV6
for _, entry := range distributionV4 {
addAtMaskLength(c.distributionV4[:], entry.MaskLength, entry.Count)
}
for _, entry := range distributionV6 {
addAtMaskLength(c.distributionV6[:], entry.MaskLength, entry.Count)
}
@@ -102,6 +106,7 @@ func (c *liveCounts) addToDistribution(maskLengthsV4, maskLengthsV6 []int, n int
for _, maskLength := range maskLengthsV4 {
addAtMaskLength(c.distributionV4[:], maskLength, n)
}
for _, maskLength := range maskLengthsV6 {
addAtMaskLength(c.distributionV6[:], maskLength, n)
}
@@ -138,9 +143,11 @@ func (c *liveCounts) fill(s *Stats) {
// ascending order, the way the distribution query returns them.
func distributionList(counts []int) []PrefixDistribution {
var list []PrefixDistribution
for maskLength, count := range counts {
if count > 0 {
list = append(list, PrefixDistribution{MaskLength: maskLength, Count: count})
list = append(list,
PrefixDistribution{MaskLength: maskLength, Count: count})
}
}
@@ -152,7 +159,8 @@ func distributionList(counts []int) []PrefixDistribution {
func (d *Database) countRows(ctx context.Context, table string) (int, error) {
var n int
// table is one of a fixed set of literals below, never external input.
if err := d.db.QueryRowContext(ctx, "SELECT COUNT(*) FROM "+table).Scan(&n); err != nil {
err := d.db.QueryRowContext(ctx, "SELECT COUNT(*) FROM "+table).Scan(&n)
if err != nil {
return 0, fmt.Errorf("failed to count %s: %w", table, err)
}
@@ -167,30 +175,37 @@ func (d *Database) seedCounts(ctx context.Context) error {
if err != nil {
return err
}
prefixesV4, err := d.countRows(ctx, "prefixes_v4")
if err != nil {
return err
}
prefixesV6, err := d.countRows(ctx, "prefixes_v6")
if err != nil {
return err
}
peerings, err := d.countRows(ctx, "peerings")
if err != nil {
return err
}
peers, err := d.countRows(ctx, "bgp_peers")
if err != nil {
return err
}
routesV4, err := d.countRows(ctx, "live_routes_v4")
routesV4, err := d.countRows(ctx, tableLiveRoutesV4)
if err != nil {
return err
}
routesV6, err := d.countRows(ctx, "live_routes_v6")
routesV6, err := d.countRows(ctx, tableLiveRoutesV6)
if err != nil {
return err
}
distributionV4, distributionV6, err := d.GetPrefixDistributionContext(ctx)
if err != nil {
return err
internal/database/counts_test.go → internal/database/counts_internal_test.go
+222 -95
View File
@@ -12,6 +12,12 @@ import (
"github.com/google/uuid"
)
// testV4PeerIP is the peer the IPv4 routes in these tests are learned from.
const testV4PeerIP = "192.0.2.1"
// testV6Prefix is the prefix of the IPv6 routes in these tests.
const testV6Prefix = "2001:db8::/32"
// mkV4Route builds an IPv4 live route with its mask length taken from the
// prefix.
func mkV4Route(t *testing.T, prefix string, asn int, ts time.Time) *LiveRoute {
@@ -28,7 +34,7 @@ func mkV4Route(t *testing.T, prefix string, asn int, ts time.Time) *LiveRoute {
MaskLength: maskLength,
IPVersion: ipVersionV4,
OriginASN: asn,
PeerIP: "192.0.2.1",
PeerIP: testV4PeerIP,
ASPath: []int{asn},
NextHop: "192.0.2.254",
LastUpdated: ts,
@@ -56,6 +62,8 @@ func mkV6Route(prefix string, asn int, ts time.Time) *LiveRoute {
// deleted. These counts are what /api/v1/stats reports; before this change the
// endpoint recomputed them with a COUNT(*) over each table on every request.
func TestLiveCountsTrackWritesInRealtime(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -70,6 +78,7 @@ func TestLiveCountsTrackWritesInRealtime(t *testing.T) {
if err != nil {
t.Fatalf("GetStatsContext on empty database: %v", err)
}
if empty.ASNs != 0 || empty.Prefixes != 0 || empty.Peerings != 0 ||
empty.Peers != 0 || empty.LiveRoutes != 0 {
t.Fatalf("empty database counts nonzero: %+v", empty)
@@ -77,30 +86,16 @@ func TestLiveCountsTrackWritesInRealtime(t *testing.T) {
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
if err := db.GetOrCreateASNBatch(map[int]time.Time{64500: ts, 64501: ts}); err != nil {
t.Fatalf("GetOrCreateASNBatch: %v", err)
}
if err := db.UpdatePrefixesBatch(map[string]time.Time{
"198.51.100.0/24": ts,
"2001:db8::/32": ts,
}); err != nil {
t.Fatalf("UpdatePrefixesBatch: %v", err)
}
if err := db.UpdatePeerBatch(map[string]PeerUpdate{
"192.0.2.1": {PeerIP: "192.0.2.1", PeerASN: 64500, MessageType: "UPDATE", Timestamp: ts},
}); err != nil {
t.Fatalf("UpdatePeerBatch: %v", err)
}
if err := db.RecordPeering(64500, 64501, ts); err != nil {
t.Fatalf("RecordPeering: %v", err)
}
writeASNsPrefixesAndPeers(t, db, ts)
routes := []*LiveRoute{
mkV4Route(t, "198.51.100.0/24", 64500, ts),
mkV4Route(t, "203.0.113.0/24", 64501, ts.Add(time.Minute)),
mkV6Route("2001:db8::/32", 64502, ts.Add(2*time.Minute)),
mkV6Route(testV6Prefix, 64502, ts.Add(2*time.Minute)),
}
if err := db.UpsertLiveRouteBatch(routes); err != nil {
err = db.UpsertLiveRouteBatch(routes)
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
@@ -108,44 +103,87 @@ func TestLiveCountsTrackWritesInRealtime(t *testing.T) {
if err != nil {
t.Fatalf("GetStatsContext: %v", err)
}
assertCounts(t, "after inserts", stats, wantCounts{
asns: 2, prefixes: 2, peerings: 1, peers: 1,
ipv4Routes: 2, ipv6Routes: 1, liveRoutes: 3,
})
// Re-announcing the same routes is an update, not an insert: counts hold.
if err := db.UpsertLiveRouteBatch(routes); err != nil {
err = db.UpsertLiveRouteBatch(routes)
if err != nil {
t.Fatalf("UpsertLiveRouteBatch (re-announce): %v", err)
}
stats, err = db.GetStatsContext(ctx)
if err != nil {
t.Fatalf("GetStatsContext: %v", err)
}
assertCounts(t, "after re-announce", stats, wantCounts{
asns: 2, prefixes: 2, peerings: 1, peers: 1,
ipv4Routes: 2, ipv6Routes: 1, liveRoutes: 3,
})
// A withdrawal removes one route.
if err := db.DeleteLiveRouteBatch([]LiveRouteDeletion{
{Prefix: "203.0.113.0/24", PeerIP: "192.0.2.1", IPVersion: ipVersionV4},
}); err != nil {
err = db.DeleteLiveRouteBatch([]LiveRouteDeletion{
{Prefix: "203.0.113.0/24", PeerIP: testV4PeerIP, IPVersion: ipVersionV4},
})
if err != nil {
t.Fatalf("DeleteLiveRouteBatch: %v", err)
}
stats, err = db.GetStatsContext(ctx)
if err != nil {
t.Fatalf("GetStatsContext: %v", err)
}
assertCounts(t, "after delete", stats, wantCounts{
asns: 2, prefixes: 2, peerings: 1, peers: 1,
ipv4Routes: 1, ipv6Routes: 1, liveRoutes: 2,
})
}
// writeASNsPrefixesAndPeers writes two ASNs, two prefixes, one peer and one
// peering: the rows besides live routes that the stats count.
func writeASNsPrefixesAndPeers(t *testing.T, db *Database, ts time.Time) {
t.Helper()
err := db.GetOrCreateASNBatch(map[int]time.Time{64500: ts, 64501: ts})
if err != nil {
t.Fatalf("GetOrCreateASNBatch: %v", err)
}
err = db.UpdatePrefixesBatch(map[string]time.Time{
"198.51.100.0/24": ts,
testV6Prefix: ts,
})
if err != nil {
t.Fatalf("UpdatePrefixesBatch: %v", err)
}
err = db.UpdatePeerBatch(map[string]PeerUpdate{
testV4PeerIP: {
PeerIP: testV4PeerIP, PeerASN: 64500,
MessageType: "UPDATE", Timestamp: ts,
},
})
if err != nil {
t.Fatalf("UpdatePeerBatch: %v", err)
}
err = db.RecordPeering(64500, 64501, ts)
if err != nil {
t.Fatalf("RecordPeering: %v", err)
}
}
// TestLiveCountsSeededFromDatabaseAtStartup writes rows, reopens the same
// database file, and checks the counts and the prefix distribution come back
// from the seed scan rather than starting at zero.
func TestLiveCountsSeededFromDatabaseAtStartup(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -154,16 +192,22 @@ func TestLiveCountsSeededFromDatabaseAtStartup(t *testing.T) {
}
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
if err := db.GetOrCreateASNBatch(map[int]time.Time{64500: ts, 64501: ts, 64502: ts}); err != nil {
err = db.GetOrCreateASNBatch(map[int]time.Time{64500: ts, 64501: ts, 64502: ts})
if err != nil {
t.Fatalf("GetOrCreateASNBatch: %v", err)
}
if err := db.UpsertLiveRouteBatch([]*LiveRoute{
err = db.UpsertLiveRouteBatch([]*LiveRoute{
mkV4Route(t, "198.51.100.0/24", 64500, ts),
mkV6Route("2001:db8::/32", 64502, ts),
}); err != nil {
mkV6Route(testV6Prefix, 64502, ts),
})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
if err := db.Close(); err != nil {
err = db.Close()
if err != nil {
t.Fatalf("Close: %v", err)
}
@@ -177,13 +221,16 @@ func TestLiveCountsSeededFromDatabaseAtStartup(t *testing.T) {
if err != nil {
t.Fatalf("GetStatsContext after reopen: %v", err)
}
if stats.ASNs != 3 {
t.Errorf("seeded ASNs = %d, want 3", stats.ASNs)
}
if stats.IPv4Routes != 1 || stats.IPv6Routes != 1 || stats.LiveRoutes != 2 {
t.Errorf("seeded routes = (v4 %d, v6 %d, total %d), want (1, 1, 2)",
stats.IPv4Routes, stats.IPv6Routes, stats.LiveRoutes)
}
assertDistribution(t, "seeded IPv4 distribution", stats.IPv4PrefixDistribution,
[]PrefixDistribution{{MaskLength: 24, Count: 1}})
assertDistribution(t, "seeded IPv6 distribution", stats.IPv6PrefixDistribution,
@@ -196,26 +243,91 @@ func TestLiveCountsSeededFromDatabaseAtStartup(t *testing.T) {
// tables. The steps run once through the batch methods the prefix handler uses
// and once through the single-route methods.
func TestPrefixDistributionTracksWrites(t *testing.T) {
t.Parallel()
steps := prefixDistributionSteps(t)
for _, batch := range []bool{true, false} {
name := "single-route writes"
if batch {
name = "batch writes"
}
t.Run(name, func(t *testing.T) {
t.Parallel()
db, err := New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
defer func() { _ = db.Close() }()
ctx := context.Background()
for _, step := range steps {
err := announceRoutes(db, batch, step.announce)
if err != nil {
t.Fatalf("%s: announce: %v", step.name, err)
}
err = withdrawRoutes(db, batch, step.withdraw)
if err != nil {
t.Fatalf("%s: withdraw: %v", step.name, err)
}
stats, err := db.GetStatsContext(ctx)
if err != nil {
t.Fatalf("%s: GetStatsContext: %v", step.name, err)
}
queryV4, queryV6, err := db.GetPrefixDistributionContext(ctx)
if err != nil {
t.Fatalf("%s: GetPrefixDistributionContext: %v", step.name, err)
}
assertDistribution(t, step.name+": IPv4 distribution",
stats.IPv4PrefixDistribution, step.wantV4)
assertDistribution(t, step.name+": IPv6 distribution",
stats.IPv6PrefixDistribution, step.wantV6)
assertDistribution(t, step.name+": IPv4 distribution query",
queryV4, step.wantV4)
assertDistribution(t, step.name+": IPv6 distribution query",
queryV6, step.wantV6)
}
})
}
}
// prefixDistributionStep is one step of TestPrefixDistributionTracksWrites:
// the routes it announces, then the routes it withdraws, then the
// distributions both the stats and the distribution query must report.
type prefixDistributionStep struct {
name string
announce []*LiveRoute
withdraw []*LiveRoute
wantV4 []PrefixDistribution
wantV6 []PrefixDistribution
}
// prefixDistributionSteps returns the steps of
// TestPrefixDistributionTracksWrites, in the order they run.
func prefixDistributionSteps(t *testing.T) []prefixDistributionStep {
t.Helper()
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
shared := mkV4Route(t, "198.51.100.0/24", 64500, ts)
sharedSecondPeer := mkV4Route(t, "198.51.100.0/24", 64500, ts)
sharedSecondPeer.PeerIP = "192.0.2.2"
other := mkV4Route(t, "203.0.113.0/24", 64501, ts)
wide := mkV4Route(t, "172.16.0.0/16", 64502, ts)
v6 := mkV6Route("2001:db8::/32", 64503, ts)
v6SecondPeer := mkV6Route("2001:db8::/32", 64503, ts)
v6 := mkV6Route(testV6Prefix, 64503, ts)
v6SecondPeer := mkV6Route(testV6Prefix, 64503, ts)
v6SecondPeer.PeerIP = "2001:db8::2"
all := []PrefixDistribution{{MaskLength: 16, Count: 1}, {MaskLength: 24, Count: 2}}
v6Only := []PrefixDistribution{{MaskLength: 32, Count: 1}}
steps := []struct {
name string
announce []*LiveRoute
withdraw []*LiveRoute
wantV4 []PrefixDistribution
wantV6 []PrefixDistribution
}{
return []prefixDistributionStep{
{
name: "new routes",
announce: []*LiveRoute{shared, other, wide, v6},
@@ -239,7 +351,9 @@ func TestPrefixDistributionTracksWrites(t *testing.T) {
{
name: "withdrawal of the last route for a prefix",
withdraw: []*LiveRoute{sharedSecondPeer, v6SecondPeer},
wantV4: []PrefixDistribution{{MaskLength: 16, Count: 1}, {MaskLength: 24, Count: 1}},
wantV4: []PrefixDistribution{
{MaskLength: 16, Count: 1}, {MaskLength: 24, Count: 1},
},
},
{
name: "withdrawal of every remaining route",
@@ -270,45 +384,6 @@ func TestPrefixDistributionTracksWrites(t *testing.T) {
wantV6: v6Only,
},
}
for _, batch := range []bool{true, false} {
name := "single-route writes"
if batch {
name = "batch writes"
}
t.Run(name, func(t *testing.T) {
db, err := New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
defer func() { _ = db.Close() }()
ctx := context.Background()
for _, step := range steps {
if err := announceRoutes(db, batch, step.announce); err != nil {
t.Fatalf("%s: announce: %v", step.name, err)
}
if err := withdrawRoutes(db, batch, step.withdraw); err != nil {
t.Fatalf("%s: withdraw: %v", step.name, err)
}
stats, err := db.GetStatsContext(ctx)
if err != nil {
t.Fatalf("%s: GetStatsContext: %v", step.name, err)
}
queryV4, queryV6, err := db.GetPrefixDistributionContext(ctx)
if err != nil {
t.Fatalf("%s: GetPrefixDistributionContext: %v", step.name, err)
}
assertDistribution(t, step.name+": IPv4 distribution", stats.IPv4PrefixDistribution, step.wantV4)
assertDistribution(t, step.name+": IPv6 distribution", stats.IPv6PrefixDistribution, step.wantV6)
assertDistribution(t, step.name+": IPv4 distribution query", queryV4, step.wantV4)
assertDistribution(t, step.name+": IPv6 distribution query", queryV6, step.wantV6)
}
})
}
}
// announceRoutes writes routes in one UpsertLiveRouteBatch, or with one
@@ -317,8 +392,10 @@ func announceRoutes(db *Database, batch bool, routes []*LiveRoute) error {
if batch {
return db.UpsertLiveRouteBatch(routes)
}
for _, route := range routes {
if err := db.UpsertLiveRoute(route); err != nil {
err := db.UpsertLiveRoute(route)
if err != nil {
return err
}
}
@@ -331,7 +408,8 @@ func announceRoutes(db *Database, batch bool, routes []*LiveRoute) error {
func withdrawRoutes(db *Database, batch bool, routes []*LiveRoute) error {
if !batch {
for _, route := range routes {
if err := db.DeleteLiveRoute(route.Prefix, route.PeerIP); err != nil {
err := db.DeleteLiveRoute(route.Prefix, route.PeerIP)
if err != nil {
return err
}
}
@@ -354,6 +432,8 @@ func withdrawRoutes(db *Database, batch bool, routes []*LiveRoute) error {
// routes stay, and the route counts and prefix distribution held in memory
// equal what is read from the route tables.
func TestDeleteLiveRoutesOfPeer(t *testing.T) {
t.Parallel()
const (
peerIP = "192.0.2.1"
otherPeerIP = "192.0.2.2"
@@ -375,14 +455,16 @@ func TestDeleteLiveRoutesOfPeer(t *testing.T) {
}
onlyV6 := mkV6Route("2001:db8:1::/48", 64503, ts)
onlyV6.MaskLength = 48
if err := db.UpsertLiveRouteBatch([]*LiveRoute{
err = db.UpsertLiveRouteBatch([]*LiveRoute{
ofPeer(peerIP, mkV4Route(t, sharedV4, 64500, ts)),
ofPeer(otherPeerIP, mkV4Route(t, sharedV4, 64500, ts)),
ofPeer(peerIP, mkV4Route(t, "203.0.113.0/24", 64501, ts)),
ofPeer(peerIP, mkV6Route(sharedV6, 64502, ts)),
ofPeer(otherPeerIP, mkV6Route(sharedV6, 64502, ts)),
ofPeer(peerIP, onlyV6),
}); err != nil {
})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
@@ -392,40 +474,61 @@ func TestDeleteLiveRoutesOfPeer(t *testing.T) {
if err != nil {
t.Fatalf("DeleteLiveRoutesOfPeer: %v", err)
}
if deleted != want {
t.Errorf("DeleteLiveRoutesOfPeer deleted %d routes, want %d", deleted, want)
t.Errorf("DeleteLiveRoutesOfPeer deleted %d routes, want %d",
deleted, want)
}
}
ctx := context.Background()
for _, prefix := range []string{sharedV4, sharedV6} {
routes, _, err := db.GetPrefixDetailsContext(ctx, prefix)
routes, _, err := db.GetPrefixDetailsContext(context.Background(), prefix)
if err != nil {
t.Fatalf("GetPrefixDetailsContext(%s): %v", prefix, err)
}
if len(routes) != 1 || routes[0].PeerIP != otherPeerIP {
t.Errorf("live routes for %s = %+v, want only the one from %s", prefix, routes, otherPeerIP)
t.Errorf("live routes for %s = %+v, want only the one from %s",
prefix, routes, otherPeerIP)
}
}
checkOneLiveRoutePerIPVersion(t, db)
}
// checkOneLiveRoutePerIPVersion checks that the database has one IPv4 live
// route, for a /24, and one IPv6 live route, for a /32, both as counted in
// memory and as read from the route tables.
func checkOneLiveRoutePerIPVersion(t *testing.T, db *Database) {
t.Helper()
ctx := context.Background()
stats, err := db.GetStatsContext(ctx)
if err != nil {
t.Fatalf("GetStatsContext: %v", err)
}
tableV4, tableV6, err := db.GetLiveRouteCountsContext(ctx)
if err != nil {
t.Fatalf("GetLiveRouteCountsContext: %v", err)
}
if stats.IPv4Routes != 1 || tableV4 != 1 || stats.IPv6Routes != 1 || tableV6 != 1 {
t.Errorf("live routes (IPv4, IPv6): (%d, %d) in memory, (%d, %d) in the tables, want (1, 1)",
if stats.IPv4Routes != 1 || tableV4 != 1 || stats.IPv6Routes != 1 ||
tableV6 != 1 {
t.Errorf("live routes (IPv4, IPv6): (%d, %d) in memory, (%d, %d) in "+
"the tables, want (1, 1)",
stats.IPv4Routes, stats.IPv6Routes, tableV4, tableV6)
}
queryV4, queryV6, err := db.GetPrefixDistributionContext(ctx)
if err != nil {
t.Fatalf("GetPrefixDistributionContext: %v", err)
}
wantV4 := []PrefixDistribution{{MaskLength: 24, Count: 1}}
wantV6 := []PrefixDistribution{{MaskLength: 32, Count: 1}}
assertDistribution(t, "IPv4 distribution", stats.IPv4PrefixDistribution, wantV4)
assertDistribution(t, "IPv6 distribution", stats.IPv6PrefixDistribution, wantV6)
assertDistribution(t, "IPv4 distribution query", queryV4, wantV4)
@@ -437,6 +540,8 @@ func TestDeleteLiveRoutesOfPeer(t *testing.T) {
// MIN/MAX union query read its result into *time.Time, which the driver could
// not parse, so it logged a warning every call and left both timestamps nil.
func TestStatsRouteTimestamps(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -451,6 +556,7 @@ func TestStatsRouteTimestamps(t *testing.T) {
if err != nil {
t.Fatalf("GetStatsContext on empty database: %v", err)
}
if empty.OldestRoute != nil || empty.NewestRoute != nil {
t.Fatalf("empty database timestamps = (%v, %v), want (nil, nil)",
empty.OldestRoute, empty.NewestRoute)
@@ -459,11 +565,13 @@ func TestStatsRouteTimestamps(t *testing.T) {
base := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
oldest := base
newest := base.Add(2 * time.Minute)
if err := db.UpsertLiveRouteBatch([]*LiveRoute{
err = db.UpsertLiveRouteBatch([]*LiveRoute{
mkV4Route(t, "198.51.100.0/24", 64500, base.Add(time.Minute)),
mkV4Route(t, "203.0.113.0/24", 64501, oldest),
mkV6Route("2001:db8::/32", 64502, newest),
}); err != nil {
mkV6Route(testV6Prefix, 64502, newest),
})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
@@ -471,9 +579,11 @@ func TestStatsRouteTimestamps(t *testing.T) {
if err != nil {
t.Fatalf("GetStatsContext: %v", err)
}
if stats.OldestRoute == nil || !stats.OldestRoute.Equal(oldest) {
t.Errorf("OldestRoute = %v, want %v", stats.OldestRoute, oldest)
}
if stats.NewestRoute == nil || !stats.NewestRoute.Equal(newest) {
t.Errorf("NewestRoute = %v, want %v", stats.NewestRoute, newest)
}
@@ -482,6 +592,8 @@ func TestStatsRouteTimestamps(t *testing.T) {
// TestLiveCountsConcurrentReadWrite runs writers and stats readers at once so
// the race detector proves the counters are safe under concurrent use.
func TestLiveCountsConcurrentReadWrite(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -493,17 +605,21 @@ func TestLiveCountsConcurrentReadWrite(t *testing.T) {
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
const writers = 4
var wg sync.WaitGroup
var wg sync.WaitGroup
wg.Add(writers)
for w := range writers {
go func(base int) {
defer wg.Done()
for i := range 25 {
asn := 65000 + base*100 + i
route := mkV6Route("2001:db8::/32", asn, ts)
route := mkV6Route(testV6Prefix, asn, ts)
route.PeerIP = "2001:db8::" + uuid.NewString()
if err := db.UpsertLiveRoute(route); err != nil {
err := db.UpsertLiveRoute(route)
if err != nil {
t.Errorf("UpsertLiveRoute: %v", err)
return
@@ -514,15 +630,19 @@ func TestLiveCountsConcurrentReadWrite(t *testing.T) {
var readerWG sync.WaitGroup
readerWG.Add(1)
stop := make(chan struct{})
go func() {
defer readerWG.Done()
for {
select {
case <-stop:
return
default:
if _, err := db.GetStatsContext(context.Background()); err != nil {
_, err := db.GetStatsContext(context.Background())
if err != nil {
t.Errorf("GetStatsContext: %v", err)
return
@@ -539,6 +659,7 @@ func TestLiveCountsConcurrentReadWrite(t *testing.T) {
if err != nil {
t.Fatalf("GetStatsContext: %v", err)
}
if want := writers * 25; stats.IPv6Routes != want {
t.Errorf("IPv6Routes = %d, want %d", stats.IPv6Routes, want)
}
@@ -563,21 +684,27 @@ func assertCounts(t *testing.T, when string, got Stats, want wantCounts) {
if got.ASNs != want.asns {
t.Errorf("%s: ASNs = %d, want %d", when, got.ASNs, want.asns)
}
if got.Prefixes != want.prefixes {
t.Errorf("%s: Prefixes = %d, want %d", when, got.Prefixes, want.prefixes)
}
if got.Peerings != want.peerings {
t.Errorf("%s: Peerings = %d, want %d", when, got.Peerings, want.peerings)
}
if got.Peers != want.peers {
t.Errorf("%s: Peers = %d, want %d", when, got.Peers, want.peers)
}
if got.IPv4Routes != want.ipv4Routes {
t.Errorf("%s: IPv4Routes = %d, want %d", when, got.IPv4Routes, want.ipv4Routes)
}
if got.IPv6Routes != want.ipv6Routes {
t.Errorf("%s: IPv6Routes = %d, want %d", when, got.IPv6Routes, want.ipv6Routes)
}
if got.LiveRoutes != want.liveRoutes {
t.Errorf("%s: LiveRoutes = %d, want %d", when, got.LiveRoutes, want.liveRoutes)
}
internal/database/database.go
+1301 -971
View File
File diff suppressed because it is too large Load Diff
internal/database/database_test.go → internal/database/database_internal_test.go
+231 -114
View File
@@ -78,6 +78,8 @@ const (
// address, ErrNoRoute when none covers it, and the next less specific prefix
// once the only route of the most specific one is withdrawn.
func TestGetIPInfoFindsMostSpecificLiveRoute(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -95,24 +97,7 @@ func TestGetIPInfoFindsMostSpecificLiveRoute(t *testing.T) {
"2001:db8:1::/48": 64501,
"2001:db8:1:2::/64": 64502,
}
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
routes := make([]*LiveRoute, 0, len(origins))
for prefix, asn := range origins {
routes = append(routes, &LiveRoute{
ID: uuid.New(),
Prefix: prefix,
MaskLength: netip.MustParsePrefix(prefix).Bits(),
IPVersion: detectIPVersion(prefix),
OriginASN: asn,
PeerIP: testPeerIP,
ASPath: []int{asn},
NextHop: testPeerIP,
LastUpdated: ts,
})
}
if err := db.UpsertLiveRouteBatch(routes); err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
writeOriginRoutes(t, db, origins)
// lookup checks that ip resolves to the live prefix want, or to ErrNoRoute
// when want is empty.
@@ -127,11 +112,13 @@ func TestGetIPInfoFindsMostSpecificLiveRoute(t *testing.T) {
return
}
if err != nil {
t.Errorf("GetIPInfo(%s): %v", ip, err)
return
}
if info.Netblock != want || info.MaskLength != netip.MustParsePrefix(want).Bits() ||
info.ASN != origins[want] {
t.Errorf("GetIPInfo(%s) = %s (mask %d) AS%d, want %s AS%d",
@@ -166,6 +153,8 @@ func TestGetIPInfoFindsMostSpecificLiveRoute(t *testing.T) {
// route for an address, and that a route of exactly /8 or /16, or longer, still
// is. The prefixes are written out so that a wrong limit fails the test.
func TestGetIPInfoIgnoresShortRoutes(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -174,24 +163,11 @@ func TestGetIPInfoIgnoresShortRoutes(t *testing.T) {
}
defer func() { _ = db.Close() }()
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
// announce writes one live route for prefix, originated by asn.
announce := func(prefix string, asn int) {
t.Helper()
err := db.UpsertLiveRouteBatch([]*LiveRoute{{
ID: uuid.New(),
Prefix: prefix,
MaskLength: netip.MustParsePrefix(prefix).Bits(),
IPVersion: detectIPVersion(prefix),
OriginASN: asn,
PeerIP: testPeerIP,
ASPath: []int{asn},
NextHop: testPeerIP,
LastUpdated: ts,
}})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch(%s): %v", prefix, err)
}
writeOriginRoutes(t, db, map[string]int{prefix: asn})
}
// lookup checks that ip resolves to the live prefix netblock from asn.
@@ -204,8 +180,10 @@ func TestGetIPInfoIgnoresShortRoutes(t *testing.T) {
return
}
if info.Netblock != netblock || info.ASN != asn {
t.Errorf("GetIPInfo(%s) = %s AS%d, want %s AS%d", ip, info.Netblock, info.ASN, netblock, asn)
t.Errorf("GetIPInfo(%s) = %s AS%d, want %s AS%d",
ip, info.Netblock, info.ASN, netblock, asn)
}
}
@@ -216,8 +194,10 @@ func TestGetIPInfoIgnoresShortRoutes(t *testing.T) {
announce("192.0.0.0/7", 64500)
for _, ip := range []string{"2001:db8::1", "192.0.2.1"} {
if info, err := db.GetIPInfo(ip); !errors.Is(err, ErrNoRoute) {
t.Errorf("GetIPInfo(%s) with only short routes = %+v, %v; want ErrNoRoute", ip, info, err)
info, err := db.GetIPInfo(ip)
if !errors.Is(err, ErrNoRoute) {
t.Errorf("GetIPInfo(%s) with only short routes = %+v, %v; "+
"want ErrNoRoute", ip, info, err)
}
}
@@ -238,6 +218,8 @@ func TestGetIPInfoIgnoresShortRoutes(t *testing.T) {
// origin AS. An upsert updates the peer's route before it would insert one, so
// only an insert straight into the table shows what the table allows.
func TestLiveRouteTablesHoldOneRoutePerPrefixAndPeer(t *testing.T) {
t.Parallel()
db, err := New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
@@ -245,33 +227,76 @@ func TestLiveRouteTablesHoldOneRoutePerPrefixAndPeer(t *testing.T) {
defer func() { _ = db.Close() }()
const otherOriginASN = 64501
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
for _, tt := range []struct {
route *LiveRoute
insertSQL string
}{
{route: mkV4Route(t, "198.51.100.0/24", 64500, ts), insertSQL: insertLiveRouteV4SQL},
{route: mkV6Route("2001:db8::/32", 64500, ts), insertSQL: insertLiveRouteV6SQL},
{
route: mkV4Route(t, "198.51.100.0/24", 64500, ts),
insertSQL: insertLiveRouteV4SQL,
},
{
route: mkV6Route(testV6Prefix, 64500, ts),
insertSQL: insertLiveRouteV6SQL,
},
} {
if err := db.UpsertLiveRoute(tt.route); err != nil {
err := db.UpsertLiveRoute(tt.route)
if err != nil {
t.Fatalf("UpsertLiveRoute(%s): %v", tt.route.Prefix, err)
}
_, err := db.db.Exec(tt.insertSQL, uuid.NewString(), tt.route.Prefix, tt.route.MaskLength,
otherOriginASN, tt.route.PeerIP, "[]", tt.route.NextHop, ts)
_, err = db.db.ExecContext(t.Context(), tt.insertSQL, uuid.NewString(),
tt.route.Prefix, tt.route.MaskLength, otherOriginASN,
tt.route.PeerIP, "[]", tt.route.NextHop, ts)
var sqliteErr sqlite3.Error
if !errors.As(err, &sqliteErr) || sqliteErr.ExtendedCode != sqlite3.ErrConstraintUnique {
t.Errorf("second route from peer %s for %s: error %v, want a UNIQUE constraint error",
if !errors.As(err, &sqliteErr) ||
sqliteErr.ExtendedCode != sqlite3.ErrConstraintUnique {
t.Errorf("second route from peer %s for %s: error %v, "+
"want a UNIQUE constraint error",
tt.route.PeerIP, tt.route.Prefix, err)
}
}
}
// writeOriginRoutes writes one live route from testPeerIP for each prefix in
// origins, originated by the AS it maps to.
func writeOriginRoutes(t *testing.T, db *Database, origins map[string]int) {
t.Helper()
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
routes := make([]*LiveRoute, 0, len(origins))
for prefix, asn := range origins {
routes = append(routes, &LiveRoute{
ID: uuid.New(),
Prefix: prefix,
MaskLength: netip.MustParsePrefix(prefix).Bits(),
IPVersion: detectIPVersion(prefix),
OriginASN: asn,
PeerIP: testPeerIP,
ASPath: []int{asn},
NextHop: testPeerIP,
LastUpdated: ts,
})
}
err := db.UpsertLiveRouteBatch(routes)
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
}
// TestConnectionPoolPragmas holds several pooled connections open at once and
// checks each one carries the per-connection settings from the DSN, plus the
// process-wide hard heap limit, and sees the new file with auto_vacuum
// incremental.
func TestConnectionPoolPragmas(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -291,57 +316,64 @@ func TestConnectionPoolPragmas(t *testing.T) {
}
}()
for i := 0; i < heldConnections; i++ {
for i := range heldConnections {
c, err := db.db.Conn(ctx)
if err != nil {
t.Fatalf("failed to open connection %d: %v", i, err)
}
conns = append(conns, c)
}
for i, c := range conns {
var cacheSize int
if err := c.QueryRowContext(ctx, "PRAGMA cache_size").Scan(&cacheSize); err != nil {
t.Fatalf("conn %d: failed to read cache_size: %v", i, err)
}
cacheSize := readPragma(ctx, t, c, i, "cache_size")
if cacheSize != sqliteCacheSizeKiB {
t.Errorf("conn %d: cache_size = %d, want %d", i, cacheSize, sqliteCacheSizeKiB)
t.Errorf("conn %d: cache_size = %d, want %d",
i, cacheSize, sqliteCacheSizeKiB)
}
var busyTimeout int
if err := c.QueryRowContext(ctx, "PRAGMA busy_timeout").Scan(&busyTimeout); err != nil {
t.Fatalf("conn %d: failed to read busy_timeout: %v", i, err)
}
busyTimeout := readPragma(ctx, t, c, i, "busy_timeout")
if busyTimeout != sqliteBusyTimeoutMs {
t.Errorf("conn %d: busy_timeout = %d, want %d", i, busyTimeout, sqliteBusyTimeoutMs)
t.Errorf("conn %d: busy_timeout = %d, want %d",
i, busyTimeout, sqliteBusyTimeoutMs)
}
var tempStore int
if err := c.QueryRowContext(ctx, "PRAGMA temp_store").Scan(&tempStore); err != nil {
t.Fatalf("conn %d: failed to read temp_store: %v", i, err)
}
tempStore := readPragma(ctx, t, c, i, "temp_store")
if tempStore == tempStoreMemory {
t.Errorf("conn %d: temp_store = %d, want anything but %d (MEMORY)", i, tempStore, tempStoreMemory)
t.Errorf("conn %d: temp_store = %d, want anything but %d (MEMORY)",
i, tempStore, tempStoreMemory)
}
var hardHeapLimit int64
if err := c.QueryRowContext(ctx, "PRAGMA hard_heap_limit").Scan(&hardHeapLimit); err != nil {
t.Fatalf("conn %d: failed to read hard_heap_limit: %v", i, err)
}
hardHeapLimit := readPragma(ctx, t, c, i, "hard_heap_limit")
if hardHeapLimit != sqliteHardHeapLimitBytes {
t.Errorf("conn %d: hard_heap_limit = %d, want %d", i, hardHeapLimit, sqliteHardHeapLimitBytes)
t.Errorf("conn %d: hard_heap_limit = %d, want %d",
i, hardHeapLimit, sqliteHardHeapLimitBytes)
}
var autoVacuum int
if err := c.QueryRowContext(ctx, "PRAGMA auto_vacuum").Scan(&autoVacuum); err != nil {
t.Fatalf("conn %d: failed to read auto_vacuum: %v", i, err)
}
autoVacuum := readPragma(ctx, t, c, i, "auto_vacuum")
if autoVacuum != autoVacuumIncremental {
t.Errorf("conn %d: auto_vacuum = %d, want %d (incremental)", i, autoVacuum, autoVacuumIncremental)
t.Errorf("conn %d: auto_vacuum = %d, want %d (incremental)",
i, autoVacuum, autoVacuumIncremental)
}
}
}
// readPragma reads the integer value of PRAGMA name on connection i, c.
func readPragma(
ctx context.Context, t *testing.T, c *sql.Conn, i int, name string,
) int64 {
t.Helper()
var value int64
err := c.QueryRowContext(ctx, "PRAGMA "+name).Scan(&value)
if err != nil {
t.Fatalf("conn %d: failed to read %s: %v", i, name, err)
}
return value
}
// TestVacuumReturnsFreePages checks that after routes are deleted from a new
// database, one Vacuum call returns every page they used. The deletes leave
// fewer free pages than the 1000 Vacuum frees per call, so none may remain.
@@ -349,6 +381,8 @@ func TestConnectionPoolPragmas(t *testing.T) {
// the free pages stayed in the file, and with the PRAGMA run by ExecContext
// Vacuum freed only one page per call.
func TestVacuumReturnsFreePages(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -359,10 +393,12 @@ func TestVacuumReturnsFreePages(t *testing.T) {
ctx := context.Background()
ts := time.Now().UTC()
const asn = 64500
routes := make([]*LiveRoute, 0, vacuumTestRoutes)
deletions := make([]LiveRouteDeletion, 0, vacuumTestRoutes)
for i := range vacuumTestRoutes {
route := mkV4Route(t, fmt.Sprintf("10.%d.%d.0/24", i/256, i%256), asn, ts)
routes = append(routes, route)
@@ -373,29 +409,39 @@ func TestVacuumReturnsFreePages(t *testing.T) {
})
}
if err := db.UpsertLiveRouteBatch(routes); err != nil {
err = db.UpsertLiveRouteBatch(routes)
if err != nil {
t.Fatalf("failed to write routes: %v", err)
}
if err := db.DeleteLiveRouteBatch(deletions); err != nil {
err = db.DeleteLiveRouteBatch(deletions)
if err != nil {
t.Fatalf("failed to delete routes: %v", err)
}
var before int
if err := db.db.QueryRowContext(ctx, "PRAGMA freelist_count").Scan(&before); err != nil {
err = db.db.QueryRowContext(ctx, "PRAGMA freelist_count").Scan(&before)
if err != nil {
t.Fatalf("failed to read freelist_count: %v", err)
}
if before == 0 {
t.Fatalf("no free pages after deleting %d routes", vacuumTestRoutes)
}
if err := db.Vacuum(ctx); err != nil {
err = db.Vacuum(ctx)
if err != nil {
t.Fatalf("Vacuum failed: %v", err)
}
var after int
if err := db.db.QueryRowContext(ctx, "PRAGMA freelist_count").Scan(&after); err != nil {
err = db.db.QueryRowContext(ctx, "PRAGMA freelist_count").Scan(&after)
if err != nil {
t.Fatalf("failed to read freelist_count: %v", err)
}
if after != 0 {
t.Errorf("free pages after Vacuum = %d of %d, want 0", after, before)
}
@@ -410,6 +456,8 @@ func TestVacuumReturnsFreePages(t *testing.T) {
// no batch is dropped. The checkpoint runs without the Database mutex, exactly
// as the background maintainer does in production.
func TestBatchWriteDuringCheckpoint(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -422,8 +470,10 @@ func TestBatchWriteDuringCheckpoint(t *testing.T) {
var wg sync.WaitGroup
wg.Add(1)
go func() {
defer wg.Done()
for {
select {
case <-ctx.Done():
@@ -435,12 +485,14 @@ func TestBatchWriteDuringCheckpoint(t *testing.T) {
}()
ts := time.Now().UTC()
for i := 0; i < contentionIterations; i++ {
for i := range contentionIterations {
asns := map[int]time.Time{
i % contendedASNCount: ts,
(i % contendedASNCount) + asnSecondBand: ts,
}
if err := db.GetOrCreateASNBatch(asns); err != nil {
err := db.GetOrCreateASNBatch(asns)
if err != nil {
cancel()
wg.Wait()
t.Fatalf("batch write failed under checkpoint contention: %v", err)
@@ -455,6 +507,8 @@ func TestBatchWriteDuringCheckpoint(t *testing.T) {
// family learned from two peers in the same second and checks that the
// prefix-length sample lists each of them once.
func TestGetRandomPrefixesByLengthListsPrefixOnce(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -467,18 +521,24 @@ func TestGetRandomPrefixesByLengthListsPrefixOnce(t *testing.T) {
v4First := mkV4Route(t, "198.51.100.0/24", 64500, ts)
v4Second := mkV4Route(t, "198.51.100.0/24", 64500, ts)
v4Second.PeerIP = secondV4PeerIP
v6First := mkV6Route("2001:db8::/32", 64500, ts)
v6Second := mkV6Route("2001:db8::/32", 64500, ts)
v6First := mkV6Route(testV6Prefix, 64500, ts)
v6Second := mkV6Route(testV6Prefix, 64500, ts)
v6Second.PeerIP = secondV6PeerIP
if err := db.UpsertLiveRouteBatch([]*LiveRoute{v4First, v4Second, v6First, v6Second}); err != nil {
err = db.UpsertLiveRouteBatch(
[]*LiveRoute{v4First, v4Second, v6First, v6Second})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
for _, want := range []*LiveRoute{v4First, v6First} {
routes, err := db.GetRandomPrefixesByLength(want.MaskLength, want.IPVersion, sampleLimit)
routes, err := db.GetRandomPrefixesByLength(want.MaskLength,
want.IPVersion, sampleLimit)
if err != nil {
t.Fatalf("GetRandomPrefixesByLength(%d, %d): %v", want.MaskLength, want.IPVersion, err)
t.Fatalf("GetRandomPrefixesByLength(%d, %d): %v",
want.MaskLength, want.IPVersion, err)
}
if len(routes) != 1 || routes[0].Prefix != want.Prefix {
t.Errorf("GetRandomPrefixesByLength(%d, %d) = %+v, want %s listed once",
want.MaskLength, want.IPVersion, routes, want.Prefix)
@@ -490,6 +550,8 @@ func TestGetRandomPrefixesByLengthListsPrefixOnce(t *testing.T) {
// two peers at different times, the older route first, and checks that the
// prefix-length sample lists the newer route.
func TestGetRandomPrefixesByLengthListsNewestRoute(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -502,17 +564,21 @@ func TestGetRandomPrefixesByLengthListsNewestRoute(t *testing.T) {
olderRoute := mkV4Route(t, "198.51.100.0/24", 64500, older)
newerRoute := mkV4Route(t, "198.51.100.0/24", 64500, older.Add(time.Minute))
newerRoute.PeerIP = secondV4PeerIP
if err := db.UpsertLiveRouteBatch([]*LiveRoute{olderRoute, newerRoute}); err != nil {
err = db.UpsertLiveRouteBatch([]*LiveRoute{olderRoute, newerRoute})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
routes, err := db.GetRandomPrefixesByLength(newerRoute.MaskLength, ipVersionV4, sampleLimit)
routes, err := db.GetRandomPrefixesByLength(newerRoute.MaskLength,
ipVersionV4, sampleLimit)
if err != nil {
t.Fatalf("GetRandomPrefixesByLength: %v", err)
}
if len(routes) != 1 || routes[0].PeerIP != newerRoute.PeerIP {
t.Errorf("GetRandomPrefixesByLength = %+v, want only the newer route, from %s",
routes, newerRoute.PeerIP)
t.Errorf("GetRandomPrefixesByLength = %+v, want only the newer route, "+
"from %s", routes, newerRoute.PeerIP)
}
}
@@ -523,6 +589,8 @@ func TestGetRandomPrefixesByLengthListsNewestRoute(t *testing.T) {
// shuffled prefixes, and the random lookups used for a mask length with more
// than maxPrefixesToShuffle prefixes.
func TestGetRandomPrefixesByLengthSamplesManyPrefixes(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -532,8 +600,11 @@ func TestGetRandomPrefixesByLengthSamplesManyPrefixes(t *testing.T) {
defer func() { _ = db.Close() }()
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
var routes []*LiveRoute
var stored []string
stored := make([]string, 0, sampledPrefixCount)
for i := range sampledPrefixCount {
prefix := fmt.Sprintf("10.0.%d.0/24", i)
for _, p := range []string{prefix, fmt.Sprintf("172.%d.0.0/16", i)} {
@@ -542,22 +613,29 @@ func TestGetRandomPrefixesByLengthSamplesManyPrefixes(t *testing.T) {
second.PeerIP = secondV4PeerIP
routes = append(routes, first, second)
}
stored = append(stored, prefix)
}
if err := db.UpsertLiveRouteBatch(routes); err != nil {
err = db.UpsertLiveRouteBatch(routes)
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
slices.Sort(stored)
maskLength := routes[0].MaskLength
sample, err := db.GetRandomPrefixesByLength(maskLength, ipVersionV4, sampleLimit)
if err != nil {
t.Fatalf("GetRandomPrefixesByLength: %v", err)
}
var listed []string
listed := make([]string, 0, len(sample))
for _, route := range sample {
listed = append(listed, route.Prefix)
}
checkPrefixSample(t, "GetRandomPrefixesByLength", listed, stored)
looked, err := db.samplePrefixesByLength(
@@ -565,6 +643,7 @@ func TestGetRandomPrefixesByLengthSamplesManyPrefixes(t *testing.T) {
if err != nil {
t.Fatalf("samplePrefixesByLength: %v", err)
}
checkPrefixSample(t, "samplePrefixesByLength", looked, stored)
}
@@ -577,15 +656,20 @@ func checkPrefixSample(t *testing.T, name string, sample, stored []string) {
if len(sample) != sampleLimit {
t.Errorf("%s listed %d prefixes, want %d", name, len(sample), sampleLimit)
}
listed := make(map[string]bool)
for _, prefix := range sample {
if listed[prefix] || !slices.Contains(stored, prefix) {
t.Errorf("%s lists %s twice or it is not one of the /24s", name, prefix)
}
listed[prefix] = true
}
if slices.Equal(slices.Sorted(slices.Values(sample)), stored[:sampleLimit]) {
t.Errorf("%s lists just the first %d prefixes in text order", name, sampleLimit)
t.Errorf("%s lists just the first %d prefixes in text order",
name, sampleLimit)
}
}
@@ -595,6 +679,8 @@ func checkPrefixSample(t *testing.T, name string, sample, stored []string) {
// order, and that the sample lists the prefixes asked for, each once, some of
// them after those first ones, which only the random lookups reach.
func TestGetRandomPrefixesByLengthSamplesPastFirstPrefixes(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -604,51 +690,64 @@ func TestGetRandomPrefixesByLengthSamplesPastFirstPrefixes(t *testing.T) {
defer func() { _ = db.Close() }()
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
var routes []*LiveRoute
var stored []string
routes := make([]*LiveRoute, 0, largeSamplePrefixCount)
stored := make([]string, 0, largeSamplePrefixCount)
for i := range largeSamplePrefixCount {
route := mkV6Route(fmt.Sprintf("2001:%x::/32", i), 64500, ts)
routes = append(routes, route)
stored = append(stored, route.Prefix)
}
if err := db.UpsertLiveRouteBatch(routes); err != nil {
err = db.UpsertLiveRouteBatch(routes)
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
slices.Sort(stored)
firstPrefixes := stored[:maxPrefixesToShuffle+1]
maskLength := routes[0].MaskLength
read, err := db.firstPrefixesByLength(
context.Background(), liveRoutesV6Table, maskLength, maxPrefixesToShuffle+1)
read, err := db.firstPrefixesByLength(context.Background(),
liveRoutesV6Table, maskLength, maxPrefixesToShuffle+1)
if err != nil {
t.Fatalf("firstPrefixesByLength: %v", err)
}
if !slices.Equal(read, firstPrefixes) {
t.Errorf("firstPrefixesByLength read %d prefixes, want the first %d in text order",
len(read), len(firstPrefixes))
t.Errorf("firstPrefixesByLength read %d prefixes, want the first %d "+
"in text order", len(read), len(firstPrefixes))
}
sample, err := db.GetRandomPrefixesByLength(maskLength, ipVersionV6, largeSampleLimit)
sample, err := db.GetRandomPrefixesByLength(maskLength, ipVersionV6,
largeSampleLimit)
if err != nil {
t.Fatalf("GetRandomPrefixesByLength: %v", err)
}
if len(sample) != largeSampleLimit {
t.Errorf("GetRandomPrefixesByLength listed %d prefixes, want %d", len(sample), largeSampleLimit)
t.Errorf("GetRandomPrefixesByLength listed %d prefixes, want %d",
len(sample), largeSampleLimit)
}
listed := make(map[string]bool)
pastFirstPrefixes := 0
for _, route := range sample {
if listed[route.Prefix] {
t.Errorf("GetRandomPrefixesByLength lists %s twice", route.Prefix)
}
listed[route.Prefix] = true
if !slices.Contains(firstPrefixes, route.Prefix) {
pastFirstPrefixes++
}
}
if pastFirstPrefixes == 0 {
t.Errorf("GetRandomPrefixesByLength lists only prefixes among the first %d in text order",
len(firstPrefixes))
t.Errorf("GetRandomPrefixesByLength lists only prefixes among the first "+
"%d in text order", len(firstPrefixes))
}
}
@@ -657,6 +756,8 @@ func TestGetRandomPrefixesByLengthSamplesPastFirstPrefixes(t *testing.T) {
// sampleLimit prefixes: the stored one, then the shuffled ones in order. The
// stored prefix is the second shuffled one, so the fill-up must skip it.
func TestSamplePrefixesByLengthFillsUpFromShuffled(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -665,20 +766,25 @@ func TestSamplePrefixesByLengthFillsUpFromShuffled(t *testing.T) {
}
defer func() { _ = db.Close() }()
var shuffled []string
shuffled := make([]string, 0, sampledPrefixCount)
for i := range sampledPrefixCount {
shuffled = append(shuffled, fmt.Sprintf("10.0.%d.0/24", i))
}
route := mkV4Route(t, shuffled[1], 64500, time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC))
if err := db.UpsertLiveRouteBatch([]*LiveRoute{route}); err != nil {
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
route := mkV4Route(t, shuffled[1], 64500, ts)
err = db.UpsertLiveRouteBatch([]*LiveRoute{route})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
sample, err := db.samplePrefixesByLength(
context.Background(), liveRoutesV4Table, route.MaskLength, sampleLimit, shuffled)
sample, err := db.samplePrefixesByLength(context.Background(),
liveRoutesV4Table, route.MaskLength, sampleLimit, shuffled)
if err != nil {
t.Fatalf("samplePrefixesByLength: %v", err)
}
want := append([]string{shuffled[1], shuffled[0]}, shuffled[2:sampleLimit]...)
if !slices.Equal(sample, want) {
t.Errorf("samplePrefixesByLength = %v, want %v", sample, want)
@@ -691,6 +797,8 @@ func TestSamplePrefixesByLengthFillsUpFromShuffled(t *testing.T) {
// left, and that only the prefixes that still have a route get their newest
// route listed.
func TestGetRandomPrefixesByLengthSkipsWithdrawnRoutes(t *testing.T) {
t.Parallel()
cfg := &config.Config{StateDir: t.TempDir()}
db, err := New(cfg, logger.New())
@@ -702,29 +810,38 @@ func TestGetRandomPrefixesByLengthSkipsWithdrawnRoutes(t *testing.T) {
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
kept := mkV4Route(t, "198.51.100.0/24", 64500, ts)
withdrawn := mkV4Route(t, "172.16.0.0/16", 64500, ts)
if err := db.UpsertLiveRouteBatch([]*LiveRoute{kept, withdrawn}); err != nil {
err = db.UpsertLiveRouteBatch([]*LiveRoute{kept, withdrawn})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
if err := db.DeleteLiveRoute(withdrawn.Prefix, withdrawn.PeerIP); err != nil {
err = db.DeleteLiveRoute(withdrawn.Prefix, withdrawn.PeerIP)
if err != nil {
t.Fatalf("DeleteLiveRoute: %v", err)
}
ctx := context.Background()
sample, err := db.samplePrefixesByLength(
ctx, liveRoutesV4Table, withdrawn.MaskLength, sampleLimit, []string{withdrawn.Prefix})
sample, err := db.samplePrefixesByLength(ctx, liveRoutesV4Table,
withdrawn.MaskLength, sampleLimit, []string{withdrawn.Prefix})
if err != nil {
t.Fatalf("samplePrefixesByLength: %v", err)
}
if len(sample) != 0 {
t.Errorf("samplePrefixesByLength = %v, want none for a mask length with no route left", sample)
t.Errorf("samplePrefixesByLength = %v, want none for a mask length "+
"with no route left", sample)
}
routes, err := db.newestRouteOfEachPrefix(
ctx, liveRoutesV4Table, ipVersionV4, []string{withdrawn.Prefix, kept.Prefix})
routes, err := db.newestRouteOfEachPrefix(ctx, liveRoutesV4Table,
ipVersionV4, []string{withdrawn.Prefix, kept.Prefix})
if err != nil {
t.Fatalf("newestRouteOfEachPrefix: %v", err)
}
if len(routes) != 1 || routes[0].Prefix != kept.Prefix {
t.Errorf("newestRouteOfEachPrefix = %+v, want only the route of %s", routes, kept.Prefix)
t.Errorf("newestRouteOfEachPrefix = %+v, want only the route of %s",
routes, kept.Prefix)
}
}
internal/database/interface.go
+12 -4
View File
@@ -29,6 +29,8 @@ type Stats struct {
// Store defines the interface for database operations. It provides methods for
// managing ASNs, prefixes, announcements, peerings, BGP peers, and live routes.
// Implementations must be safe for concurrent use.
//
//nolint:interfacebloat // the one seam handlers and tests depend on
type Store interface {
// ASN operations
GetOrCreateASN(number int, timestamp time.Time) (*ASN, error)
@@ -58,8 +60,10 @@ type Store interface {
DeleteLiveRoute(prefix, peerIP string) error
DeleteLiveRouteBatch(deletions []LiveRouteDeletion) error
DeleteLiveRoutesOfPeer(peerIP string) (deleted int, err error)
GetPrefixDistribution() (ipv4 []PrefixDistribution, ipv6 []PrefixDistribution, err error)
GetPrefixDistributionContext(ctx context.Context) (ipv4 []PrefixDistribution, ipv6 []PrefixDistribution, err error)
GetPrefixDistribution() (
ipv4 []PrefixDistribution, ipv6 []PrefixDistribution, err error)
GetPrefixDistributionContext(ctx context.Context) (
ipv4 []PrefixDistribution, ipv6 []PrefixDistribution, err error)
GetLiveRouteCounts() (ipv4Count, ipv6Count int, err error)
GetLiveRouteCountsContext(ctx context.Context) (ipv4Count, ipv6Count int, err error)
@@ -78,9 +82,13 @@ type Store interface {
GetASPeers(asn int) ([]ASPeer, error)
GetASPeersContext(ctx context.Context, asn int) ([]ASPeer, error)
GetPrefixDetails(prefix string) ([]LiveRoute, map[int]ASN, error)
GetPrefixDetailsContext(ctx context.Context, prefix string) ([]LiveRoute, map[int]ASN, error)
GetPrefixDetailsContext(
ctx context.Context, prefix string,
) ([]LiveRoute, map[int]ASN, error)
GetRandomPrefixesByLength(maskLength, ipVersion, limit int) ([]LiveRoute, error)
GetRandomPrefixesByLengthContext(ctx context.Context, maskLength, ipVersion, limit int) ([]LiveRoute, error)
GetRandomPrefixesByLengthContext(
ctx context.Context, maskLength, ipVersion, limit int,
) ([]LiveRoute, error)
// Lifecycle
Close() error
internal/database/models.go
+26 -19
View File
@@ -1,4 +1,3 @@
// Package database provides SQLite storage for BGP routing data.
package database
import (
@@ -9,6 +8,8 @@ import (
// ASN represents an Autonomous System Number with its metadata including
// handle, description, WHOIS data, and first/last seen timestamps.
//
//nolint:tagliatelle // field names are the existing JSON API
type ASN struct {
ASN int `json:"asn"`
Handle string `json:"handle"`
@@ -36,37 +37,39 @@ type ASN struct {
// Prefix represents an IP prefix (CIDR block) with its IP version (4 or 6)
// and first/last seen timestamps.
type Prefix struct {
ID uuid.UUID `json:"id"`
Prefix string `json:"prefix"`
IPVersion int `json:"ip_version"` // 4 or 6
FirstSeen time.Time `json:"first_seen"`
LastSeen time.Time `json:"last_seen"`
ID uuid.UUID
Prefix string
IPVersion int // 4 or 6
FirstSeen time.Time
LastSeen time.Time
}
// Announcement represents a BGP announcement or withdrawal event,
// containing the prefix, AS path, origin ASN, peer ASN, next hop, and timestamp.
type Announcement struct {
ID uuid.UUID `json:"id"`
PrefixID uuid.UUID `json:"prefix_id"`
PeerASN int `json:"peer_asn"`
OriginASN int `json:"origin_asn"`
Path string `json:"path"` // JSON-encoded AS path
NextHop string `json:"next_hop"`
Timestamp time.Time `json:"timestamp"`
IsWithdrawal bool `json:"is_withdrawal"`
ID uuid.UUID
PrefixID uuid.UUID
PeerASN int
OriginASN int
Path string // JSON-encoded AS path
NextHop string
Timestamp time.Time
IsWithdrawal bool
}
// ASNPeering represents a peering relationship between two ASNs,
// stored with the lower ASN as ASA and the higher as ASB.
type ASNPeering struct {
ID uuid.UUID `json:"id"`
ASA int `json:"as_a"`
ASB int `json:"as_b"`
FirstSeen time.Time `json:"first_seen"`
LastSeen time.Time `json:"last_seen"`
ID uuid.UUID
ASA int
ASB int
FirstSeen time.Time
LastSeen time.Time
}
// LiveRoute represents a route in the live routing table
//
//nolint:tagliatelle // field names are the existing JSON API
type LiveRoute struct {
ID uuid.UUID `json:"id"`
Prefix string `json:"prefix"`
@@ -80,12 +83,16 @@ type LiveRoute struct {
}
// PrefixDistribution represents the distribution of prefixes by mask length
//
//nolint:tagliatelle // field names are the existing JSON API
type PrefixDistribution struct {
MaskLength int `json:"mask_length"`
Count int `json:"count"`
}
// IPInfo represents comprehensive IP information for the /ip endpoint
//
//nolint:tagliatelle // field names are the existing JSON API
type IPInfo struct {
IP string `json:"ip"`
PTR []string `json:"ptr,omitempty"`
internal/database/slowquery.go
+25 -43
View File
@@ -18,30 +18,12 @@ func logSlowQuery(logger *logger.Logger, query string, start time.Time) {
}
}
// queryRow wraps QueryRow with slow query logging
// nolint:unused // kept for consistency with other query wrappers
func (d *Database) queryRow(query string, args ...interface{}) *sql.Row {
start := time.Now()
defer logSlowQuery(d.logger, query, start)
return d.db.QueryRow(query, args...)
}
// query wraps Query with slow query logging
// nolint:unused // kept for future use to ensure all queries go through slow query logging
func (d *Database) query(query string, args ...interface{}) (*sql.Rows, error) {
start := time.Now()
defer logSlowQuery(d.logger, query, start)
return d.db.Query(query, args...)
}
// exec wraps Exec with slow query logging
func (d *Database) exec(query string, args ...interface{}) error {
func (d *Database) exec(query string, args ...any) error {
start := time.Now()
defer logSlowQuery(d.logger, query, start)
_, err := d.db.Exec(query, args...)
_, err := d.db.ExecContext(context.Background(), query, args...)
return err
}
@@ -49,51 +31,51 @@ func (d *Database) exec(query string, args ...interface{}) error {
// loggingTx wraps sql.Tx to log slow queries
type loggingTx struct {
*sql.Tx
logger *logger.Logger
}
// QueryRow wraps sql.Tx.QueryRow to log slow queries
func (tx *loggingTx) QueryRow(query string, args ...interface{}) *sql.Row {
func (tx *loggingTx) QueryRow(query string, args ...any) *sql.Row {
start := time.Now()
defer logSlowQuery(tx.logger, query, start)
return tx.Tx.QueryRow(query, args...)
}
// Query wraps sql.Tx.Query to log slow queries
func (tx *loggingTx) Query(query string, args ...interface{}) (*sql.Rows, error) {
start := time.Now()
defer logSlowQuery(tx.logger, query, start)
return tx.Tx.Query(query, args...)
}
// QueryContext wraps sql.Tx.QueryContext to log slow queries
func (tx *loggingTx) QueryContext(ctx context.Context, query string, args ...interface{}) (*sql.Rows, error) {
start := time.Now()
defer logSlowQuery(tx.logger, query, start)
return tx.Tx.QueryContext(ctx, query, args...)
return tx.Tx.QueryRowContext(context.Background(), query, args...)
}
// QueryRowContext wraps sql.Tx.QueryRowContext to log slow queries
func (tx *loggingTx) QueryRowContext(ctx context.Context, query string, args ...interface{}) *sql.Row {
func (tx *loggingTx) QueryRowContext(
ctx context.Context, query string, args ...any,
) *sql.Row {
start := time.Now()
defer logSlowQuery(tx.logger, query, start)
return tx.Tx.QueryRowContext(ctx, query, args...)
}
// Exec wraps sql.Tx.Exec to log slow queries
func (tx *loggingTx) Exec(query string, args ...interface{}) (sql.Result, error) {
// QueryContext wraps sql.Tx.QueryContext to log slow queries
func (tx *loggingTx) QueryContext(
ctx context.Context, query string, args ...any,
) (*sql.Rows, error) {
start := time.Now()
defer logSlowQuery(tx.logger, query, start)
return tx.Tx.Exec(query, args...)
//nolint:sqlclosecheck // returns the rows, which the caller closes
return tx.Tx.QueryContext(ctx, query, args...)
}
// Exec wraps sql.Tx.Exec to log slow queries
func (tx *loggingTx) Exec(query string, args ...any) (sql.Result, error) {
start := time.Now()
defer logSlowQuery(tx.logger, query, start)
return tx.Tx.ExecContext(context.Background(), query, args...)
}
// ExecContext wraps sql.Tx.ExecContext to log slow queries
func (tx *loggingTx) ExecContext(ctx context.Context, query string, args ...interface{}) (sql.Result, error) {
func (tx *loggingTx) ExecContext(
ctx context.Context, query string, args ...any,
) (sql.Result, error) {
start := time.Now()
defer logSlowQuery(tx.logger, query, start)
internal/database/utils.go
+1
View File
@@ -27,6 +27,7 @@ func prefixMaskLength(prefix string) (int, error) {
if err != nil {
return 0, fmt.Errorf("invalid prefix %s: %w", prefix, err)
}
maskLength, _ := network.Mask.Size()
return maskLength, nil
internal/logger/logger.go
+1
View File
@@ -74,6 +74,7 @@ func getSourceAttrs() []slog.Attr {
// Get the function name
fn := runtime.FuncForPC(pc)
var funcName string
if fn != nil {
funcName = filepath.Base(fn.Name())
internal/metrics/metrics.go
+10 -4
View File
@@ -64,6 +64,7 @@ func New() *Tracker {
// first counts one reconnect.
func (t *Tracker) SetConnected(connected bool) {
t.isConnected.Store(connected)
if !connected {
return
}
@@ -120,9 +121,11 @@ func (t *Tracker) GetStreamMetrics() StreamMetrics {
if msgCount >= 0 {
totalMessages = uint64(msgCount)
}
if byteCount >= 0 {
totalBytes = uint64(byteCount)
}
if wireByteCount >= 0 {
totalWireBytes = uint64(wireByteCount)
}
@@ -166,7 +169,7 @@ func (t *Tracker) RecordWithdrawal() {
func (t *Tracker) SetBGPPeerCount(count int) {
// BGP peer count is always small (< 1000), so int32 is safe
if count > 0 && count < 1<<31 {
t.bgpPeerCount.Store(int32(count)) //nolint:gosec // count is validated
t.bgpPeerCount.Store(int32(count))
}
}
@@ -214,7 +217,8 @@ type StreamMetrics struct {
TotalMessages uint64
// TotalBytes is the total number of decompressed bytes received since startup
TotalBytes uint64
// TotalWireBytes is the total number of bytes received on the wire (before decompression)
// TotalWireBytes is the total number of bytes received on the wire (before
// decompression)
TotalWireBytes uint64
// ConnectedSince is the time when the current connection was established
ConnectedSince time.Time
@@ -222,9 +226,11 @@ type StreamMetrics struct {
Connected bool
// MessagesPerSec is the rate of messages received per second (1-minute average)
MessagesPerSec float64
// BitsPerSec is the rate of decompressed bits received per second (1-minute average)
// BitsPerSec is the rate of decompressed bits received per second (1-minute
// average)
BitsPerSec float64
// WireBitsPerSec is the rate of bits received on the wire per second (1-minute average)
// WireBitsPerSec is the rate of bits received on the wire per second
// (1-minute average)
WireBitsPerSec float64
// ReconnectCount is the number of reconnections since startup
ReconnectCount uint64
internal/metrics/metrics_test.go
+9 -3
View File
@@ -1,12 +1,18 @@
package metrics
package metrics_test
import "testing"
import (
"testing"
"git.eeqj.de/sneak/routewatch/internal/metrics"
)
// TestReconnectCount connects three times with a disconnect between each, the
// way the streamer does, and expects every connection after the first to count
// one reconnect.
func TestReconnectCount(t *testing.T) {
tracker := New()
t.Parallel()
tracker := metrics.New()
tracker.SetConnected(true)
tracker.SetConnected(false)
internal/ristypes/ris.go
+15 -5
View File
@@ -1,4 +1,5 @@
// Package ristypes defines the data structures for RIS Live BGP messages and announcements.
// Package ristypes defines the data structures for RIS Live BGP messages and
// announcements.
package ristypes
import (
@@ -33,25 +34,30 @@ type ASPath []int
func (p *ASPath) UnmarshalJSON(data []byte) error {
// First try to unmarshal as a simple array of integers
var simple []int
if err := json.Unmarshal(data, &simple); err == nil {
err := json.Unmarshal(data, &simple)
if err == nil {
*p = ASPath(simple)
return nil
}
// If that fails, unmarshal as array of interfaces and flatten
var raw []interface{}
if err := json.Unmarshal(data, &raw); err != nil {
var raw []any
err = json.Unmarshal(data, &raw)
if err != nil {
return err
}
// Flatten the array
result := make([]int, 0)
for _, item := range raw {
switch v := item.(type) {
case float64:
result = append(result, int(v))
case []interface{}:
case []any:
// Nested array - flatten it
for _, nested := range v {
if num, ok := nested.(float64); ok {
@@ -78,6 +84,8 @@ type RISLiveMessage struct {
// It contains metadata about the BGP session (peer, ASN, host) along with
// the actual BGP update data including AS path, communities, announcements,
// and withdrawals.
//
//nolint:tagliatelle // field names are the RIS Live feed's
type RISMessage struct {
Type string `json:"type"`
Timestamp float64 `json:"timestamp"`
@@ -108,6 +116,8 @@ type RISMessage struct {
// RISAnnouncement represents a BGP route announcement within a RIS message.
// It contains the next hop IP address and the list of prefixes being announced
// via that next hop.
//
//nolint:tagliatelle // field names are the RIS Live feed's
type RISAnnouncement struct {
NextHop string `json:"next_hop"`
Prefixes []string `json:"prefixes"`
internal/ristypes/ris_test.go
+24 -7
View File
@@ -1,10 +1,12 @@
package ristypes
package ristypes_test
import (
"bufio"
"encoding/json"
"os"
"testing"
"git.eeqj.de/sneak/routewatch/internal/ristypes"
)
// messageExamplesPath is the captured RIS Live feed used as decode fixtures,
@@ -15,32 +17,43 @@ const messageExamplesPath = "../../docs/message-examples.json"
// streamer does and checks that the fields no handler reads (Community, Raw)
// stay empty while the fields handlers use (Path, Announcements) still decode.
func TestDecodeDropsCommunityAndRaw(t *testing.T) {
t.Parallel()
f, err := os.Open(messageExamplesPath)
if err != nil {
t.Fatalf("open fixtures: %v", err)
}
defer f.Close()
defer func() {
_ = f.Close()
}()
var messages, withPath, withAnnouncements int
scanner := bufio.NewScanner(f)
scanner.Buffer(make([]byte, 0, 64*1024), 1024*1024)
for scanner.Scan() {
line := scanner.Bytes()
if len(line) == 0 {
continue
}
var wrapper RISLiveMessage
if err := json.Unmarshal(line, &wrapper); err != nil {
var wrapper ristypes.RISLiveMessage
err := json.Unmarshal(line, &wrapper)
if err != nil {
t.Fatalf("unmarshal message %d: %v", messages+1, err)
}
messages++
msg := wrapper.Data
if msg.Community != nil {
t.Errorf("message %d: Community decoded, want empty: %v", messages, msg.Community)
t.Errorf("message %d: Community decoded, want empty: %v",
messages, msg.Community)
}
if msg.Raw != "" {
t.Errorf("message %d: Raw decoded, want empty", messages)
}
@@ -48,11 +61,14 @@ func TestDecodeDropsCommunityAndRaw(t *testing.T) {
if len(msg.Path) > 0 {
withPath++
}
if len(msg.Announcements) > 0 {
withAnnouncements++
}
}
if err := scanner.Err(); err != nil {
err = scanner.Err()
if err != nil {
t.Fatalf("scan fixtures: %v", err)
}
@@ -62,6 +78,7 @@ func TestDecodeDropsCommunityAndRaw(t *testing.T) {
// The fixtures include announcement messages; a used field must still decode,
// otherwise an empty Community/Raw would prove nothing.
if withPath == 0 || withAnnouncements == 0 {
t.Fatalf("used fields did not decode: withPath=%d withAnnouncements=%d", withPath, withAnnouncements)
t.Fatalf("used fields did not decode: withPath=%d withAnnouncements=%d",
withPath, withAnnouncements)
}
}
internal/routewatch/app.go
+34 -16
View File
@@ -1,10 +1,13 @@
// Package routewatch contains the primary RouteWatch type that represents a running instance
// of the application and contains pointers to its core dependencies, and is responsible for initialization.
// Package routewatch contains the primary RouteWatch type that represents a
// running instance of the application and contains pointers to its core
// dependencies, and is responsible for initialization. It also contains the
// ASN WHOIS fetcher for background updates and the database maintainer for
// background maintenance tasks.
package routewatch
import (
"context"
"fmt"
"errors"
"sync"
"time"
@@ -18,6 +21,10 @@ import (
"go.uber.org/fx"
)
// errNonBatchedNotImplemented is returned by Run when batched database writes
// are disabled, since only the batched handlers exist.
var errNonBatchedNotImplemented = errors.New("non-batched handlers not implemented")
// Dependencies contains all dependencies for RouteWatch
type Dependencies struct {
fx.In
@@ -68,8 +75,10 @@ func (rw *RouteWatch) Run(ctx context.Context) error {
// Apply runtime limit if specified
if rw.maxRuntime > 0 {
var cancel context.CancelFunc
ctx, cancel = context.WithTimeout(ctx, rw.maxRuntime)
defer cancel()
rw.logger.Info("Running with time limit", "max_runtime", rw.maxRuntime)
}
@@ -96,29 +105,31 @@ func (rw *RouteWatch) Run(ctx context.Context) error {
// Non-batched handlers not implemented yet
rw.logger.Error("Non-batched handlers not implemented")
return fmt.Errorf("non-batched handlers not implemented")
return errNonBatchedNotImplemented
}
// No longer need routing table handler - PrefixHandler maintains live_routes table
// Start streaming
if err := rw.streamer.Start(); err != nil {
err := rw.streamer.Start(ctx)
if err != nil {
return err
}
// Start HTTP server
if err := rw.server.Start(); err != nil {
err = rw.server.Start(ctx)
if err != nil {
return err
}
// Start ASN WHOIS fetcher for background updates
rw.asnFetcher = NewASNFetcher(rw.db, rw.logger.Logger)
rw.asnFetcher.Start()
rw.asnFetcher.Start(ctx)
rw.server.SetASNFetcher(rw.asnFetcher)
// Start database maintenance goroutine
rw.dbMaintainer = NewDBMaintainer(rw.db, rw.logger.Logger)
rw.dbMaintainer.Start()
rw.dbMaintainer.Start(ctx)
// Wait for context cancellation
<-ctx.Done()
@@ -127,13 +138,14 @@ func (rw *RouteWatch) Run(ctx context.Context) error {
}
// Shutdown performs graceful shutdown of all services
func (rw *RouteWatch) Shutdown() {
func (rw *RouteWatch) Shutdown(ctx context.Context) {
rw.mu.Lock()
if rw.shutdown {
rw.mu.Unlock()
return
}
rw.shutdown = true
rw.mu.Unlock()
@@ -142,14 +154,17 @@ func (rw *RouteWatch) Shutdown() {
rw.logger.Info("Flushing database handler")
rw.dbHandler.Stop()
}
if rw.peerHandler != nil {
rw.logger.Info("Flushing peer handler")
rw.peerHandler.Stop()
}
if rw.prefixHandler != nil {
rw.logger.Info("Flushing prefix handler")
rw.prefixHandler.Stop()
}
if rw.peeringHandler != nil {
rw.logger.Info("Flushing peering handler")
rw.peeringHandler.Stop()
@@ -168,11 +183,15 @@ func (rw *RouteWatch) Shutdown() {
// Stop services
rw.streamer.Stop()
// Stop HTTP server with a timeout
// Stop HTTP server with a timeout of its own, however much of ctx is left
const serverStopTimeout = 5 * time.Second
stopCtx, cancel := context.WithTimeout(context.Background(), serverStopTimeout)
stopCtx, cancel := context.WithTimeout(
context.WithoutCancel(ctx), serverStopTimeout)
defer cancel()
if err := rw.server.Stop(stopCtx); err != nil {
err := rw.server.Stop(stopCtx)
if err != nil {
rw.logger.Error("Failed to stop HTTP server gracefully", "error", err)
}
@@ -185,12 +204,11 @@ func (rw *RouteWatch) Shutdown() {
"bits_per_sec", metrics.BitsPerSec,
"duration", time.Since(metrics.ConnectedSince),
)
}
// getModule provides all fx dependencies
func getModule() fx.Option {
return fx.Options(
func getModule() []fx.Option {
return []fx.Option{
fx.Provide(
logger.New,
config.New,
@@ -203,5 +221,5 @@ func getModule() fx.Option {
server.New,
New,
),
)
}
}
internal/routewatch/app_integration_test.go → internal/routewatch/app_integration_internal_test.go
+74 -32
View File
@@ -50,7 +50,9 @@ func newMockStore() *mockStore {
}
// GetOrCreateASN mock implementation
func (m *mockStore) GetOrCreateASN(number int, timestamp time.Time) (*database.ASN, error) {
func (m *mockStore) GetOrCreateASN(
number int, timestamp time.Time,
) (*database.ASN, error) {
m.mu.Lock()
defer m.mu.Unlock()
@@ -99,11 +101,14 @@ func (m *mockStore) UpdatePrefixesBatch(prefixes map[string]time.Time) error {
}
}
}
return nil
}
// GetOrCreatePrefix mock implementation
func (m *mockStore) GetOrCreatePrefix(prefix string, timestamp time.Time) (*database.Prefix, error) {
func (m *mockStore) GetOrCreatePrefix(
prefix string, timestamp time.Time,
) (*database.Prefix, error) {
m.mu.Lock()
defer m.mu.Unlock()
@@ -168,7 +173,7 @@ func (m *mockStore) RecordPeering(asA, asB int, _ time.Time) error {
}
// UpdatePeer mock implementation
func (m *mockStore) UpdatePeer(peerIP string, peerASN int, messageType string, timestamp time.Time) error {
func (m *mockStore) UpdatePeer(_ string, _ int, _ string, _ time.Time) error {
// Simple mock - just return nil
return nil
}
@@ -194,41 +199,45 @@ func (m *mockStore) GetStats() (database.Stats, error) {
}
// GetStatsContext returns statistics about the mock store with context support
func (m *mockStore) GetStatsContext(ctx context.Context) (database.Stats, error) {
func (m *mockStore) GetStatsContext(_ context.Context) (database.Stats, error) {
return m.GetStats()
}
// UpsertLiveRoute mock implementation
func (m *mockStore) UpsertLiveRoute(route *database.LiveRoute) error {
func (m *mockStore) UpsertLiveRoute(_ *database.LiveRoute) error {
// Simple mock - just return nil
return nil
}
// DeleteLiveRoute mock implementation
func (m *mockStore) DeleteLiveRoute(prefix, peerIP string) error {
func (m *mockStore) DeleteLiveRoute(_, _ string) error {
// Simple mock - just return nil
return nil
}
// GetPrefixDistribution mock implementation
func (m *mockStore) GetPrefixDistribution() (ipv4 []database.PrefixDistribution, ipv6 []database.PrefixDistribution, err error) {
func (m *mockStore) GetPrefixDistribution() (
[]database.PrefixDistribution, []database.PrefixDistribution, error,
) {
// Return empty distributions for now
return nil, nil, nil
}
// GetPrefixDistributionContext mock implementation with context support
func (m *mockStore) GetPrefixDistributionContext(ctx context.Context) (ipv4 []database.PrefixDistribution, ipv6 []database.PrefixDistribution, err error) {
func (m *mockStore) GetPrefixDistributionContext(_ context.Context) (
[]database.PrefixDistribution, []database.PrefixDistribution, error,
) {
return m.GetPrefixDistribution()
}
// GetLiveRouteCounts mock implementation
func (m *mockStore) GetLiveRouteCounts() (ipv4Count, ipv6Count int, err error) {
func (m *mockStore) GetLiveRouteCounts() (int, int, error) {
// Return mock counts
return m.RouteCount / 2, m.RouteCount / 2, nil
}
// GetLiveRouteCountsContext mock implementation with context support
func (m *mockStore) GetLiveRouteCountsContext(ctx context.Context) (ipv4Count, ipv6Count int, err error) {
func (m *mockStore) GetLiveRouteCountsContext(_ context.Context) (int, int, error) {
return m.GetLiveRouteCounts()
}
@@ -247,41 +256,51 @@ func (m *mockStore) GetASDetails(asn int) (*database.ASN, []database.LiveRoute,
}
// GetASDetailsContext mock implementation with context support
func (m *mockStore) GetASDetailsContext(ctx context.Context, asn int) (*database.ASN, []database.LiveRoute, error) {
func (m *mockStore) GetASDetailsContext(
_ context.Context, asn int,
) (*database.ASN, []database.LiveRoute, error) {
return m.GetASDetails(asn)
}
// GetPrefixDetails mock implementation
func (m *mockStore) GetPrefixDetails(prefix string) ([]database.LiveRoute, map[int]database.ASN, error) {
func (m *mockStore) GetPrefixDetails(
_ string,
) ([]database.LiveRoute, map[int]database.ASN, error) {
// Return empty routes for now
return []database.LiveRoute{}, map[int]database.ASN{}, nil
}
// GetPrefixDetailsContext mock implementation with context support
func (m *mockStore) GetPrefixDetailsContext(
ctx context.Context, prefix string,
_ context.Context, prefix string,
) ([]database.LiveRoute, map[int]database.ASN, error) {
return m.GetPrefixDetails(prefix)
}
func (m *mockStore) GetRandomPrefixesByLength(maskLength, ipVersion, limit int) ([]database.LiveRoute, error) {
func (m *mockStore) GetRandomPrefixesByLength(
_, _, _ int,
) ([]database.LiveRoute, error) {
// Return empty routes for now
return []database.LiveRoute{}, nil
}
// GetRandomPrefixesByLengthContext mock implementation with context support
func (m *mockStore) GetRandomPrefixesByLengthContext(ctx context.Context, maskLength, ipVersion, limit int) ([]database.LiveRoute, error) {
func (m *mockStore) GetRandomPrefixesByLengthContext(
_ context.Context, maskLength, ipVersion, limit int,
) ([]database.LiveRoute, error) {
return m.GetRandomPrefixesByLength(maskLength, ipVersion, limit)
}
// GetASPeers mock implementation
func (m *mockStore) GetASPeers(asn int) ([]database.ASPeer, error) {
func (m *mockStore) GetASPeers(_ int) ([]database.ASPeer, error) {
// Return empty peers for now
return []database.ASPeer{}, nil
}
// GetASPeersContext mock implementation with context support
func (m *mockStore) GetASPeersContext(ctx context.Context, asn int) ([]database.ASPeer, error) {
func (m *mockStore) GetASPeersContext(
_ context.Context, asn int,
) ([]database.ASPeer, error) {
return m.GetASPeers(asn)
}
@@ -291,8 +310,11 @@ func (m *mockStore) GetIPInfo(ip string) (*database.IPInfo, error) {
}
// GetIPInfoContext mock implementation with context support
func (m *mockStore) GetIPInfoContext(ctx context.Context, ip string) (*database.IPInfo, error) {
func (m *mockStore) GetIPInfoContext(
_ context.Context, ip string,
) (*database.IPInfo, error) {
now := time.Now()
return &database.IPInfo{
IP: ip,
Netblock: "8.8.8.0/24",
@@ -309,17 +331,21 @@ func (m *mockStore) GetIPInfoContext(ctx context.Context, ip string) (*database.
}
// GetNextStaleASN mock implementation
func (m *mockStore) GetNextStaleASN(ctx context.Context, staleThreshold time.Duration) (int, error) {
func (m *mockStore) GetNextStaleASN(_ context.Context, _ time.Duration) (int, error) {
return 0, database.ErrNoStaleASN
}
// UpdateASNWHOIS mock implementation
func (m *mockStore) UpdateASNWHOIS(ctx context.Context, update *database.ASNWHOISUpdate) error {
func (m *mockStore) UpdateASNWHOIS(
_ context.Context, _ *database.ASNWHOISUpdate,
) error {
return nil
}
// GetWHOISStats mock implementation
func (m *mockStore) GetWHOISStats(ctx context.Context, staleThreshold time.Duration) (*database.WHOISStats, error) {
func (m *mockStore) GetWHOISStats(
_ context.Context, _ time.Duration,
) (*database.WHOISStats, error) {
m.mu.Lock()
defer m.mu.Unlock()
@@ -347,25 +373,28 @@ func (m *mockStore) UpsertLiveRouteBatch(routes []*database.LiveRoute) error {
LastSeen: route.LastUpdated,
}
m.PrefixCount++
if route.IPVersion == 4 {
m.IPv4Prefixes++
} else {
m.IPv6Prefixes++
}
}
m.RouteCount++
}
return nil
}
// DeleteLiveRouteBatch mock implementation
func (m *mockStore) DeleteLiveRouteBatch(deletions []database.LiveRouteDeletion) error {
func (m *mockStore) DeleteLiveRouteBatch(_ []database.LiveRouteDeletion) error {
// Simple mock - just return nil
return nil
}
// DeleteLiveRoutesOfPeer mock implementation
func (m *mockStore) DeleteLiveRoutesOfPeer(peerIP string) (int, error) {
func (m *mockStore) DeleteLiveRoutesOfPeer(_ string) (int, error) {
// Simple mock - deletes nothing
return 0, nil
}
@@ -385,43 +414,53 @@ func (m *mockStore) GetOrCreateASNBatch(asns map[int]time.Time) error {
m.ASNCount++
}
}
return nil
}
// UpdatePeerBatch mock implementation
func (m *mockStore) UpdatePeerBatch(peers map[string]database.PeerUpdate) error {
func (m *mockStore) UpdatePeerBatch(_ map[string]database.PeerUpdate) error {
// Simple mock - just return nil
return nil
}
// Vacuum mock implementation
func (m *mockStore) Vacuum(ctx context.Context) error {
func (m *mockStore) Vacuum(_ context.Context) error {
return nil
}
// Analyze mock implementation
func (m *mockStore) Analyze(ctx context.Context) error {
func (m *mockStore) Analyze(_ context.Context) error {
return nil
}
// Checkpoint mock implementation
func (m *mockStore) Checkpoint(ctx context.Context) error {
func (m *mockStore) Checkpoint(_ context.Context) error {
return nil
}
// Ping mock implementation
func (m *mockStore) Ping(ctx context.Context) error {
func (m *mockStore) Ping(_ context.Context) error {
return nil
}
func TestRouteWatchLiveFeed(t *testing.T) {
t.Parallel()
if testing.Short() {
t.Skip("skipping live RIPE RIS network feed test in short mode; run without -short to include it")
t.Skip("skipping live RIPE RIS network feed test in short mode; " +
"run without -short to include it")
}
// Create mock database
mockDB := newMockStore()
defer mockDB.Close()
defer func() {
err := mockDB.Close()
if err != nil {
t.Errorf("failed to close mock database: %v", err)
}
}()
logger := logger.New()
@@ -474,16 +513,19 @@ func TestRouteWatchLiveFeed(t *testing.T) {
if stats.ASNs == 0 {
t.Error("Expected to receive some ASNs from live feed")
}
t.Logf("Received %d unique ASNs in 5 seconds", stats.ASNs)
if stats.Prefixes == 0 {
t.Error("Expected to receive some prefixes from live feed")
}
t.Logf("Received %d unique prefixes (%d IPv4, %d IPv6) in 5 seconds", stats.Prefixes, stats.IPv4Prefixes, stats.IPv6Prefixes)
t.Logf("Received %d unique prefixes (%d IPv4, %d IPv6) in 5 seconds",
stats.Prefixes, stats.IPv4Prefixes, stats.IPv6Prefixes)
if stats.Peerings == 0 {
t.Error("Expected to receive some peerings from live feed")
}
t.Logf("Recorded %d AS peering relationships in 5 seconds", stats.Peerings)
t.Logf("Recorded %d AS peering relationships in 5 seconds", stats.Peerings)
}
internal/routewatch/app_test.go
+2 -2
View File
@@ -1,3 +1,3 @@
package routewatch
package routewatch_test
// Tests for routewatch package are in app_integration_test.go
// Tests for routewatch package are in app_integration_internal_test.go
internal/routewatch/ashandler.go
+22 -16
View File
@@ -41,9 +41,11 @@ type ASHandler struct {
// Batching. mu guards batch only: the flush goroutine takes the batch
// under mu and writes it after releasing mu, so HandleMessage waits for
// a database write only while batch is at asnBatchMaxSize.
mu sync.Mutex
batch []asnOp
batchTaken sync.Cond // signalled when the flush goroutine takes batch; its L is &mu
mu sync.Mutex
batch []asnOp
// batchTaken is signalled when the flush goroutine takes batch; its L is
// &mu.
batchTaken sync.Cond
batchFullCh chan struct{} // wakes the flush goroutine when batch is full
lastFlush time.Time // used only by the flush goroutine
stopCh chan struct{}
@@ -83,7 +85,7 @@ func NewASHandler(db database.Store, logger *logger.Logger) *ASHandler {
// the AS path information needed to track autonomous systems.
func (h *ASHandler) WantsMessage(messageType string) bool {
// We only care about UPDATE messages for the database
return messageType == "UPDATE"
return messageType == ristypes.MessageTypeUpdate
}
// QueueCapacity returns the recommended message queue size for this handler.
@@ -148,10 +150,21 @@ func (h *ASHandler) HandleMessage(msg *ristypes.RISMessage) {
}
}
// Stop gracefully shuts down the ASHandler by signaling the background flush
// goroutine to terminate and waiting for it to complete. Any pending ASN
// operations in the current batch are flushed to the database before Stop
// returns. This method should be called during application shutdown to ensure
// no data is lost.
func (h *ASHandler) Stop() {
close(h.stopCh)
h.wg.Wait()
}
// flushLoop runs in a goroutine, the only one that writes the batch: it
// flushes when the batch is full, periodically, and once more on Stop.
func (h *ASHandler) flushLoop() {
defer h.wg.Done()
ticker := time.NewTicker(asnBatchTimeout)
defer ticker.Stop()
@@ -181,6 +194,7 @@ func (h *ASHandler) takeBatch() []asnOp {
if len(h.batch) == 0 {
return nil
}
batch := h.batch
h.batch = make([]asnOp, 0, asnBatchSize)
h.batchTaken.Broadcast()
@@ -205,19 +219,11 @@ func (h *ASHandler) flushBatch() {
}
// Process all ASNs in a single batch transaction
if err := h.db.GetOrCreateASNBatch(asnMap); err != nil {
h.logger.Error("Failed to process ASN batch", "error", err, "count", len(asnMap))
err := h.db.GetOrCreateASNBatch(asnMap)
if err != nil {
h.logger.Error("Failed to process ASN batch",
"error", err, "count", len(asnMap))
}
h.lastFlush = time.Now()
}
// Stop gracefully shuts down the ASHandler by signaling the background flush
// goroutine to terminate and waiting for it to complete. Any pending ASN
// operations in the current batch are flushed to the database before Stop
// returns. This method should be called during application shutdown to ensure
// no data is lost.
func (h *ASHandler) Stop() {
close(h.stopCh)
h.wg.Wait()
}
internal/routewatch/asnfetcher.go
+21 -24
View File
@@ -1,8 +1,8 @@
// Package routewatch contains the ASN WHOIS fetcher for background updates.
package routewatch
import (
"context"
"errors"
"log/slog"
"sync"
"time"
@@ -73,10 +73,13 @@ func NewASNFetcher(db database.Store, logger *slog.Logger) *ASNFetcher {
}
}
// Start begins the background ASN fetcher goroutine.
func (f *ASNFetcher) Start() {
// Start begins the background ASN fetcher goroutine. It runs until Stop:
// the end of ctx does not stop it.
func (f *ASNFetcher) Start(ctx context.Context) {
f.wg.Add(1)
go f.run()
go f.run(context.WithoutCancel(ctx))
f.logger.Info("ASN fetcher started",
"base_interval", baseInterval,
"min_interval", minInterval,
@@ -151,10 +154,7 @@ func (f *ASNFetcher) recordSuccess() {
f.consecutiveFails = 0
// Decrease interval by half, but not below minimum
newInterval := f.currentInterval / backoffMultiplier
if newInterval < minInterval {
newInterval = minInterval
}
newInterval := max(f.currentInterval/backoffMultiplier, minInterval)
if newInterval != f.currentInterval {
f.logger.Debug("Decreased fetch interval",
@@ -177,10 +177,7 @@ func (f *ASNFetcher) recordFailure() {
f.consecutiveFails++
// Exponential backoff: multiply by 2, capped at max
newInterval := f.currentInterval * backoffMultiplier
if newInterval > maxInterval {
newInterval = maxInterval
}
newInterval := min(f.currentInterval*backoffMultiplier, maxInterval)
if newInterval != f.currentInterval {
f.logger.Debug("Increased fetch interval due to failure",
@@ -199,7 +196,7 @@ func (f *ASNFetcher) recordFailure() {
}
// run is the main background loop.
func (f *ASNFetcher) run() {
func (f *ASNFetcher) run(ctx context.Context) {
defer f.wg.Done()
timer := time.NewTimer(f.getInterval())
@@ -212,13 +209,13 @@ func (f *ASNFetcher) run() {
case asn := <-f.immediateQueue:
// Process immediate request (respects lock)
f.tryFetch(asn)
f.tryFetch(ctx, asn)
// Reset timer after immediate fetch
timer.Reset(f.getInterval())
case <-timer.C:
// Background fetch of stale/missing ASN
f.fetchNextStale()
f.fetchNextStale(ctx)
// Reset timer with potentially updated interval
timer.Reset(f.getInterval())
}
@@ -227,7 +224,7 @@ func (f *ASNFetcher) run() {
// tryFetch attempts to fetch and update an ASN, respecting the fetch lock.
// Returns true if fetch was successful.
func (f *ASNFetcher) tryFetch(asn int) bool {
func (f *ASNFetcher) tryFetch(ctx context.Context, asn int) bool {
// Try to acquire lock, skip if another fetch is running
if !f.fetchMu.TryLock() {
f.logger.Debug("Skipping fetch, another fetch in progress", "asn", asn)
@@ -236,11 +233,11 @@ func (f *ASNFetcher) tryFetch(asn int) bool {
}
defer f.fetchMu.Unlock()
return f.fetchAndUpdate(asn)
return f.fetchAndUpdate(ctx, asn)
}
// fetchNextStale finds and fetches the next ASN needing WHOIS data.
func (f *ASNFetcher) fetchNextStale() {
func (f *ASNFetcher) fetchNextStale(ctx context.Context) {
// Try to acquire lock, skip if another fetch is running
if !f.fetchMu.TryLock() {
f.logger.Debug("Skipping stale fetch, another fetch in progress")
@@ -249,12 +246,12 @@ func (f *ASNFetcher) fetchNextStale() {
}
defer f.fetchMu.Unlock()
ctx, cancel := context.WithTimeout(context.Background(), time.Minute)
lookupCtx, cancel := context.WithTimeout(ctx, time.Minute)
defer cancel()
asn, err := f.db.GetNextStaleASN(ctx, whoisStaleThreshold)
asn, err := f.db.GetNextStaleASN(lookupCtx, whoisStaleThreshold)
if err != nil {
if err != database.ErrNoStaleASN {
if !errors.Is(err, database.ErrNoStaleASN) {
f.logger.Error("Failed to get stale ASN", "error", err)
f.recordFailure()
}
@@ -263,13 +260,13 @@ func (f *ASNFetcher) fetchNextStale() {
return
}
f.fetchAndUpdate(asn)
f.fetchAndUpdate(ctx, asn)
}
// fetchAndUpdate performs a WHOIS lookup and updates the database.
// Returns true if successful.
func (f *ASNFetcher) fetchAndUpdate(asn int) bool {
ctx, cancel := context.WithTimeout(context.Background(), time.Minute)
func (f *ASNFetcher) fetchAndUpdate(ctx context.Context, asn int) bool {
ctx, cancel := context.WithTimeout(ctx, time.Minute)
defer cancel()
f.logger.Info("Fetching WHOIS data", "asn", asn)
internal/routewatch/batchwrite_test.go → internal/routewatch/batchwrite_internal_test.go
+97 -46
View File
@@ -49,21 +49,11 @@ func newBlockingStore() *blockingStore {
}
}
func (s *blockingStore) waitForRelease() {
s.startOnce.Do(func() { close(s.writeStarted) })
<-s.release
}
// releaseWrites lets the held writes, and all later ones, finish. A second call
// does nothing.
func (s *blockingStore) releaseWrites() {
s.releaseOnce.Do(func() { close(s.release) })
}
func (s *blockingStore) UpsertLiveRouteBatch(routes []*database.LiveRoute) error {
s.waitForRelease()
s.mu.Lock()
defer s.mu.Unlock()
for _, route := range routes {
s.prefixes[route.Prefix] = true
}
@@ -75,6 +65,7 @@ func (s *blockingStore) GetOrCreateASNBatch(asns map[int]time.Time) error {
s.waitForRelease()
s.mu.Lock()
defer s.mu.Unlock()
for asn := range asns {
s.asns[asn] = true
}
@@ -86,6 +77,7 @@ func (s *blockingStore) UpdatePeerBatch(peers map[string]database.PeerUpdate) er
s.waitForRelease()
s.mu.Lock()
defer s.mu.Unlock()
for peerIP := range peers {
s.peers[peerIP] = true
}
@@ -93,9 +85,21 @@ func (s *blockingStore) UpdatePeerBatch(peers map[string]database.PeerUpdate) er
return nil
}
func (s *blockingStore) waitForRelease() {
s.startOnce.Do(func() { close(s.writeStarted) })
<-s.release
}
// releaseWrites lets the held writes, and all later ones, finish. A second call
// does nothing.
func (s *blockingStore) releaseWrites() {
s.releaseOnce.Do(func() { close(s.release) })
}
// returnsInTime reports whether f returns within handleMessageTimeout.
func returnsInTime(f func()) bool {
done := make(chan struct{})
go func() {
f()
close(done)
@@ -115,7 +119,9 @@ func returnsInTime(f func()) bool {
func startBlockedWrite(
t *testing.T,
store *blockingStore,
h interface{ HandleMessage(*ristypes.RISMessage) },
h interface {
HandleMessage(msg *ristypes.RISMessage)
},
fill []*ristypes.RISMessage,
) {
t.Helper()
@@ -125,7 +131,8 @@ func startBlockedWrite(
h.HandleMessage(msg)
}
}) {
t.Fatal("HandleMessage did not return after filling the batch: it waits for the database write")
t.Fatal("HandleMessage did not return after filling the batch: " +
"it waits for the database write")
}
select {
@@ -143,13 +150,14 @@ func handleWhileAWriteIsBlocked(
t *testing.T,
store *blockingStore,
h interface {
HandleMessage(*ristypes.RISMessage)
HandleMessage(msg *ristypes.RISMessage)
Stop()
},
fill []*ristypes.RISMessage,
extra *ristypes.RISMessage,
) {
t.Helper()
defer h.Stop()
defer store.releaseWrites()
@@ -169,22 +177,25 @@ func handleAtMaxBatchSize(
t *testing.T,
store *blockingStore,
h interface {
HandleMessage(*ristypes.RISMessage)
HandleMessage(msg *ristypes.RISMessage)
Stop()
},
fill, toMax []*ristypes.RISMessage,
extra *ristypes.RISMessage,
) {
t.Helper()
defer h.Stop()
defer store.releaseWrites()
startBlockedWrite(t, store, h, fill)
for _, msg := range toMax {
h.HandleMessage(msg)
}
returned := make(chan struct{})
go func() {
h.HandleMessage(extra)
close(returned)
@@ -192,11 +203,13 @@ func handleAtMaxBatchSize(
select {
case <-returned:
t.Fatal("HandleMessage returned while the batch was at its maximum size and a write was blocked")
t.Fatal("HandleMessage returned while the batch was at its maximum " +
"size and a write was blocked")
case <-time.After(stillWaitingAfter):
}
store.releaseWrites()
select {
case <-returned:
case <-time.After(handleMessageTimeout):
@@ -208,46 +221,57 @@ func handleAtMaxBatchSize(
// with origin AS testASNA.
func prefixAnnouncement(prefixes []string) *ristypes.RISMessage {
return &ristypes.RISMessage{
Peer: testPeerIP,
Path: ristypes.ASPath{testASNA},
Announcements: []ristypes.RISAnnouncement{{NextHop: testPeerIP, Prefixes: prefixes}},
Peer: testPeerIP,
Path: ristypes.ASPath{testASNA},
Announcements: []ristypes.RISAnnouncement{
{NextHop: testPeerIP, Prefixes: prefixes},
},
}
}
func TestPrefixHandlerHandlesMessagesWhileAWriteIsBlocked(t *testing.T) {
const (
fillPrefix = "192.0.2.0/24"
extraPrefix = "198.51.100.0/24"
)
t.Parallel()
const extraPrefix = "198.51.100.0/24"
store := newBlockingStore()
h := NewPrefixHandler(store, logger.New())
fill := slices.Repeat([]string{testV4Prefix}, prefixBatchSize)
handleWhileAWriteIsBlocked(t, store, h,
[]*ristypes.RISMessage{prefixAnnouncement(slices.Repeat([]string{fillPrefix}, prefixBatchSize))},
[]*ristypes.RISMessage{prefixAnnouncement(fill)},
prefixAnnouncement([]string{extraPrefix}))
store.mu.Lock()
defer store.mu.Unlock()
if !store.prefixes[extraPrefix] {
t.Errorf("Stop did not write %s, handled while a write was blocked", extraPrefix)
t.Errorf("Stop did not write %s, handled while a write was blocked",
extraPrefix)
}
}
func TestASHandlerHandlesMessagesWhileAWriteIsBlocked(t *testing.T) {
t.Parallel()
store := newBlockingStore()
h := NewASHandler(store, logger.New())
fill := slices.Repeat(ristypes.ASPath{testASNA}, asnBatchSize)
handleWhileAWriteIsBlocked(t, store, h,
[]*ristypes.RISMessage{{Path: slices.Repeat(ristypes.ASPath{testASNA}, asnBatchSize)}},
[]*ristypes.RISMessage{{Path: fill}},
&ristypes.RISMessage{Path: ristypes.ASPath{testASNB}})
store.mu.Lock()
defer store.mu.Unlock()
if !store.asns[testASNB] {
t.Errorf("Stop did not write AS%d, handled while a write was blocked", testASNB)
t.Errorf("Stop did not write AS%d, handled while a write was blocked",
testASNB)
}
}
func TestPeerHandlerHandlesMessagesWhileAWriteIsBlocked(t *testing.T) {
t.Parallel()
const otherPeerIP = "2001:db8:ffff::2"
store := newBlockingStore()
@@ -258,60 +282,76 @@ func TestPeerHandlerHandlesMessagesWhileAWriteIsBlocked(t *testing.T) {
store.mu.Lock()
defer store.mu.Unlock()
if !store.peers[otherPeerIP] {
t.Errorf("Stop did not write peer %s, handled while a write was blocked", otherPeerIP)
t.Errorf("Stop did not write peer %s, handled while a write was blocked",
otherPeerIP)
}
}
func TestPrefixHandlerWaitsAtTheMaximumBatchSize(t *testing.T) {
const (
fillPrefix = "192.0.2.0/24"
extraPrefix = "198.51.100.0/24"
)
t.Parallel()
const extraPrefix = "198.51.100.0/24"
store := newBlockingStore()
h := NewPrefixHandler(store, logger.New())
fill := slices.Repeat([]string{testV4Prefix}, prefixBatchSize)
toMax := slices.Repeat([]string{testV4Prefix}, prefixBatchMaxSize)
handleAtMaxBatchSize(t, store, h,
[]*ristypes.RISMessage{prefixAnnouncement(slices.Repeat([]string{fillPrefix}, prefixBatchSize))},
[]*ristypes.RISMessage{prefixAnnouncement(slices.Repeat([]string{fillPrefix}, prefixBatchMaxSize))},
[]*ristypes.RISMessage{prefixAnnouncement(fill)},
[]*ristypes.RISMessage{prefixAnnouncement(toMax)},
prefixAnnouncement([]string{extraPrefix}))
store.mu.Lock()
defer store.mu.Unlock()
if !store.prefixes[extraPrefix] {
t.Errorf("Stop did not write %s, handled after the batch reached its maximum size", extraPrefix)
t.Errorf("Stop did not write %s, handled after the batch reached its "+
"maximum size", extraPrefix)
}
}
func TestASHandlerWaitsAtTheMaximumBatchSize(t *testing.T) {
t.Parallel()
store := newBlockingStore()
h := NewASHandler(store, logger.New())
fill := slices.Repeat(ristypes.ASPath{testASNA}, asnBatchSize)
toMax := slices.Repeat(ristypes.ASPath{testASNA}, asnBatchMaxSize)
handleAtMaxBatchSize(t, store, h,
[]*ristypes.RISMessage{{Path: slices.Repeat(ristypes.ASPath{testASNA}, asnBatchSize)}},
[]*ristypes.RISMessage{{Path: slices.Repeat(ristypes.ASPath{testASNA}, asnBatchMaxSize)}},
[]*ristypes.RISMessage{{Path: fill}},
[]*ristypes.RISMessage{{Path: toMax}},
&ristypes.RISMessage{Path: ristypes.ASPath{testASNB}})
store.mu.Lock()
defer store.mu.Unlock()
if !store.asns[testASNB] {
t.Errorf("Stop did not write AS%d, handled after the batch reached its maximum size", testASNB)
t.Errorf("Stop did not write AS%d, handled after the batch reached its "+
"maximum size", testASNB)
}
}
func TestPeerHandlerWaitsAtTheMaximumBatchSize(t *testing.T) {
t.Parallel()
const otherPeerIP = "2001:db8:ffff::2"
store := newBlockingStore()
h := NewPeerHandler(store, logger.New())
message := []*ristypes.RISMessage{{Peer: testPeerIP}}
handleAtMaxBatchSize(t, store, h,
slices.Repeat([]*ristypes.RISMessage{{Peer: testPeerIP}}, peerBatchSize),
slices.Repeat([]*ristypes.RISMessage{{Peer: testPeerIP}}, peerBatchMaxSize),
slices.Repeat(message, peerBatchSize),
slices.Repeat(message, peerBatchMaxSize),
&ristypes.RISMessage{Peer: otherPeerIP})
store.mu.Lock()
defer store.mu.Unlock()
if !store.peers[otherPeerIP] {
t.Errorf("Stop did not write peer %s, handled after the batch reached its maximum size", otherPeerIP)
t.Errorf("Stop did not write peer %s, handled after the batch reached "+
"its maximum size", otherPeerIP)
}
}
@@ -320,6 +360,8 @@ func TestPeerHandlerWaitsAtTheMaximumBatchSize(t *testing.T) {
// of a full batch, and a second one exactly a full batch. Only the full batch
// leaves a wake-up in batchFullCh.
func TestHandlersWakeTheFlushGoroutineWhenTheBatchIsFull(t *testing.T) {
t.Parallel()
tests := []struct {
name string
batchSize int
@@ -331,7 +373,8 @@ func TestHandlersWakeTheFlushGoroutineWhenTheBatchIsFull(t *testing.T) {
batchSize: prefixBatchSize,
handle: func(n int) chan struct{} {
h := &PrefixHandler{batchFullCh: make(chan struct{}, 1)}
h.HandleMessage(prefixAnnouncement(slices.Repeat([]string{"192.0.2.0/24"}, n)))
h.HandleMessage(prefixAnnouncement(
slices.Repeat([]string{testV4Prefix}, n)))
return h.batchFullCh
},
@@ -342,7 +385,9 @@ func TestHandlersWakeTheFlushGoroutineWhenTheBatchIsFull(t *testing.T) {
handle: func(n int) chan struct{} {
h := &ASHandler{batchFullCh: make(chan struct{}, 1)}
// The origin AS and each AS in the path are one update each.
h.HandleMessage(&ristypes.RISMessage{Path: slices.Repeat(ristypes.ASPath{testASNA}, n-1)})
h.HandleMessage(&ristypes.RISMessage{
Path: slices.Repeat(ristypes.ASPath{testASNA}, n-1),
})
return h.batchFullCh
},
@@ -363,10 +408,16 @@ func TestHandlersWakeTheFlushGoroutineWhenTheBatchIsFull(t *testing.T) {
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
if got := len(tt.handle(tt.batchSize - 1)); got != 0 {
t.Errorf("one update short of a full batch left %d wake-ups, want none", got)
t.Parallel()
got := len(tt.handle(tt.batchSize - 1))
if got != 0 {
t.Errorf("one update short of a full batch left %d wake-ups, "+
"want none", got)
}
if got := len(tt.handle(tt.batchSize)); got != 1 {
got = len(tt.handle(tt.batchSize))
if got != 1 {
t.Errorf("a full batch left %d wake-ups, want 1", got)
}
})
internal/routewatch/cli.go
+22 -11
View File
@@ -57,14 +57,19 @@ func logDebugStats(logger *logger.Logger) {
// CLIEntry is the main entry point for the routewatch command-line interface.
// It initializes the application using the fx dependency injection framework,
// sets up signal handling for graceful shutdown, and starts the RouteWatch service.
// This function blocks until the application receives a shutdown signal or encounters
// a fatal error.
// sets up signal handling for graceful shutdown, and starts the RouteWatch
// service. This function blocks until the application receives a shutdown
// signal or encounters a fatal error.
func CLIEntry() {
app := fx.New(
getModule(),
fx.Options(getModule()...),
fx.StopTimeout(shutdownTimeout), // Allow 60 seconds for graceful shutdown
fx.Invoke(func(lc fx.Lifecycle, rw *RouteWatch, logger *logger.Logger, shutdowner fx.Shutdowner) {
fx.Invoke(func(
lc fx.Lifecycle,
rw *RouteWatch,
logger *logger.Logger,
shutdowner fx.Shutdowner,
) {
lc.Append(fx.Hook{
OnStart: func(ctx context.Context) error {
// Start debug stats logging
@@ -77,25 +82,31 @@ func CLIEntry() {
go func() {
<-sigCh
logger.Info("Received shutdown signal")
if err := shutdowner.Shutdown(); err != nil {
err := shutdowner.Shutdown()
if err != nil {
logger.Error("Failed to shutdown gracefully", "error", err)
}
}()
go func() {
if err := rw.Run(ctx); err != nil {
err := rw.Run(ctx)
if err != nil {
logger.Error("RouteWatch error", "error", err)
if err := shutdowner.Shutdown(fx.ExitCode(failureExitCode)); err != nil {
logger.Error("Failed to shutdown gracefully", "error", err)
err = shutdowner.Shutdown(fx.ExitCode(failureExitCode))
if err != nil {
logger.Error("Failed to shutdown gracefully",
"error", err)
}
}
}()
return nil
},
OnStop: func(_ context.Context) error {
OnStop: func(ctx context.Context) error {
logger.Info("Shutting down RouteWatch")
rw.Shutdown()
rw.Shutdown(ctx)
return nil
},
internal/routewatch/dbmaintainer.go
+61 -59
View File
@@ -1,4 +1,3 @@
// Package routewatch contains the database maintainer for background maintenance tasks.
package routewatch
import (
@@ -64,10 +63,13 @@ func NewDBMaintainer(db database.Store, logger *slog.Logger) *DBMaintainer {
}
}
// Start begins the background maintenance goroutine.
func (m *DBMaintainer) Start() {
// Start begins the background maintenance goroutine. It runs until Stop:
// the end of ctx does not stop it.
func (m *DBMaintainer) Start(ctx context.Context) {
m.wg.Add(1)
go m.run()
go m.run(context.WithoutCancel(ctx))
m.logger.Info("Database maintainer started",
"checkpoint_interval", checkpointInterval,
"vacuum_interval", vacuumInterval,
@@ -83,15 +85,17 @@ func (m *DBMaintainer) Stop() {
}
// run is the main background loop.
func (m *DBMaintainer) run() {
func (m *DBMaintainer) run(ctx context.Context) {
defer m.wg.Done()
// Use different timers for each task
checkpointTimer := time.NewTimer(checkpointInterval)
vacuumTimer := time.NewTimer(vacuumInterval)
analyzeTimer := time.NewTimer(analyzeInterval)
defer checkpointTimer.Stop()
vacuumTimer := time.NewTimer(vacuumInterval)
defer vacuumTimer.Stop()
analyzeTimer := time.NewTimer(analyzeInterval)
defer analyzeTimer.Stop()
for {
@@ -100,90 +104,88 @@ func (m *DBMaintainer) run() {
return
case <-checkpointTimer.C:
m.runCheckpoint()
m.runCheckpoint(ctx)
checkpointTimer.Reset(checkpointInterval)
case <-vacuumTimer.C:
m.runVacuum()
m.runVacuum(ctx)
vacuumTimer.Reset(vacuumInterval)
case <-analyzeTimer.C:
m.runAnalyze()
m.runAnalyze(ctx)
analyzeTimer.Reset(analyzeInterval)
}
}
}
// runCheckpoint performs a WAL checkpoint to keep the WAL file small.
func (m *DBMaintainer) runCheckpoint() {
ctx, cancel := context.WithTimeout(context.Background(), checkpointTimeout)
defer cancel()
startTime := time.Now()
err := m.db.Checkpoint(ctx)
m.statsMu.Lock()
m.lastCheckpoint = time.Now()
m.lastCheckpointError = err
if err == nil {
m.checkpointCount++
}
m.statsMu.Unlock()
func (m *DBMaintainer) runCheckpoint(ctx context.Context) {
duration, err := runWithTimeout(ctx, checkpointTimeout, m.db.Checkpoint)
m.recordResult(&m.lastCheckpoint, &m.lastCheckpointError, &m.checkpointCount, err)
if err != nil {
m.logger.Error("WAL checkpoint failed", "error", err, "duration", time.Since(startTime))
m.logger.Error("WAL checkpoint failed", "error", err, "duration", duration)
} else {
m.logger.Debug("WAL checkpoint completed", "duration", time.Since(startTime))
m.logger.Debug("WAL checkpoint completed", "duration", duration)
}
}
// runVacuum performs an incremental vacuum operation on the database.
func (m *DBMaintainer) runVacuum() {
ctx, cancel := context.WithTimeout(context.Background(), vacuumTimeout)
defer cancel()
func (m *DBMaintainer) runVacuum(ctx context.Context) {
m.logger.Debug("Running incremental vacuum")
startTime := time.Now()
err := m.db.Vacuum(ctx)
m.statsMu.Lock()
m.lastVacuum = time.Now()
m.lastVacuumError = err
if err == nil {
m.vacuumCount++
}
m.statsMu.Unlock()
duration, err := runWithTimeout(ctx, vacuumTimeout, m.db.Vacuum)
m.recordResult(&m.lastVacuum, &m.lastVacuumError, &m.vacuumCount, err)
if err != nil {
m.logger.Error("Incremental vacuum failed", "error", err, "duration", time.Since(startTime))
m.logger.Error("Incremental vacuum failed", "error", err, "duration", duration)
} else {
m.logger.Debug("Incremental vacuum completed", "duration", time.Since(startTime))
m.logger.Debug("Incremental vacuum completed", "duration", duration)
}
}
// runAnalyze performs an ANALYZE operation on the database.
func (m *DBMaintainer) runAnalyze() {
ctx, cancel := context.WithTimeout(context.Background(), analyzeTimeout)
defer cancel()
func (m *DBMaintainer) runAnalyze(ctx context.Context) {
m.logger.Info("Starting database ANALYZE")
startTime := time.Now()
err := m.db.Analyze(ctx)
m.statsMu.Lock()
m.lastAnalyze = time.Now()
m.lastAnalyzeError = err
if err == nil {
m.analyzeCount++
}
m.statsMu.Unlock()
duration, err := runWithTimeout(ctx, analyzeTimeout, m.db.Analyze)
m.recordResult(&m.lastAnalyze, &m.lastAnalyzeError, &m.analyzeCount, err)
if err != nil {
m.logger.Error("ANALYZE failed", "error", err, "duration", time.Since(startTime))
m.logger.Error("ANALYZE failed", "error", err, "duration", duration)
} else {
m.logger.Info("ANALYZE completed", "duration", time.Since(startTime))
m.logger.Info("ANALYZE completed", "duration", duration)
}
}
// runWithTimeout runs operation with a context that ends after timeout and
// returns how long it took and its error.
func runWithTimeout(
ctx context.Context,
timeout time.Duration,
operation func(context.Context) error,
) (time.Duration, error) {
ctx, cancel := context.WithTimeout(ctx, timeout)
defer cancel()
startTime := time.Now()
err := operation(ctx)
return time.Since(startTime), err
}
// recordResult stores when a maintenance task last ran and its error, and
// counts it when it succeeded.
func (m *DBMaintainer) recordResult(
lastRun *time.Time, lastErr *error, successCount *int, err error,
) {
m.statsMu.Lock()
defer m.statsMu.Unlock()
*lastRun = time.Now()
*lastErr = err
if err == nil {
*successCount++
}
}
internal/routewatch/handler.go
+12 -14
View File
@@ -1,6 +1,8 @@
package routewatch
import (
"slices"
"git.eeqj.de/sneak/routewatch/internal/logger"
"git.eeqj.de/sneak/routewatch/internal/ristypes"
)
@@ -15,10 +17,11 @@ type SimpleHandler struct {
callback func(*ristypes.RISMessage)
}
// NewSimpleHandler creates a new SimpleHandler that accepts specific message types.
// The messageTypes parameter specifies which RIS message types this handler will process.
// If messageTypes is empty, the handler will accept all message types.
// The callback function is invoked for each message that passes the type filter.
// NewSimpleHandler creates a new SimpleHandler that accepts specific message
// types. The messageTypes parameter specifies which RIS message types this
// handler will process. If messageTypes is empty, the handler will accept all
// message types. The callback function is invoked for each message that
// passes the type filter.
func NewSimpleHandler(
logger *logger.Logger,
messageTypes []string,
@@ -31,22 +34,17 @@ func NewSimpleHandler(
}
}
// WantsMessage returns true if this handler wants to process messages of the given type.
// It checks whether messageType is in the handler's configured list of accepted types.
// If no specific types were configured (empty messageTypes slice), it returns true for all types.
// WantsMessage returns true if this handler wants to process messages of the
// given type. It checks whether messageType is in the handler's configured
// list of accepted types. If no specific types were configured (empty
// messageTypes slice), it returns true for all types.
func (h *SimpleHandler) WantsMessage(messageType string) bool {
// If no specific types are set, accept all messages
if len(h.messageTypes) == 0 {
return true
}
for _, t := range h.messageTypes {
if t == messageType {
return true
}
}
return false
return slices.Contains(h.messageTypes, messageType)
}
// HandleMessage processes a RIS message by invoking the configured callback function.
internal/routewatch/peerhandler.go
+25 -15
View File
@@ -1,7 +1,8 @@
package routewatch
// peerhandler.go provides batched peer tracking functionality for BGP route monitoring.
// It tracks BGP peers from incoming UPDATE messages and maintains peer state in the database.
// peerhandler.go provides batched peer tracking functionality for BGP route
// monitoring. It tracks BGP peers from incoming UPDATE messages and maintains
// peer state in the database.
import (
"strconv"
@@ -44,9 +45,11 @@ type PeerHandler struct {
// Batching. mu guards peerBatch only: the flush goroutine takes the batch
// under mu and writes it after releasing mu, so HandleMessage waits for
// a database write only while peerBatch is at peerBatchMaxSize.
mu sync.Mutex
peerBatch []peerUpdate
batchTaken sync.Cond // signalled when the flush goroutine takes peerBatch; its L is &mu
mu sync.Mutex
peerBatch []peerUpdate
// batchTaken is signalled when the flush goroutine takes peerBatch; its L
// is &mu.
batchTaken sync.Cond
batchFullCh chan struct{} // wakes the flush goroutine when peerBatch is full
lastFlush time.Time // used only by the flush goroutine
stopCh chan struct{}
@@ -104,8 +107,10 @@ func (h *PeerHandler) QueueCapacity() int {
func (h *PeerHandler) HandleMessage(msg *ristypes.RISMessage) {
// Parse peer ASN from string
peerASN := 0
if msg.PeerASN != "" {
if asn, err := strconv.Atoi(msg.PeerASN); err == nil {
asn, err := strconv.Atoi(msg.PeerASN)
if err == nil {
peerASN = asn
}
}
@@ -139,10 +144,17 @@ func (h *PeerHandler) HandleMessage(msg *ristypes.RISMessage) {
}
}
// Stop gracefully stops the handler and flushes remaining batches
func (h *PeerHandler) Stop() {
close(h.stopCh)
h.wg.Wait()
}
// flushLoop runs in a goroutine, the only one that writes the batch: it
// flushes when the batch is full, periodically, and once more on Stop.
func (h *PeerHandler) flushLoop() {
defer h.wg.Done()
ticker := time.NewTicker(peerBatchTimeout)
defer ticker.Stop()
@@ -172,6 +184,7 @@ func (h *PeerHandler) takeBatch() []peerUpdate {
if len(h.peerBatch) == 0 {
return nil
}
batch := h.peerBatch
h.peerBatch = make([]peerUpdate, 0, peerBatchSize)
h.batchTaken.Broadcast()
@@ -190,7 +203,8 @@ func (h *PeerHandler) flushBatch() {
// Deduplicate by peer IP, keeping the latest update for each peer
peerMap := make(map[string]peerUpdate)
for _, update := range batch {
if existing, ok := peerMap[update.peerIP]; !ok || update.timestamp.After(existing.timestamp) {
existing, ok := peerMap[update.peerIP]
if !ok || update.timestamp.After(existing.timestamp) {
peerMap[update.peerIP] = update
}
}
@@ -207,15 +221,11 @@ func (h *PeerHandler) flushBatch() {
}
// Process all peers in a single batch transaction
if err := h.db.UpdatePeerBatch(dbPeerMap); err != nil {
h.logger.Error("Failed to process peer batch", "error", err, "count", len(dbPeerMap))
err := h.db.UpdatePeerBatch(dbPeerMap)
if err != nil {
h.logger.Error("Failed to process peer batch",
"error", err, "count", len(dbPeerMap))
}
h.lastFlush = time.Now()
}
// Stop gracefully stops the handler and flushes remaining batches
func (h *PeerHandler) Stop() {
close(h.stopCh)
h.wg.Wait()
}
internal/routewatch/peeringhandler.go
+24 -19
View File
@@ -72,7 +72,7 @@ func NewPeeringHandler(db database.Store, logger *logger.Logger) *PeeringHandler
// the AS path information needed to extract peering relationships.
func (h *PeeringHandler) WantsMessage(messageType string) bool {
// We only care about UPDATE messages that have AS paths
return messageType == "UPDATE"
return messageType == ristypes.MessageTypeUpdate
}
// QueueCapacity returns the buffer size for the handler's message queue.
@@ -101,6 +101,7 @@ func (h *PeeringHandler) HandleMessage(msg *ristypes.RISMessage) {
return
}
key := string(pathJSON)
h.mu.Lock()
@@ -116,6 +117,23 @@ func (h *PeeringHandler) HandleMessage(msg *ristypes.RISMessage) {
h.mu.Unlock()
}
// ProcessPeeringsNow triggers immediate processing of all accumulated AS
// paths into peering records. This bypasses the normal periodic processing
// schedule and is primarily intended for testing purposes.
func (h *PeeringHandler) ProcessPeeringsNow() {
h.processPeerings()
}
// Stop gracefully shuts down the handler by signaling the background
// goroutine to stop and performing a final synchronous processing of
// any remaining AS paths. This ensures no peering data is lost during
// shutdown.
func (h *PeeringHandler) Stop() {
close(h.stopCh)
// Process any remaining peerings synchronously
h.processPeerings()
}
// processLoop runs periodically to process AS paths into peerings
func (h *PeeringHandler) processLoop() {
ticker := time.NewTicker(peeringProcessInterval)
@@ -134,13 +152,6 @@ func (h *PeeringHandler) processLoop() {
}
}
// ProcessPeeringsNow triggers immediate processing of all accumulated AS
// paths into peering records. This bypasses the normal periodic processing
// schedule and is primarily intended for testing purposes.
func (h *PeeringHandler) ProcessPeeringsNow() {
h.processPeerings()
}
// processPeerings extracts peerings from AS paths and writes to database
func (h *PeeringHandler) processPeerings() {
// Take the accumulated paths and replace the map with a fresh empty one
@@ -160,11 +171,14 @@ func (h *PeeringHandler) processPeerings() {
type peeringKey struct {
low, high int
}
peerings := make(map[peeringKey]time.Time)
for pathJSON, timestamp := range paths {
var path []int
if err := json.Unmarshal([]byte(pathJSON), &path); err != nil {
err := json.Unmarshal([]byte(pathJSON), &path)
if err != nil {
h.logger.Error("Failed to decode AS path", "error", err)
continue
@@ -197,6 +211,7 @@ func (h *PeeringHandler) processPeerings() {
// Record peerings in database
start := time.Now()
successCount := 0
for key, ts := range peerings {
err := h.db.RecordPeering(key.low, key.high, ts)
if err != nil {
@@ -218,13 +233,3 @@ func (h *PeeringHandler) processPeerings() {
"duration", time.Since(start),
)
}
// Stop gracefully shuts down the handler by signaling the background
// goroutine to stop and performing a final synchronous processing of
// any remaining AS paths. This ensures no peering data is lost during
// shutdown.
func (h *PeeringHandler) Stop() {
close(h.stopCh)
// Process any remaining peerings synchronously
h.processPeerings()
}
internal/routewatch/peeringhandler_test.go → internal/routewatch/peeringhandler_internal_test.go
+14 -2
View File
@@ -53,7 +53,8 @@ func newTestHandler(db database.Store) *PeeringHandler {
}
}
func snapshot(h *PeeringHandler) (tracked, dropped int) {
// snapshot returns how many paths the handler tracks and how many it dropped.
func snapshot(h *PeeringHandler) (int, int) {
h.mu.Lock()
defer h.mu.Unlock()
@@ -82,6 +83,8 @@ func handle(h *PeeringHandler, ts time.Time, asns ...int) {
// empties the path map (the swap) and that a path seen again after a run is
// recorded in the next run too.
func TestPeeringHandlerProcessesEachRunAndEmpties(t *testing.T) {
t.Parallel()
store := &recordingStore{mockStore: newMockStore()}
h := newTestHandler(store)
@@ -94,6 +97,7 @@ func TestPeeringHandlerProcessesEachRunAndEmpties(t *testing.T) {
if tracked, _ := snapshot(h); tracked != 0 {
t.Fatalf("map not empty after first run: %d paths remain", tracked)
}
if got := store.callCount(); got != 1 {
t.Fatalf("want 1 RecordPeering call after first run, got %d", got)
}
@@ -105,6 +109,7 @@ func TestPeeringHandlerProcessesEachRunAndEmpties(t *testing.T) {
if tracked, _ := snapshot(h); tracked != 0 {
t.Fatalf("map not empty after second run: %d paths remain", tracked)
}
if got := store.callCount(); got != 2 {
t.Fatalf("want 2 RecordPeering calls after second run, got %d", got)
}
@@ -114,6 +119,8 @@ func TestPeeringHandlerProcessesEachRunAndEmpties(t *testing.T) {
// and counts them, while a path already tracked is refreshed rather than
// dropped.
func TestPeeringHandlerCapDropsAndCounts(t *testing.T) {
t.Parallel()
store := &recordingStore{mockStore: newMockStore()}
h := newTestHandler(store)
@@ -126,6 +133,7 @@ func TestPeeringHandlerCapDropsAndCounts(t *testing.T) {
h.mu.Lock()
h.asPaths[existing] = now
for i := 0; len(h.asPaths) < maxTrackedPaths; i++ {
h.asPaths[strconv.Itoa(i)] = now
}
@@ -138,6 +146,7 @@ func TestPeeringHandlerCapDropsAndCounts(t *testing.T) {
if tracked != maxTrackedPaths {
t.Fatalf("want map size %d after drop, got %d", maxTrackedPaths, tracked)
}
if dropped != 1 {
t.Fatalf("want dropped count 1, got %d", dropped)
}
@@ -150,6 +159,7 @@ func TestPeeringHandlerCapDropsAndCounts(t *testing.T) {
if tracked != maxTrackedPaths {
t.Fatalf("want map size %d after refresh, got %d", maxTrackedPaths, tracked)
}
if dropped != 1 {
t.Fatalf("want dropped count still 1 after refresh, got %d", dropped)
}
@@ -157,7 +167,9 @@ func TestPeeringHandlerCapDropsAndCounts(t *testing.T) {
h.mu.Lock()
gotTS := h.asPaths[existing]
h.mu.Unlock()
if !gotTS.Equal(refreshed) {
t.Fatalf("existing path timestamp not refreshed: want %v, got %v", refreshed, gotTS)
t.Fatalf("existing path timestamp not refreshed: want %v, got %v",
refreshed, gotTS)
}
}
internal/routewatch/prefixhandler.go
+104 -174
View File
@@ -41,9 +41,9 @@ const (
ipv6Version = 6
)
// PrefixHandler tracks BGP prefixes and maintains a live routing table in the database.
// Routes are added on announcement and deleted on withdrawal, and all routes of
// a peer are deleted when its BGP session goes down.
// PrefixHandler tracks BGP prefixes and maintains a live routing table in the
// database. Routes are added on announcement and deleted on withdrawal, and all
// routes of a peer are deleted when its BGP session goes down.
type PrefixHandler struct {
db database.Store
logger *logger.Logger
@@ -52,10 +52,14 @@ type PrefixHandler struct {
// Batching. mu guards batch and peersDown only: the flush goroutine takes
// both under mu and writes them after releasing mu, so HandleMessage
// waits for a database write only while batch is at prefixBatchMaxSize.
mu sync.Mutex
batch []prefixUpdate
peersDown []string // peers whose BGP session went down since the batch was taken
batchTaken sync.Cond // signalled when the flush goroutine takes batch; its L is &mu
mu sync.Mutex
batch []prefixUpdate
// peersDown holds the peers whose BGP session went down since the batch
// was taken.
peersDown []string
// batchTaken is signalled when the flush goroutine takes batch; its L is
// &mu.
batchTaken sync.Cond
batchFullCh chan struct{} // wakes the flush goroutine when batch is full
lastFlush time.Time // used only by the flush goroutine
stopCh chan struct{}
@@ -95,11 +99,13 @@ func (h *PrefixHandler) SetMetricsTracker(metrics *metrics.Tracker) {
h.metrics = metrics
}
// WantsMessage returns true if this handler wants to process messages of the given type
// WantsMessage returns true if this handler wants to process messages of the
// given type
func (h *PrefixHandler) WantsMessage(messageType string) bool {
// UPDATE messages change routes; a STATE message can say a peer's
// session went down, which removes all of its routes.
return messageType == ristypes.MessageTypeUpdate || messageType == ristypes.MessageTypeState
return messageType == ristypes.MessageTypeUpdate ||
messageType == ristypes.MessageTypeState
}
// QueueCapacity returns the desired queue capacity for this handler
@@ -183,6 +189,12 @@ func (h *PrefixHandler) HandleMessage(msg *ristypes.RISMessage) {
}
}
// Stop gracefully stops the handler and flushes remaining batches
func (h *PrefixHandler) Stop() {
close(h.stopCh)
h.wg.Wait()
}
// peerSessionDown notes that the peer's BGP session went down, so that the next
// flush deletes all of the peer's live routes before it writes the batch. The
// peer's updates still in the batch came before the down, which replaces them,
@@ -192,7 +204,9 @@ func (h *PrefixHandler) peerSessionDown(peer string) {
h.mu.Lock()
defer h.mu.Unlock()
h.batch = slices.DeleteFunc(h.batch, func(update prefixUpdate) bool { return update.peer == peer })
h.batch = slices.DeleteFunc(h.batch, func(update prefixUpdate) bool {
return update.peer == peer
})
h.peersDown = append(h.peersDown, peer)
}
@@ -200,6 +214,7 @@ func (h *PrefixHandler) peerSessionDown(peer string) {
// flushes when the batch is full, periodically, and once more on Stop.
func (h *PrefixHandler) flushLoop() {
defer h.wg.Done()
ticker := time.NewTicker(prefixBatchTimeout)
defer ticker.Stop()
@@ -230,6 +245,7 @@ func (h *PrefixHandler) takeBatch() ([]prefixUpdate, []string) {
if len(h.batch) == 0 && len(h.peersDown) == 0 {
return nil, nil
}
batch, peersDown := h.batch, h.peersDown
h.batch = make([]prefixUpdate, 0, prefixBatchSize)
h.peersDown = nil
@@ -250,43 +266,34 @@ func (h *PrefixHandler) flushBatch() {
batchSize := len(batch)
// The routes of a peer whose session went down are deleted first: the
// batch holds only the updates the peer sent after the down. Most downs
// come from peers that have no routes, so only a delete that removed
// routes is logged.
for _, peer := range peersDown {
deleted, err := h.db.DeleteLiveRoutesOfPeer(peer)
if err != nil {
h.logger.Error("Failed to delete the live routes of a peer whose session went down",
"peer", peer, "error", err)
continue
}
if deleted > 0 {
h.logger.Info("Deleted the live routes of a peer whose session went down",
"peer", peer, "routes", deleted)
}
}
// batch holds only the updates the peer sent after the down.
h.deleteLiveRoutesOfPeers(peersDown)
// For each prefix and peer, keep only the last update the feed delivered.
// Updates from different peers for the same prefix are all kept. Arrival
// order decides, not the timestamp: timestamps are whole seconds, so
// updates within one second carry the same one.
type prefixAndPeer struct{ prefix, peer string }
lastUpdates := make(map[prefixAndPeer]prefixUpdate)
for _, update := range batch {
lastUpdates[prefixAndPeer{update.prefix, update.peer}] = update
}
// Collect routes to upsert and delete
var routesToUpsert []*database.LiveRoute
var routesToDelete []database.LiveRouteDeletion
var (
routesToUpsert []*database.LiveRoute
routesToDelete []database.LiveRouteDeletion
)
// Collect unique prefixes to update
prefixesToUpdate := make(map[string]time.Time)
for _, update := range lastUpdates {
// Track prefix for both announcements and withdrawals
if _, exists := prefixesToUpdate[update.prefix]; !exists || update.timestamp.After(prefixesToUpdate[update.prefix]) {
seen, exists := prefixesToUpdate[update.prefix]
if !exists || update.timestamp.After(seen) {
prefixesToUpdate[update.prefix] = update.timestamp
}
@@ -300,7 +307,8 @@ func (h *PrefixHandler) flushBatch() {
// Parse CIDR to get IP version
_, ipVersion, err := parseCIDR(update.prefix)
if err != nil {
h.logger.Error("Failed to parse CIDR for withdrawal", "prefix", update.prefix, "error", err)
h.logger.Error("Failed to parse CIDR for withdrawal",
"prefix", update.prefix, "error", err)
continue
}
@@ -314,30 +322,7 @@ func (h *PrefixHandler) flushBatch() {
}
}
// Process batch operations
successCount := 0
if len(routesToUpsert) > 0 {
if err := h.db.UpsertLiveRouteBatch(routesToUpsert); err != nil {
h.logger.Error("Failed to upsert route batch", "error", err, "count", len(routesToUpsert))
} else {
successCount += len(routesToUpsert)
}
}
if len(routesToDelete) > 0 {
if err := h.db.DeleteLiveRouteBatch(routesToDelete); err != nil {
h.logger.Error("Failed to delete route batch", "error", err, "count", len(routesToDelete))
} else {
successCount += len(routesToDelete)
}
}
// Update prefix tables
if len(prefixesToUpdate) > 0 {
if err := h.db.UpdatePrefixesBatch(prefixesToUpdate); err != nil {
h.logger.Error("Failed to update prefix batch", "error", err, "count", len(prefixesToUpdate))
}
}
successCount := h.writeBatch(routesToUpsert, routesToDelete, prefixesToUpdate)
elapsed := time.Since(startTime)
h.logger.Debug("Flushed prefix batch",
@@ -351,6 +336,69 @@ func (h *PrefixHandler) flushBatch() {
h.lastFlush = time.Now()
}
// deleteLiveRoutesOfPeers deletes all live routes of each of peers, the peers
// whose BGP session went down. Most downs come from peers that have no routes,
// so only a delete that removed routes is logged.
func (h *PrefixHandler) deleteLiveRoutesOfPeers(peers []string) {
for _, peer := range peers {
deleted, err := h.db.DeleteLiveRoutesOfPeer(peer)
if err != nil {
h.logger.Error(
"Failed to delete the live routes of a peer whose session went down",
"peer", peer, "error", err)
continue
}
if deleted > 0 {
h.logger.Info(
"Deleted the live routes of a peer whose session went down",
"peer", peer, "routes", deleted)
}
}
}
// writeBatch writes the collected route upserts, route deletions and prefix
// updates to the database and returns how many routes it wrote.
func (h *PrefixHandler) writeBatch(
routesToUpsert []*database.LiveRoute,
routesToDelete []database.LiveRouteDeletion,
prefixesToUpdate map[string]time.Time,
) int {
successCount := 0
if len(routesToUpsert) > 0 {
err := h.db.UpsertLiveRouteBatch(routesToUpsert)
if err != nil {
h.logger.Error("Failed to upsert route batch",
"error", err, "count", len(routesToUpsert))
} else {
successCount += len(routesToUpsert)
}
}
if len(routesToDelete) > 0 {
err := h.db.DeleteLiveRouteBatch(routesToDelete)
if err != nil {
h.logger.Error("Failed to delete route batch",
"error", err, "count", len(routesToDelete))
} else {
successCount += len(routesToDelete)
}
}
// Update prefix tables
if len(prefixesToUpdate) > 0 {
err := h.db.UpdatePrefixesBatch(prefixesToUpdate)
if err != nil {
h.logger.Error("Failed to update prefix batch",
"error", err, "count", len(prefixesToUpdate))
}
}
return successCount
}
// canonicalPrefix returns prefix in the text form net/netip prints, the form
// the IP lookup builds when it looks a prefix up. The feed sends IPv6
// announcements compressed ("2001:db8::/32") but IPv6 withdrawals uncompressed
@@ -366,8 +414,8 @@ func canonicalPrefix(prefix string) string {
return p.Masked().String()
}
// parseCIDR extracts the mask length and IP version from a prefix string
func parseCIDR(prefix string) (maskLength int, ipVersion int, err error) {
// parseCIDR returns the mask length and the IP version of a prefix string
func parseCIDR(prefix string) (int, int, error) {
_, ipNet, err := net.ParseCIDR(prefix)
if err != nil {
return 0, 0, err
@@ -381,50 +429,6 @@ func parseCIDR(prefix string) (maskLength int, ipVersion int, err error) {
return ones, ipv4Version, nil
}
// processAnnouncement handles storing an announcement in the database
// nolint:unused // kept for potential future use
func (h *PrefixHandler) processAnnouncement(_ *database.Prefix, update prefixUpdate) {
// Parse CIDR to get mask length
maskLength, ipVersion, err := parseCIDR(update.prefix)
if err != nil {
h.logger.Error("Failed to parse CIDR",
"prefix", update.prefix,
"error", err,
)
return
}
// Track route update metrics
if h.metrics != nil {
if ipVersion == ipv4Version {
h.metrics.RecordIPv4Update()
} else {
h.metrics.RecordIPv6Update()
}
}
// Create live route record
liveRoute := &database.LiveRoute{
ID: uuid.New(),
Prefix: update.prefix,
MaskLength: maskLength,
IPVersion: ipVersion,
OriginASN: update.originASN,
PeerIP: update.peer,
ASPath: update.path,
NextHop: update.peer, // Using peer as next hop
LastUpdated: update.timestamp,
}
if err := h.db.UpsertLiveRoute(liveRoute); err != nil {
h.logger.Error("Failed to upsert live route",
"prefix", update.prefix,
"error", err,
)
}
}
// createLiveRoute creates a LiveRoute from a prefix update
func (h *PrefixHandler) createLiveRoute(update prefixUpdate) *database.LiveRoute {
// Parse CIDR to get mask length
@@ -462,77 +466,3 @@ func (h *PrefixHandler) createLiveRoute(update prefixUpdate) *database.LiveRoute
return liveRoute
}
// processAnnouncementDirect handles storing an announcement directly without prefix table
// nolint:unused // kept for potential future use
func (h *PrefixHandler) processAnnouncementDirect(update prefixUpdate) {
// Parse CIDR to get mask length
maskLength, ipVersion, err := parseCIDR(update.prefix)
if err != nil {
h.logger.Error("Failed to parse CIDR",
"prefix", update.prefix,
"error", err,
)
return
}
// Track route update metrics
if h.metrics != nil {
if ipVersion == ipv4Version {
h.metrics.RecordIPv4Update()
} else {
h.metrics.RecordIPv6Update()
}
}
// Create live route record
liveRoute := &database.LiveRoute{
ID: uuid.New(),
Prefix: update.prefix,
MaskLength: maskLength,
IPVersion: ipVersion,
OriginASN: update.originASN,
PeerIP: update.peer,
ASPath: update.path,
NextHop: update.peer, // Using peer as next hop
LastUpdated: update.timestamp,
}
if err := h.db.UpsertLiveRoute(liveRoute); err != nil {
h.logger.Error("Failed to upsert live route",
"prefix", update.prefix,
"error", err,
)
}
}
// processWithdrawalDirect handles removing a route directly without prefix table
// nolint:unused // kept for potential future use
func (h *PrefixHandler) processWithdrawalDirect(update prefixUpdate) {
if err := h.db.DeleteLiveRoute(update.prefix, update.peer); err != nil {
h.logger.Error("Failed to delete live route",
"prefix", update.prefix,
"peer", update.peer,
"error", err,
)
}
}
// processWithdrawal handles removing a route from the live routing table
// nolint:unused // kept for potential future use
func (h *PrefixHandler) processWithdrawal(_ *database.Prefix, update prefixUpdate) {
if err := h.db.DeleteLiveRoute(update.prefix, update.peer); err != nil {
h.logger.Error("Failed to delete live route",
"prefix", update.prefix,
"peer", update.peer,
"error", err,
)
}
}
// Stop gracefully stops the handler and flushes remaining batches
func (h *PrefixHandler) Stop() {
close(h.stopCh)
h.wg.Wait()
}
internal/routewatch/prefixhandler_test.go → internal/routewatch/prefixhandler_internal_test.go
+260 -153
View File
@@ -17,6 +17,12 @@ import (
const testPeerIP = "2001:db8:ffff::1"
// An IPv4 and an IPv6 prefix the tests announce.
const (
testV4Prefix = "192.0.2.0/24"
testV6Prefix = "2001:db8:1::/48"
)
// TestPrefixHandlerStoresPrefixesTheIPLookupFinds runs announcements and
// withdrawals through the prefix handler into a real database and looks the
// addresses up. The prefixes are written the way the feed sends them: IPv6
@@ -24,11 +30,9 @@ const testPeerIP = "2001:db8:ffff::1"
// remove the route the announcement stored, and the IP lookup must find the
// stored prefix.
func TestPrefixHandlerStoresPrefixesTheIPLookupFinds(t *testing.T) {
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
defer func() { _ = db.Close() }()
t.Parallel()
db := newTestDatabase(t)
// Built without NewPrefixHandler's flush timer, so the test flushes itself.
h := &PrefixHandler{db: db, logger: logger.New()}
@@ -40,34 +44,38 @@ func TestPrefixHandlerStoresPrefixesTheIPLookupFinds(t *testing.T) {
ParsedTimestamp: ts,
Announcements: []ristypes.RISAnnouncement{{
NextHop: testPeerIP,
Prefixes: []string{"2001:db8:1::/48", "192.0.2.0/24"},
Prefixes: []string{testV6Prefix, testV4Prefix},
}},
})
h.flushBatch()
for ip, want := range map[string]string{
"2001:db8:1::1": "2001:db8:1::/48",
"192.0.2.1": "192.0.2.0/24",
"2001:db8:1::1": testV6Prefix,
"192.0.2.1": testV4Prefix,
} {
info, err := db.GetIPInfo(ip)
if err != nil {
t.Fatalf("GetIPInfo(%s) after announcement: %v", ip, err)
}
if info.Netblock != want || info.ASN != testASNB {
t.Errorf("GetIPInfo(%s) = %s AS%d, want %s AS%d", ip, info.Netblock, info.ASN, want, testASNB)
t.Errorf("GetIPInfo(%s) = %s AS%d, want %s AS%d",
ip, info.Netblock, info.ASN, want, testASNB)
}
}
h.HandleMessage(&ristypes.RISMessage{
Peer: testPeerIP,
ParsedTimestamp: ts.Add(time.Minute),
Withdrawals: []string{"2001:db8:1:0:0:0:0:0/48", "192.0.2.0/24"},
Withdrawals: []string{"2001:db8:1:0:0:0:0:0/48", testV4Prefix},
})
h.flushBatch()
for _, ip := range []string{"2001:db8:1::1", "192.0.2.1"} {
if info, err := db.GetIPInfo(ip); !errors.Is(err, database.ErrNoRoute) {
t.Errorf("GetIPInfo(%s) after withdrawal = %+v, %v; want ErrNoRoute", ip, info, err)
info, err := db.GetIPInfo(ip)
if !errors.Is(err, database.ErrNoRoute) {
t.Errorf("GetIPInfo(%s) after withdrawal = %+v, %v; want ErrNoRoute",
ip, info, err)
}
}
}
@@ -79,6 +87,8 @@ func TestPrefixHandlerStoresPrefixesTheIPLookupFinds(t *testing.T) {
// live route for it, and updates from different peers are all applied. The
// in-memory route count and prefix distribution must match the route table.
func TestPrefixHandlerAppliesEachPeersLastUpdateInABatch(t *testing.T) {
t.Parallel()
const (
prefix = "198.51.100.0/24"
prefixMaskLength = 24
@@ -91,11 +101,14 @@ func TestPrefixHandlerAppliesEachPeersLastUpdateInABatch(t *testing.T) {
Peer: peer,
Path: ristypes.ASPath{testASNA, testASNB},
ParsedTimestamp: ts,
Announcements: []ristypes.RISAnnouncement{{NextHop: peer, Prefixes: []string{prefix}}},
Announcements: []ristypes.RISAnnouncement{
{NextHop: peer, Prefixes: []string{prefix}},
},
}
}
withdraw := func(peer string) *ristypes.RISMessage {
return &ristypes.RISMessage{Peer: peer, ParsedTimestamp: ts, Withdrawals: []string{prefix}}
return &ristypes.RISMessage{Peer: peer, ParsedTimestamp: ts,
Withdrawals: []string{prefix}}
}
announceAndWithdraw := func(peer string) *ristypes.RISMessage {
msg := announce(peer)
@@ -112,8 +125,10 @@ func TestPrefixHandlerAppliesEachPeersLastUpdateInABatch(t *testing.T) {
wantPeers []string
}{
{
name: "two peers announce the prefix",
batches: [][]*ristypes.RISMessage{{announce(testPeerIP), announce(otherPeerIP)}},
name: "two peers announce the prefix",
batches: [][]*ristypes.RISMessage{
{announce(testPeerIP), announce(otherPeerIP)},
},
wantPeers: []string{testPeerIP, otherPeerIP},
},
{
@@ -149,39 +164,10 @@ func TestPrefixHandlerAppliesEachPeersLastUpdateInABatch(t *testing.T) {
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
defer func() { _ = db.Close() }()
t.Parallel()
// Built without NewPrefixHandler's flush timer, so the test flushes itself.
h := &PrefixHandler{db: db, logger: logger.New()}
for _, batch := range tt.batches {
for _, msg := range batch {
h.HandleMessage(msg)
}
h.flushBatch()
}
ctx := context.Background()
routes, _, err := db.GetPrefixDetailsContext(ctx, prefix)
if err != nil && !errors.Is(err, database.ErrNoRoute) {
t.Fatalf("GetPrefixDetailsContext: %v", err)
}
var gotPeers []string
for _, route := range routes {
gotPeers = append(gotPeers, route.PeerIP)
}
if !slices.Equal(gotPeers, tt.wantPeers) {
t.Errorf("peers with a live route for %s = %v, want %v", prefix, gotPeers, tt.wantPeers)
}
var wantDistribution []database.PrefixDistribution
if len(tt.wantPeers) > 0 {
wantDistribution = []database.PrefixDistribution{{MaskLength: prefixMaskLength, Count: 1}}
}
assertCounts(t, db, ipv4Version, len(tt.wantPeers), wantDistribution)
db := handleBatches(t, tt.batches)
checkLivePeers(t, db, prefix, prefixMaskLength, tt.wantPeers)
})
}
}
@@ -194,14 +180,22 @@ func TestPrefixHandlerAppliesEachPeersLastUpdateInABatch(t *testing.T) {
// in. The in-memory route count and prefix distribution must match the route
// table.
func TestPrefixHandlerKeepsOneRoutePerPrefixAndPeer(t *testing.T) {
t.Parallel()
families := []struct {
ipVersion int
prefix, otherPrefix string
// The mask length of both prefixes.
maskLength int
}{
{ipVersion: ipv4Version, prefix: "198.51.100.0/24", otherPrefix: "203.0.113.0/24", maskLength: 24},
{ipVersion: ipv6Version, prefix: "2001:db8:1::/48", otherPrefix: "2001:db8:2::/48", maskLength: 48},
{
ipVersion: ipv4Version, maskLength: 24,
prefix: "198.51.100.0/24", otherPrefix: "203.0.113.0/24",
},
{
ipVersion: ipv6Version, maskLength: 48,
prefix: testV6Prefix, otherPrefix: "2001:db8:2::/48",
},
}
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
@@ -210,7 +204,9 @@ func TestPrefixHandlerKeepsOneRoutePerPrefixAndPeer(t *testing.T) {
Peer: testPeerIP,
Path: path,
ParsedTimestamp: ts,
Announcements: []ristypes.RISAnnouncement{{NextHop: testPeerIP, Prefixes: []string{announced}}},
Announcements: []ristypes.RISAnnouncement{
{NextHop: testPeerIP, Prefixes: []string{announced}},
},
}
}
@@ -219,71 +215,50 @@ func TestPrefixHandlerKeepsOneRoutePerPrefixAndPeer(t *testing.T) {
// One message that announces otherPrefix with origin AS testASNB and
// withdraws prefix.
announceOtherAndWithdraw := announce(ristypes.ASPath{testASNA, testASNB}, otherPrefix)
announceOtherAndWithdraw := announce(
ristypes.ASPath{testASNA, testASNB}, otherPrefix)
announceOtherAndWithdraw.Withdrawals = []string{prefix}
tests := []struct {
name string
messages []*ristypes.RISMessage
name string
// Each message is handled and then flushed as its own batch.
batches [][]*ristypes.RISMessage
// The origin AS of the peer's live route for each prefix. A prefix
// left out must have no live route.
wantOrigins map[string]int
}{
{
name: "the peer announces the prefix with a new origin AS",
messages: []*ristypes.RISMessage{
announce(ristypes.ASPath{testASNA}, prefix),
announce(ristypes.ASPath{testASNA, testASNB}, prefix),
batches: [][]*ristypes.RISMessage{
{announce(ristypes.ASPath{testASNA}, prefix)},
{announce(ristypes.ASPath{testASNA, testASNB}, prefix)},
},
wantOrigins: map[string]int{prefix: testASNB},
},
{
name: "one message announces a prefix with another origin AS and withdraws the prefix",
messages: []*ristypes.RISMessage{
announce(ristypes.ASPath{testASNA}, prefix),
announceOtherAndWithdraw,
name: "one message announces a prefix with another origin AS " +
"and withdraws the prefix",
batches: [][]*ristypes.RISMessage{
{announce(ristypes.ASPath{testASNA}, prefix)},
{announceOtherAndWithdraw},
},
wantOrigins: map[string]int{otherPrefix: testASNB},
},
}
for _, tt := range tests {
t.Run(fmt.Sprintf("IPv%d %s", family.ipVersion, tt.name), func(t *testing.T) {
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
defer func() { _ = db.Close() }()
name := fmt.Sprintf("IPv%d %s", family.ipVersion, tt.name)
t.Run(name, func(t *testing.T) {
t.Parallel()
// Built without NewPrefixHandler's flush timer, so the test flushes itself.
h := &PrefixHandler{db: db, logger: logger.New()}
for _, msg := range tt.messages {
h.HandleMessage(msg)
h.flushBatch()
}
for _, p := range []string{prefix, otherPrefix} {
routes, _, err := db.GetPrefixDetailsContext(context.Background(), p)
if err != nil && !errors.Is(err, database.ErrNoRoute) {
t.Fatalf("GetPrefixDetailsContext(%s): %v", p, err)
}
var gotOrigins []int
for _, route := range routes {
gotOrigins = append(gotOrigins, route.OriginASN)
}
var wantOrigins []int
if origin, ok := tt.wantOrigins[p]; ok {
wantOrigins = []int{origin}
}
if !slices.Equal(gotOrigins, wantOrigins) {
t.Errorf("origin ASes of the live routes for %s = %v, want %v", p, gotOrigins, wantOrigins)
}
}
db := handleBatches(t, tt.batches)
checkLiveOrigins(t, db, tt.wantOrigins, prefix, otherPrefix)
wantDistribution := []database.PrefixDistribution{
{MaskLength: family.maskLength, Count: len(tt.wantOrigins)},
}
assertCounts(t, db, family.ipVersion, len(tt.wantOrigins), wantDistribution)
assertCounts(t, db, family.ipVersion, len(tt.wantOrigins),
wantDistribution)
})
}
}
@@ -293,12 +268,16 @@ func TestPrefixHandlerKeepsOneRoutePerPrefixAndPeer(t *testing.T) {
// messages go to the prefix handler, which acts on a down, and not to the peer
// handler, which would record a peer whose session never comes up as active.
func TestStateMessagesGoToThePrefixHandlerNotThePeerHandler(t *testing.T) {
t.Parallel()
if !(&PrefixHandler{}).WantsMessage(ristypes.MessageTypeState) {
t.Error("the prefix handler does not want STATE messages")
}
if (&PeerHandler{}).WantsMessage(ristypes.MessageTypeState) {
t.Error("the peer handler wants STATE messages")
}
if !(&PeerHandler{}).WantsMessage(ristypes.MessageTypeUpdate) {
t.Error("the peer handler does not want UPDATE messages")
}
@@ -313,6 +292,8 @@ func TestStateMessagesGoToThePrefixHandlerNotThePeerHandler(t *testing.T) {
// in the same batch, is kept. The in-memory route counts and prefix
// distribution must match the route tables.
func TestPrefixHandlerDeletesAPeersRoutesWhenItsSessionGoesDown(t *testing.T) {
t.Parallel()
const (
otherPeerIP = "2001:db8:ffff::2"
// Both peers announce the shared prefixes, only testPeerIP the others.
@@ -335,14 +316,17 @@ func TestPrefixHandlerDeletesAPeersRoutesWhenItsSessionGoesDown(t *testing.T) {
wantAfterDown bool
}{
{
name: "the down ends the batch",
batch: []*ristypes.RISMessage{peerAnnouncement(testPeerIP, beforeDown), sessionDown(testPeerIP)},
name: "the down ends the batch",
batch: []*ristypes.RISMessage{
peerAnnouncement(testPeerIP, beforeDown),
sessionDown(),
},
},
{
name: "the peer announces a prefix after the down",
batch: []*ristypes.RISMessage{
peerAnnouncement(testPeerIP, beforeDown),
sessionDown(testPeerIP),
sessionDown(),
peerAnnouncement(testPeerIP, afterDown),
},
wantAfterDown: true,
@@ -351,21 +335,16 @@ func TestPrefixHandlerDeletesAPeersRoutesWhenItsSessionGoesDown(t *testing.T) {
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
defer func() { _ = db.Close() }()
t.Parallel()
// Built without NewPrefixHandler's flush timer, so the test flushes itself.
h := &PrefixHandler{db: db, logger: logger.New()}
h.HandleMessage(peerAnnouncement(testPeerIP, sharedV4, sharedV6, onlyV4, onlyV6))
h.HandleMessage(peerAnnouncement(otherPeerIP, sharedV4, sharedV6))
h.flushBatch()
for _, msg := range tt.batch {
h.HandleMessage(msg)
}
h.flushBatch()
db := handleBatches(t, [][]*ristypes.RISMessage{
{
peerAnnouncement(testPeerIP,
sharedV4, sharedV6, onlyV4, onlyV6),
peerAnnouncement(otherPeerIP, sharedV4, sharedV6),
},
tt.batch,
})
// The peers left with a live route for each prefix.
wantPeers := map[string][]string{
@@ -377,16 +356,20 @@ func TestPrefixHandlerDeletesAPeersRoutesWhenItsSessionGoesDown(t *testing.T) {
afterDown: nil,
}
wantRoutesV6 := 1
if tt.wantAfterDown {
wantPeers[afterDown] = []string{testPeerIP}
wantRoutesV6++
}
assertPeers(t, db, wantPeers)
assertCounts(t, db, ipv4Version, 1,
[]database.PrefixDistribution{{MaskLength: maskLengthV4, Count: 1}})
assertCounts(t, db, ipv4Version, 1, []database.PrefixDistribution{
{MaskLength: maskLengthV4, Count: 1},
})
assertCounts(t, db, ipv6Version, wantRoutesV6,
[]database.PrefixDistribution{{MaskLength: maskLengthV6, Count: wantRoutesV6}})
[]database.PrefixDistribution{
{MaskLength: maskLengthV6, Count: wantRoutesV6},
})
})
}
}
@@ -398,6 +381,8 @@ func TestPrefixHandlerDeletesAPeersRoutesWhenItsSessionGoesDown(t *testing.T) {
// that handles it and on no later flush, so the route the peer announced after
// it is kept.
func TestPrefixHandlerKeepsWhatAPeerAnnouncesAfterTheFlushOfItsDown(t *testing.T) {
t.Parallel()
const (
otherPeerIP = "2001:db8:ffff::2"
beforeDown = "192.0.2.0/24"
@@ -406,30 +391,22 @@ func TestPrefixHandlerKeepsWhatAPeerAnnouncesAfterTheFlushOfItsDown(t *testing.T
maskLength = 24
)
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
defer func() { _ = db.Close() }()
// Built without NewPrefixHandler's flush timer, so the test flushes itself.
h := &PrefixHandler{db: db, logger: logger.New()}
for _, msg := range []*ristypes.RISMessage{
peerAnnouncement(testPeerIP, beforeDown),
sessionDown(testPeerIP),
peerAnnouncement(testPeerIP, afterDown),
peerAnnouncement(otherPeerIP, otherPeers),
} {
h.HandleMessage(msg)
h.flushBatch()
}
// Each message is handled and then flushed as its own batch.
db := handleBatches(t, [][]*ristypes.RISMessage{
{peerAnnouncement(testPeerIP, beforeDown)},
{sessionDown()},
{peerAnnouncement(testPeerIP, afterDown)},
{peerAnnouncement(otherPeerIP, otherPeers)},
})
assertPeers(t, db, map[string][]string{
beforeDown: nil,
afterDown: {testPeerIP},
otherPeers: {otherPeerIP},
})
assertCounts(t, db, ipv4Version, 2, []database.PrefixDistribution{{MaskLength: maskLength, Count: 2}})
assertCounts(t, db, ipv4Version, 2, []database.PrefixDistribution{
{MaskLength: maskLength, Count: 2},
})
}
// firstWriteHeldStore is a real database whose first live route batch write
@@ -442,7 +419,11 @@ type firstWriteHeldStore struct {
once sync.Once
}
func (s *firstWriteHeldStore) UpsertLiveRouteBatch(routes []*database.LiveRoute) error {
// UpsertLiveRouteBatch holds the first call until release is closed, then
// writes routes to the database.
func (s *firstWriteHeldStore) UpsertLiveRouteBatch(
routes []*database.LiveRoute,
) error {
s.once.Do(func() {
close(s.writeStarted)
<-s.release
@@ -457,27 +438,32 @@ func (s *firstWriteHeldStore) UpsertLiveRouteBatch(routes []*database.LiveRoute)
// must delete the route the held write stored and keep the one announced after
// the down.
func TestPrefixHandlerHandlesADownWhileAFlushWrites(t *testing.T) {
t.Parallel()
const (
beforeDown = "192.0.2.0/24"
afterDown = "198.51.100.0/24"
maskLength = 24
)
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
db := newTestDatabase(t)
store := &firstWriteHeldStore{
Database: db,
writeStarted: make(chan struct{}),
release: make(chan struct{}),
}
defer func() { _ = db.Close() }()
store := &firstWriteHeldStore{Database: db, writeStarted: make(chan struct{}), release: make(chan struct{})}
// Built without NewPrefixHandler's flush timer, so the test flushes itself.
h := &PrefixHandler{db: store, logger: logger.New()}
h.HandleMessage(peerAnnouncement(testPeerIP, beforeDown))
flushed := make(chan struct{})
go func() {
h.flushBatch()
close(flushed)
}()
select {
case <-store.writeStarted:
case <-time.After(handleMessageTimeout):
@@ -485,21 +471,67 @@ func TestPrefixHandlerHandlesADownWhileAFlushWrites(t *testing.T) {
}
if !returnsInTime(func() {
h.HandleMessage(sessionDown(testPeerIP))
h.HandleMessage(sessionDown())
h.HandleMessage(peerAnnouncement(testPeerIP, afterDown))
}) {
t.Fatal("HandleMessage did not return while a database write was held")
}
close(store.release)
select {
case <-flushed:
case <-time.After(handleMessageTimeout):
t.Fatal("the flush did not finish after its write was let go")
}
h.flushBatch()
assertPeers(t, db, map[string][]string{beforeDown: nil, afterDown: {testPeerIP}})
assertCounts(t, db, ipv4Version, 1, []database.PrefixDistribution{{MaskLength: maskLength, Count: 1}})
assertPeers(t, db, map[string][]string{
beforeDown: nil,
afterDown: {testPeerIP},
})
assertCounts(t, db, ipv4Version, 1, []database.PrefixDistribution{
{MaskLength: maskLength, Count: 1},
})
}
// newTestDatabase opens a database in a new temporary directory. It is closed
// when the test ends.
func newTestDatabase(t *testing.T) *database.Database {
t.Helper()
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("failed to create database: %v", err)
}
t.Cleanup(func() { _ = db.Close() })
return db
}
// handleBatches runs each batch of messages through a prefix handler into a
// new database, flushing after each batch, and returns the database.
func handleBatches(
t *testing.T, batches [][]*ristypes.RISMessage,
) *database.Database {
t.Helper()
db := newTestDatabase(t)
// Built without NewPrefixHandler's flush timer, so the test flushes itself.
h := &PrefixHandler{db: db, logger: logger.New()}
for _, batch := range batches {
for _, msg := range batch {
h.HandleMessage(msg)
}
h.flushBatch()
}
return db
}
// peerAnnouncement returns an UPDATE message in which peer announces prefixes
@@ -510,18 +542,48 @@ func peerAnnouncement(peer string, prefixes ...string) *ristypes.RISMessage {
Peer: peer,
Path: ristypes.ASPath{testASNA, testASNB},
ParsedTimestamp: time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC),
Announcements: []ristypes.RISAnnouncement{{NextHop: peer, Prefixes: prefixes}},
Announcements: []ristypes.RISAnnouncement{
{NextHop: peer, Prefixes: prefixes},
},
}
}
// sessionDown returns the STATE message that says peer's BGP session went down.
func sessionDown(peer string) *ristypes.RISMessage {
return &ristypes.RISMessage{Type: ristypes.MessageTypeState, Peer: peer, State: ristypes.StateDown}
// sessionDown returns the STATE message that says the BGP session of the peer
// testPeerIP went down.
func sessionDown() *ristypes.RISMessage {
return &ristypes.RISMessage{
Type: ristypes.MessageTypeState,
Peer: testPeerIP,
State: ristypes.StateDown,
}
}
// checkLivePeers checks that exactly wantPeers have a live route for the IPv4
// prefix, and that the in-memory route count and prefix distribution match
// the route table.
func checkLivePeers(
t *testing.T, db *database.Database, prefix string, maskLength int,
wantPeers []string,
) {
t.Helper()
assertPeers(t, db, map[string][]string{prefix: wantPeers})
var wantDistribution []database.PrefixDistribution
if len(wantPeers) > 0 {
wantDistribution = []database.PrefixDistribution{
{MaskLength: maskLength, Count: 1},
}
}
assertCounts(t, db, ipv4Version, len(wantPeers), wantDistribution)
}
// assertPeers checks, for each prefix in wantPeers, that the peers with a live
// route for it are the ones listed, in peer IP order; nil means none.
func assertPeers(t *testing.T, db *database.Database, wantPeers map[string][]string) {
func assertPeers(
t *testing.T, db *database.Database, wantPeers map[string][]string,
) {
t.Helper()
for prefix, want := range wantPeers {
@@ -529,12 +591,46 @@ func assertPeers(t *testing.T, db *database.Database, wantPeers map[string][]str
if err != nil && !errors.Is(err, database.ErrNoRoute) {
t.Fatalf("GetPrefixDetailsContext(%s): %v", prefix, err)
}
var got []string
got := make([]string, 0, len(routes))
for _, route := range routes {
got = append(got, route.PeerIP)
}
if !slices.Equal(got, want) {
t.Errorf("peers with a live route for %s = %v, want %v", prefix, got, want)
t.Errorf("peers with a live route for %s = %v, want %v",
prefix, got, want)
}
}
}
// checkLiveOrigins checks that each of prefixes has one live route, with the
// origin AS wantOrigins gives for it, or none when wantOrigins leaves it out.
func checkLiveOrigins(
t *testing.T, db *database.Database, wantOrigins map[string]int,
prefixes ...string,
) {
t.Helper()
for _, prefix := range prefixes {
routes, _, err := db.GetPrefixDetailsContext(context.Background(), prefix)
if err != nil && !errors.Is(err, database.ErrNoRoute) {
t.Fatalf("GetPrefixDetailsContext(%s): %v", prefix, err)
}
gotOrigins := make([]int, 0, len(routes))
for _, route := range routes {
gotOrigins = append(gotOrigins, route.OriginASN)
}
var want []int
if origin, ok := wantOrigins[prefix]; ok {
want = []int{origin}
}
if !slices.Equal(gotOrigins, want) {
t.Errorf("origin ASes of the live routes for %s = %v, want %v",
prefix, gotOrigins, want)
}
}
}
@@ -542,38 +638,49 @@ func assertPeers(t *testing.T, db *database.Database, wantPeers map[string][]str
// assertCounts checks that the live route count and prefix distribution of one
// IP version are wantRoutes and wantDistribution, both as held in memory and as
// read from the route table.
func assertCounts(t *testing.T, db *database.Database, ipVersion, wantRoutes int,
wantDistribution []database.PrefixDistribution) {
func assertCounts(
t *testing.T, db *database.Database, ipVersion, wantRoutes int,
wantDistribution []database.PrefixDistribution,
) {
t.Helper()
ctx := context.Background()
stats, err := db.GetStatsContext(ctx)
if err != nil {
t.Fatalf("GetStatsContext: %v", err)
}
tableRoutesV4, tableRoutesV6, err := db.GetLiveRouteCountsContext(ctx)
if err != nil {
t.Fatalf("GetLiveRouteCountsContext: %v", err)
}
queryDistributionV4, queryDistributionV6, err := db.GetPrefixDistributionContext(ctx)
queryDistributionV4, queryDistributionV6, err :=
db.GetPrefixDistributionContext(ctx)
if err != nil {
t.Fatalf("GetPrefixDistributionContext: %v", err)
}
memoryRoutes, tableRoutes := stats.IPv4Routes, tableRoutesV4
memoryDistribution, queryDistribution := stats.IPv4PrefixDistribution, queryDistributionV4
memoryDistribution := stats.IPv4PrefixDistribution
queryDistribution := queryDistributionV4
if ipVersion == ipv6Version {
memoryRoutes, tableRoutes = stats.IPv6Routes, tableRoutesV6
memoryDistribution, queryDistribution = stats.IPv6PrefixDistribution, queryDistributionV6
memoryDistribution = stats.IPv6PrefixDistribution
queryDistribution = queryDistributionV6
}
if memoryRoutes != wantRoutes || tableRoutes != wantRoutes {
t.Errorf("IPv%d live routes: %d in memory, %d in the table, want %d",
ipVersion, memoryRoutes, tableRoutes, wantRoutes)
}
if !slices.Equal(memoryDistribution, wantDistribution) ||
!slices.Equal(queryDistribution, wantDistribution) {
t.Errorf("IPv%d prefix distribution: %v in memory, %v from the table, want %v",
ipVersion, memoryDistribution, queryDistribution, wantDistribution)
t.Errorf("IPv%d prefix distribution: %v in memory, %v from the table, "+
"want %v", ipVersion, memoryDistribution, queryDistribution,
wantDistribution)
}
}
internal/server/handlers.go
+674 -595
View File
File diff suppressed because it is too large Load Diff
internal/server/handlers_test.go → internal/server/handlers_internal_test.go
+193 -76
View File
@@ -22,12 +22,26 @@ import (
"github.com/google/uuid"
)
// The prefix and peer addresses the prefix page tests store routes for.
const (
testPrefixV4 = "192.0.2.0/24"
testPeerV4 = "198.51.100.1"
testPeerV6 = "2001:db8::1"
)
// The names the stats handler tests report the two stats handlers by.
const (
statusJSONHandlerName = "status.json"
statsHandlerName = "stats"
)
// blockingStatsDB embeds database.Store (left nil) and overrides only
// GetStatsContext, which blocks until release is closed. The stats handlers
// call it in a goroutine; every other Store method is unused on the timeout
// path and would panic if called.
type blockingStatsDB struct {
database.Store
release chan struct{}
}
@@ -43,14 +57,16 @@ func (d blockingStatsDB) GetStatsContext(_ context.Context) (database.Stats, err
// value. Before the fix the per-request goroutine sent on an unbuffered channel
// that nothing received once the timeout won, so it blocked forever and every
// poll leaked one goroutine.
//
//nolint:paralleltest // counts goroutines, so no other test may run alongside
func TestStatsHandlersDoNotLeakOnTimeout(t *testing.T) {
release := make(chan struct{})
db := blockingStatsDB{release: release}
s := New(db, streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
handlers := map[string]http.HandlerFunc{
"status.json": s.handleStatusJSON(),
"stats": s.handleStats(),
statusJSONHandlerName: s.handleStatusJSON(),
statsHandlerName: s.handleStats(),
}
baseline := settledGoroutineCount()
@@ -62,7 +78,7 @@ func TestStatsHandlersDoNotLeakOnTimeout(t *testing.T) {
for _, handler := range handlers {
for range iterations {
ctx, cancel := context.WithTimeout(context.Background(), requestTimeout)
req := httptest.NewRequest(http.MethodGet, "/", nil).WithContext(ctx)
req := httptest.NewRequestWithContext(ctx, http.MethodGet, "/", nil)
handler(httptest.NewRecorder(), req)
cancel()
}
@@ -87,67 +103,83 @@ func TestStatsHandlersDoNotLeakOnTimeout(t *testing.T) {
// The oldest and newest route times still come from one-row lookups at the ends
// of an index; with the database closed they are left out of the answer.
func TestStatsHandlersAnswerFromMemory(t *testing.T) {
t.Parallel()
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("database.New: %v", err)
}
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
if err := db.UpsertLiveRouteBatch([]*database.LiveRoute{
err = db.UpsertLiveRouteBatch([]*database.LiveRoute{
{
ID: uuid.New(), Prefix: "198.51.100.0/24", MaskLength: 24, IPVersion: 4,
OriginASN: 64500, PeerIP: "192.0.2.1", ASPath: []int{64500}, NextHop: "192.0.2.1",
LastUpdated: ts,
OriginASN: 64500, PeerIP: "192.0.2.1", ASPath: []int{64500},
NextHop: "192.0.2.1", LastUpdated: ts,
},
{
ID: uuid.New(), Prefix: "2001:db8::/32", MaskLength: 32, IPVersion: 6,
OriginASN: 64501, PeerIP: "2001:db8::1", ASPath: []int{64501}, NextHop: "2001:db8::1",
LastUpdated: ts,
OriginASN: 64501, PeerIP: testPeerV6, ASPath: []int{64501},
NextHop: testPeerV6, LastUpdated: ts,
},
}); err != nil {
})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
if err := db.Close(); err != nil {
err = db.Close()
if err != nil {
t.Fatalf("Close: %v", err)
}
s := New(db, streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
handlers := map[string]http.HandlerFunc{
"status.json": s.handleStatusJSON(),
"stats": s.handleStats(),
statusJSONHandlerName: s.handleStatusJSON(),
statsHandlerName: s.handleStats(),
}
for name, handler := range handlers {
rec := httptest.NewRecorder()
handler(rec, httptest.NewRequest(http.MethodGet, "/", nil))
handler(rec, httptest.NewRequestWithContext(t.Context(), http.MethodGet, "/", nil))
if rec.Code != http.StatusOK {
t.Errorf("%s: status %d, want %d; body %s", name, rec.Code, http.StatusOK, rec.Body)
t.Errorf("%s: status %d, want %d; body %s",
name, rec.Code, http.StatusOK, rec.Body)
continue
}
//nolint:tagliatelle // field names are the existing JSON API
var body struct {
Data struct {
IPv4Routes int `json:"ipv4_routes"`
IPv6Routes int `json:"ipv6_routes"`
IPv4PrefixDistribution []database.PrefixDistribution `json:"ipv4_prefix_distribution"`
IPv6PrefixDistribution []database.PrefixDistribution `json:"ipv6_prefix_distribution"`
IPv4Routes int `json:"ipv4_routes"`
IPv6Routes int `json:"ipv6_routes"`
IPv4Distribution []database.PrefixDistribution `json:"ipv4_prefix_distribution"`
IPv6Distribution []database.PrefixDistribution `json:"ipv6_prefix_distribution"`
} `json:"data"`
}
if err := json.Unmarshal(rec.Body.Bytes(), &body); err != nil {
err := json.Unmarshal(rec.Body.Bytes(), &body)
if err != nil {
t.Fatalf("%s: decoding the answer: %v", name, err)
}
if body.Data.IPv4Routes != 1 || body.Data.IPv6Routes != 1 {
t.Errorf("%s: routes = (v4 %d, v6 %d), want (1, 1)", name, body.Data.IPv4Routes, body.Data.IPv6Routes)
t.Errorf("%s: routes = (v4 %d, v6 %d), want (1, 1)",
name, body.Data.IPv4Routes, body.Data.IPv6Routes)
}
wantV4 := []database.PrefixDistribution{{MaskLength: 24, Count: 1}}
if !slices.Equal(body.Data.IPv4PrefixDistribution, wantV4) {
t.Errorf("%s: IPv4 distribution = %v, want %v", name, body.Data.IPv4PrefixDistribution, wantV4)
if !slices.Equal(body.Data.IPv4Distribution, wantV4) {
t.Errorf("%s: IPv4 distribution = %v, want %v",
name, body.Data.IPv4Distribution, wantV4)
}
wantV6 := []database.PrefixDistribution{{MaskLength: 32, Count: 1}}
if !slices.Equal(body.Data.IPv6PrefixDistribution, wantV6) {
t.Errorf("%s: IPv6 distribution = %v, want %v", name, body.Data.IPv6PrefixDistribution, wantV6)
if !slices.Equal(body.Data.IPv6Distribution, wantV6) {
t.Errorf("%s: IPv6 distribution = %v, want %v",
name, body.Data.IPv6Distribution, wantV6)
}
}
}
@@ -159,25 +191,30 @@ func TestStatsHandlersAnswerFromMemory(t *testing.T) {
// "2001:DB8:0:0::/32" was not found although "2001:db8::/32" was stored, and
// "foo/24" answered 500 (https://git.eeqj.de/sneak/routewatch/issues/71).
func TestPrefixDetailHandlersParsePrefix(t *testing.T) {
t.Parallel()
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("database.New: %v", err)
}
t.Cleanup(func() { _ = db.Close() })
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
if err := db.UpsertLiveRouteBatch([]*database.LiveRoute{
err = db.UpsertLiveRouteBatch([]*database.LiveRoute{
{
ID: uuid.New(), Prefix: "192.0.2.0/24", MaskLength: 24, IPVersion: 4,
OriginASN: 64500, PeerIP: "198.51.100.1", ASPath: []int{64500}, NextHop: "198.51.100.1",
LastUpdated: ts,
ID: uuid.New(), Prefix: testPrefixV4, MaskLength: 24, IPVersion: 4,
OriginASN: 64500, PeerIP: testPeerV4, ASPath: []int{64500},
NextHop: testPeerV4, LastUpdated: ts,
},
{
ID: uuid.New(), Prefix: "2001:db8::/32", MaskLength: 32, IPVersion: 6,
OriginASN: 64501, PeerIP: "2001:db8::1", ASPath: []int{64501}, NextHop: "2001:db8::1",
LastUpdated: ts,
OriginASN: 64501, PeerIP: testPeerV6, ASPath: []int{64501},
NextHop: testPeerV6, LastUpdated: ts,
},
}); err != nil {
})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
@@ -197,20 +234,29 @@ func TestPrefixDetailHandlersParsePrefix(t *testing.T) {
for _, tt := range tests {
rec := httptest.NewRecorder()
s.router.ServeHTTP(rec, httptest.NewRequest(http.MethodGet, tt.path, nil))
s.router.ServeHTTP(rec, httptest.NewRequestWithContext(
t.Context(), http.MethodGet, tt.path, nil))
if rec.Code != tt.want {
t.Errorf("%s: status %d, want %d; body %s", tt.path, rec.Code, tt.want, rec.Body)
t.Errorf("%s: status %d, want %d; body %s",
tt.path, rec.Code, tt.want, rec.Body)
}
}
}
// errASDetailsLookupFailed is what failingASDetailsDB answers every AS details
// lookup with.
var errASDetailsLookupFailed = errors.New("AS details lookup failed")
// failingASDetailsDB wraps a real store and fails every AS details lookup.
type failingASDetailsDB struct {
database.Store
}
func (failingASDetailsDB) GetASDetailsContext(_ context.Context, _ int) (*database.ASN, []database.LiveRoute, error) {
return nil, nil, errors.New("AS details lookup failed")
func (failingASDetailsDB) GetASDetailsContext(
_ context.Context, _ int,
) (*database.ASN, []database.LiveRoute, error) {
return nil, nil, errASDetailsLookupFailed
}
// TestPrefixDetailShowsOriginNames checks that the prefix page shows each
@@ -220,51 +266,66 @@ func (failingASDetailsDB) GetASDetailsContext(_ context.Context, _ int) (*databa
// (https://git.eeqj.de/sneak/routewatch/issues/67). The AS path column shows
// handles too, so the test looks for the origin list's "(handle)" form.
func TestPrefixDetailShowsOriginNames(t *testing.T) {
t.Parallel()
db, err := database.New(&config.Config{StateDir: t.TempDir()}, logger.New())
if err != nil {
t.Fatalf("database.New: %v", err)
}
t.Cleanup(func() { _ = db.Close() })
ts := time.Date(2026, 1, 2, 3, 4, 5, 0, time.UTC)
origins := []int{15169, 13335}
for _, asn := range origins {
if _, err := db.GetOrCreateASN(asn, ts); err != nil {
_, err := db.GetOrCreateASN(asn, ts)
if err != nil {
t.Fatalf("GetOrCreateASN(%d): %v", asn, err)
}
}
if err := db.UpsertLiveRouteBatch([]*database.LiveRoute{
err = db.UpsertLiveRouteBatch([]*database.LiveRoute{
{
ID: uuid.New(), Prefix: "192.0.2.0/24", MaskLength: 24, IPVersion: 4,
OriginASN: origins[0], PeerIP: "198.51.100.1", ASPath: []int{origins[0]}, NextHop: "198.51.100.1",
LastUpdated: ts,
ID: uuid.New(), Prefix: testPrefixV4, MaskLength: 24, IPVersion: 4,
OriginASN: origins[0], PeerIP: testPeerV4,
ASPath: []int{origins[0]}, NextHop: testPeerV4, LastUpdated: ts,
},
{
ID: uuid.New(), Prefix: "192.0.2.0/24", MaskLength: 24, IPVersion: 4,
OriginASN: origins[1], PeerIP: "198.51.100.2", ASPath: []int{origins[1]}, NextHop: "198.51.100.2",
LastUpdated: ts,
ID: uuid.New(), Prefix: testPrefixV4, MaskLength: 24, IPVersion: 4,
OriginASN: origins[1], PeerIP: "198.51.100.2",
ASPath: []int{origins[1]}, NextHop: "198.51.100.2", LastUpdated: ts,
},
}); err != nil {
})
if err != nil {
t.Fatalf("UpsertLiveRouteBatch: %v", err)
}
s := New(failingASDetailsDB{db}, streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
s := New(failingASDetailsDB{db}, streamer.New(logger.New(), metrics.New()),
logger.New(), &config.Config{})
rec := httptest.NewRecorder()
s.router.ServeHTTP(rec, httptest.NewRequest(http.MethodGet, "/prefix/192.0.2.0/24", nil))
s.router.ServeHTTP(rec, httptest.NewRequestWithContext(
t.Context(), http.MethodGet, "/prefix/"+testPrefixV4, nil))
if rec.Code != http.StatusOK {
t.Fatalf("status %d, want %d; body %s", rec.Code, http.StatusOK, rec.Body)
}
page := rec.Body.String()
for _, asn := range origins {
info, ok := asinfo.Get(asn)
if !ok {
t.Fatalf("AS%d is not in the AS info data", asn)
}
if want := "(" + html.EscapeString(info.Handle) + ")"; !strings.Contains(page, want) {
want := "(" + html.EscapeString(info.Handle) + ")"
if !strings.Contains(page, want) {
t.Errorf("page does not show AS%d's handle %q", asn, want)
}
if want := html.EscapeString(info.Description); !strings.Contains(page, want) {
want = html.EscapeString(info.Description)
if !strings.Contains(page, want) {
t.Errorf("page does not show AS%d's description %q", asn, want)
}
}
@@ -282,14 +343,18 @@ const shortRequestDeadline = 50 * time.Millisecond
// written by the time limit middleware from its own goroutine while the
// handler could still be writing.
func TestRequestPastTimeLimitAnswers504(t *testing.T) {
s := New(nil, streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
t.Parallel()
s := New(nil, streamer.New(logger.New(), metrics.New()), logger.New(),
&config.Config{})
waitForTimeLimit := func(_ http.ResponseWriter, r *http.Request) {
<-r.Context().Done()
}
// Registered the way setupRoutes registers pages and JSON routes.
s.router.Get("/slow-page", waitForTimeLimit)
s.router.Get("/slow.json", JSONValidationMiddleware(http.HandlerFunc(waitForTimeLimit)).ServeHTTP)
s.router.Get("/slow.json",
JSONValidationMiddleware(http.HandlerFunc(waitForTimeLimit)).ServeHTTP)
tests := []struct {
path string
@@ -300,14 +365,18 @@ func TestRequestPastTimeLimitAnswers504(t *testing.T) {
}
for _, tt := range tests {
ctx, cancel := context.WithTimeout(context.Background(), shortRequestDeadline)
ctx, cancel := context.WithTimeout(context.Background(),
shortRequestDeadline)
rec := httptest.NewRecorder()
s.router.ServeHTTP(rec, httptest.NewRequest(http.MethodGet, tt.path, nil).WithContext(ctx))
s.router.ServeHTTP(rec,
httptest.NewRequestWithContext(ctx, http.MethodGet, tt.path, nil))
cancel()
if rec.Code != http.StatusGatewayTimeout {
t.Errorf("%s: status %d, want %d; body %s", tt.path, rec.Code, http.StatusGatewayTimeout, rec.Body)
t.Errorf("%s: status %d, want %d; body %s",
tt.path, rec.Code, http.StatusGatewayTimeout, rec.Body)
}
if tt.wantJSONBody {
checkJSONErrorCode(t, tt.path, rec, http.StatusGatewayTimeout)
}
@@ -320,25 +389,32 @@ func TestRequestPastTimeLimitAnswers504(t *testing.T) {
// body. Before the fix status.json answered 408, and /api/v1/stats wrote
// nothing, which the JSON check on its route turned into 500.
func TestStatsHandlersAnswer504WhenDatabaseReadRunsOut(t *testing.T) {
t.Parallel()
release := make(chan struct{})
defer close(release)
db := blockingStatsDB{release: release}
s := New(db, streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
s := New(db, streamer.New(logger.New(), metrics.New()), logger.New(),
&config.Config{})
handlers := map[string]http.HandlerFunc{
"status.json": s.handleStatusJSON(),
"stats": s.handleStats(),
statusJSONHandlerName: s.handleStatusJSON(),
statsHandlerName: s.handleStats(),
}
for name, handler := range handlers {
ctx, cancel := context.WithTimeout(context.Background(), shortRequestDeadline)
ctx, cancel := context.WithTimeout(context.Background(),
shortRequestDeadline)
rec := httptest.NewRecorder()
handler(rec, httptest.NewRequest(http.MethodGet, "/", nil).WithContext(ctx))
handler(rec, httptest.NewRequestWithContext(ctx, http.MethodGet, "/", nil))
cancel()
if rec.Code != http.StatusGatewayTimeout {
t.Errorf("%s: status %d, want %d; body %s", name, rec.Code, http.StatusGatewayTimeout, rec.Body)
t.Errorf("%s: status %d, want %d; body %s",
name, rec.Code, http.StatusGatewayTimeout, rec.Body)
}
checkJSONErrorCode(t, name, rec, http.StatusGatewayTimeout)
}
}
@@ -360,25 +436,33 @@ func (slowDB) GetStatsContext(ctx context.Context) (database.Stats, error) {
return database.Stats{}, nil
}
func (slowDB) GetIPInfoContext(ctx context.Context, _ string) (*database.IPInfo, error) {
func (slowDB) GetIPInfoContext(
ctx context.Context, _ string,
) (*database.IPInfo, error) {
<-ctx.Done()
return nil, ctx.Err()
}
func (slowDB) GetASDetailsContext(ctx context.Context, asn int) (*database.ASN, []database.LiveRoute, error) {
func (slowDB) GetASDetailsContext(
ctx context.Context, asn int,
) (*database.ASN, []database.LiveRoute, error) {
<-ctx.Done()
return &database.ASN{ASN: asn}, nil, nil
}
func (slowDB) GetPrefixDetailsContext(ctx context.Context, _ string) ([]database.LiveRoute, map[int]database.ASN, error) {
func (slowDB) GetPrefixDetailsContext(
ctx context.Context, _ string,
) ([]database.LiveRoute, map[int]database.ASN, error) {
<-ctx.Done()
return nil, nil, database.ErrNoRoute
}
func (slowDB) GetRandomPrefixesByLengthContext(ctx context.Context, _, _, _ int) ([]database.LiveRoute, error) {
func (slowDB) GetRandomPrefixesByLengthContext(
ctx context.Context, _, _, _ int,
) ([]database.LiveRoute, error) {
<-ctx.Done()
return nil, nil
@@ -393,7 +477,10 @@ func (slowDB) GetRandomPrefixesByLengthContext(ctx context.Context, _, _, _ int)
// body. The health check is left out: it answers 503 when its own 2-second
// ping runs out.
func TestRoutesAnswer504WhenDatabaseReadIsCutOff(t *testing.T) {
s := New(slowDB{}, streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
t.Parallel()
s := New(slowDB{}, streamer.New(logger.New(), metrics.New()), logger.New(),
&config.Config{})
tests := []struct {
path string
@@ -425,11 +512,15 @@ type slowSecondReadDB struct {
database.Store
}
func (slowSecondReadDB) GetASDetailsContext(_ context.Context, asn int) (*database.ASN, []database.LiveRoute, error) {
func (slowSecondReadDB) GetASDetailsContext(
_ context.Context, asn int,
) (*database.ASN, []database.LiveRoute, error) {
return &database.ASN{ASN: asn}, nil, nil
}
func (slowSecondReadDB) GetASPeersContext(ctx context.Context, _ int) ([]database.ASPeer, error) {
func (slowSecondReadDB) GetASPeersContext(
ctx context.Context, _ int,
) ([]database.ASPeer, error) {
<-ctx.Done()
return nil, ctx.Err()
@@ -439,7 +530,9 @@ func (slowSecondReadDB) GetStatsContext(_ context.Context) (database.Stats, erro
return database.Stats{}, nil
}
func (slowSecondReadDB) GetWHOISStats(ctx context.Context, _ time.Duration) (*database.WHOISStats, error) {
func (slowSecondReadDB) GetWHOISStats(
ctx context.Context, _ time.Duration,
) (*database.WHOISStats, error) {
<-ctx.Done()
return nil, ctx.Err()
@@ -462,7 +555,10 @@ func (idleASNFetcher) GetStats() ASNFetcherStats {
// message. Without the check after the second read, the AS page answers 504
// with an empty body and the stats routes answer 200 without the WHOIS stats.
func TestRoutesAnswer504WhenSecondDatabaseReadIsCutOff(t *testing.T) {
s := New(slowSecondReadDB{}, streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
t.Parallel()
s := New(slowSecondReadDB{}, streamer.New(logger.New(), metrics.New()),
logger.New(), &config.Config{})
s.SetASNFetcher(idleASNFetcher{})
tests := []struct {
@@ -482,27 +578,40 @@ func TestRoutesAnswer504WhenSecondDatabaseReadIsCutOff(t *testing.T) {
// checkRouteAnswers504 sends a request for path through the server's router
// with a short deadline and checks that it answers 504 with the time limit
// message: the JSON error body on JSON routes, the plain message on pages.
func checkRouteAnswers504(t *testing.T, s *Server, path string, wantJSONBody bool) {
func checkRouteAnswers504(
t *testing.T, s *Server, path string, wantJSONBody bool,
) {
t.Helper()
ctx, cancel := context.WithTimeout(context.Background(), shortRequestDeadline)
defer cancel()
rec := httptest.NewRecorder()
s.router.ServeHTTP(rec, httptest.NewRequest(http.MethodGet, path, nil).WithContext(ctx))
s.router.ServeHTTP(rec,
httptest.NewRequestWithContext(ctx, http.MethodGet, path, nil))
if rec.Code != http.StatusGatewayTimeout {
t.Errorf("%s: status %d, want %d; body %s", path, rec.Code, http.StatusGatewayTimeout, rec.Body)
t.Errorf("%s: status %d, want %d; body %s",
path, rec.Code, http.StatusGatewayTimeout, rec.Body)
}
if wantJSONBody {
checkJSONErrorCode(t, path, rec, http.StatusGatewayTimeout)
} else if body := strings.TrimSpace(rec.Body.String()); body != requestTimeoutMessage {
return
}
body := strings.TrimSpace(rec.Body.String())
if body != requestTimeoutMessage {
t.Errorf("%s: body %q, want %q", path, body, requestTimeoutMessage)
}
}
// checkJSONErrorCode checks that rec holds the JSON error body with the given
// code.
func checkJSONErrorCode(t *testing.T, name string, rec *httptest.ResponseRecorder, want int) {
func checkJSONErrorCode(
t *testing.T, name string, rec *httptest.ResponseRecorder, want int,
) {
t.Helper()
var body struct {
@@ -511,13 +620,17 @@ func checkJSONErrorCode(t *testing.T, name string, rec *httptest.ResponseRecorde
Code int `json:"code"`
} `json:"error"`
}
if err := json.Unmarshal(rec.Body.Bytes(), &body); err != nil {
err := json.Unmarshal(rec.Body.Bytes(), &body)
if err != nil {
t.Errorf("%s: decoding the answer %q: %v", name, rec.Body, err)
return
}
if body.Status != "error" || body.Error.Code != want {
t.Errorf("%s: answer %s, want status \"error\" and error code %d", name, rec.Body, want)
if body.Status != jsonStatusError || body.Error.Code != want {
t.Errorf("%s: answer %s, want status \"error\" and error code %d",
name, rec.Body, want)
}
}
@@ -525,12 +638,15 @@ func checkJSONErrorCode(t *testing.T, name string, rec *httptest.ResponseRecorde
// current count.
func settledGoroutineCount() int {
prev := runtime.NumGoroutine()
for range 20 {
time.Sleep(10 * time.Millisecond)
cur := runtime.NumGoroutine()
if cur == prev {
return cur
}
prev = cur
}
@@ -543,6 +659,7 @@ func waitForGoroutines(target int) bool {
if runtime.NumGoroutine() <= target {
return true
}
time.Sleep(10 * time.Millisecond)
}
internal/server/middleware.go
+23 -13
View File
@@ -15,6 +15,7 @@ import (
// responseWriter wraps http.ResponseWriter to capture the response
type responseWriter struct {
http.ResponseWriter
body *bytes.Buffer
statusCode int
written bool
@@ -87,8 +88,10 @@ func JSONResponseMiddleware(next http.Handler) http.Handler {
}
// Parse the original response
var originalResponse map[string]interface{}
if err := json.Unmarshal(rw.body.Bytes(), &originalResponse); err != nil {
var originalResponse map[string]any
err := json.Unmarshal(rw.body.Bytes(), &originalResponse)
if err != nil {
// If we can't parse it, just write original
w.WriteHeader(rw.statusCode)
_, _ = w.Write(rw.body.Bytes())
@@ -97,7 +100,7 @@ func JSONResponseMiddleware(next http.Handler) http.Handler {
}
// Add @meta field
originalResponse["@meta"] = map[string]interface{}{
originalResponse["@meta"] = map[string]any{
"time_zone": "UTC",
"api_version": 1,
"execution_time": fmt.Sprintf("%d ms", execTime.Milliseconds()),
@@ -160,8 +163,10 @@ func JSONValidationMiddleware(next http.Handler) http.Handler {
// For JSON responses, validate the JSON
if rw.body.Len() > 0 {
var testParse interface{}
if err := json.Unmarshal(rw.body.Bytes(), &testParse); err == nil {
var testParse any
err := json.Unmarshal(rw.body.Bytes(), &testParse)
if err == nil {
// Valid JSON, write it out
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(rw.statusCode)
@@ -180,9 +185,11 @@ func JSONValidationMiddleware(next http.Handler) http.Handler {
if statusCode == http.StatusOK {
statusCode = http.StatusInternalServerError
}
if deadlinePassed(r.Context()) {
statusCode = http.StatusGatewayTimeout
}
w.WriteHeader(statusCode)
errorMsg := "Internal server error"
@@ -190,11 +197,11 @@ func JSONValidationMiddleware(next http.Handler) http.Handler {
errorMsg = requestTimeoutMessage
}
response := map[string]interface{}{
"status": "error",
"error": map[string]interface{}{
"msg": errorMsg,
"code": statusCode,
response := map[string]any{
jsonKeyStatus: jsonStatusError,
jsonKeyError: map[string]any{
jsonKeyMsg: errorMsg,
jsonKeyCode: statusCode,
},
}
_ = json.NewEncoder(w).Encode(response)
@@ -204,6 +211,7 @@ func JSONValidationMiddleware(next http.Handler) http.Handler {
// statusWriter wraps http.ResponseWriter to capture the status code
type statusWriter struct {
http.ResponseWriter
statusCode int
written bool
}
@@ -213,6 +221,7 @@ func (sw *statusWriter) WriteHeader(statusCode int) {
sw.statusCode = statusCode
sw.written = true
}
sw.ResponseWriter.WriteHeader(statusCode)
}
@@ -252,11 +261,12 @@ func RequestLoggerMiddleware(logger *slog.Logger) func(http.Handler) http.Handle
// Slow query threshold (1 second)
const slowQueryThreshold = 1 * time.Second
if sw.statusCode >= http.StatusInternalServerError {
switch {
case sw.statusCode >= http.StatusInternalServerError:
logLevel = slog.LevelError
} else if sw.statusCode >= http.StatusBadRequest {
case sw.statusCode >= http.StatusBadRequest:
logLevel = slog.LevelWarn
} else if duration >= slowQueryThreshold {
case duration >= slowQueryThreshold:
logLevel = slog.LevelWarn
}
internal/server/routes.go
+4 -1
View File
@@ -27,7 +27,10 @@ func (s *Server) setupRoutes() {
r.Get("/", s.handleIndex())
r.Get("/status", s.handleStatusHTML())
r.Get("/status.json", JSONValidationMiddleware(s.handleStatusJSON()).ServeHTTP)
r.Get("/.well-known/healthcheck.json", JSONValidationMiddleware(s.handleHealthCheck()).ServeHTTP)
r.Get(
"/.well-known/healthcheck.json",
JSONValidationMiddleware(s.handleHealthCheck()).ServeHTTP,
)
// AS and prefix detail pages
r.Get("/as/{asn}", s.handleASDetail())
internal/server/server.go
+14 -5
View File
@@ -3,6 +3,7 @@ package server
import (
"context"
"errors"
"fmt"
"net"
"net/http"
@@ -42,7 +43,12 @@ type Server struct {
}
// New creates a new HTTP server
func New(db database.Store, streamer *streamer.Streamer, logger *logger.Logger, cfg *config.Config) *Server {
func New(
db database.Store,
streamer *streamer.Streamer,
logger *logger.Logger,
cfg *config.Config,
) *Server {
s := &Server{
db: db,
streamer: streamer,
@@ -57,8 +63,8 @@ func New(db database.Store, streamer *streamer.Streamer, logger *logger.Logger,
// Start listens on the configured port and serves HTTP in the background.
// It returns an error if it cannot listen, for example because the port is
// already in use.
func (s *Server) Start() error {
// already in use. ctx bounds only opening the port, not serving.
func (s *Server) Start(ctx context.Context) error {
const (
network = "tcp"
readHeaderTimeout = 40 * time.Second
@@ -78,7 +84,9 @@ func (s *Server) Start() error {
s.logger.Info("Starting HTTP server", "port", s.port, "addr", s.srv.Addr)
listener, err := net.Listen(network, s.srv.Addr)
var listenConfig net.ListenConfig
listener, err := listenConfig.Listen(ctx, network, s.srv.Addr)
if err != nil {
return fmt.Errorf("starting HTTP server: %w", err)
}
@@ -86,7 +94,8 @@ func (s *Server) Start() error {
s.logger.Info("HTTP server listening", "addr", listener.Addr().String())
go func() {
if err := s.srv.Serve(listener); err != nil && err != http.ErrServerClosed {
err := s.srv.Serve(listener)
if err != nil && !errors.Is(err, http.ErrServerClosed) {
s.logger.Error("HTTP server error", "error", err)
}
}()
internal/server/server_test.go
+20 -6
View File
@@ -1,4 +1,4 @@
package server
package server_test
import (
"context"
@@ -8,6 +8,7 @@ import (
"git.eeqj.de/sneak/routewatch/internal/config"
"git.eeqj.de/sneak/routewatch/internal/logger"
"git.eeqj.de/sneak/routewatch/internal/metrics"
"git.eeqj.de/sneak/routewatch/internal/server"
"git.eeqj.de/sneak/routewatch/internal/streamer"
)
@@ -16,18 +17,31 @@ import (
// listened in a goroutine, only logged the failure and returned nil, so the
// daemon kept running without its web server.
func TestStartFailsWhenPortIsHeld(t *testing.T) {
held, err := net.Listen("tcp", ":0")
t.Parallel()
var listenConfig net.ListenConfig
held, err := listenConfig.Listen(t.Context(), "tcp", ":0")
if err != nil {
t.Fatalf("failed to open the test listener: %v", err)
}
defer func() { _ = held.Close() }()
port := held.Addr().(*net.TCPAddr).Port
addr, ok := held.Addr().(*net.TCPAddr)
if !ok {
t.Fatalf("test listener address %v is not a TCP address", held.Addr())
}
port := addr.Port
log := logger.New()
s := New(nil, streamer.New(log, metrics.New()), log, &config.Config{Port: port})
s := server.New(nil, streamer.New(log, metrics.New()), log,
&config.Config{Port: port})
defer func() { _ = s.Stop(context.Background()) }()
if err := s.Start(); err == nil {
t.Fatalf("Start returned nil while port %d was held by another listener", port)
err = s.Start(t.Context())
if err == nil {
t.Fatalf("Start returned nil while port %d was held by another listener",
port)
}
}
internal/streamer/streamer.go
+292 -213
View File
@@ -1,5 +1,6 @@
// Package streamer implements an HTTP client that connects to the RIPE RIS Live streaming API,
// parses BGP UPDATE messages from the JSON stream, and dispatches them to registered handlers.
// Package streamer implements an HTTP client that connects to the RIPE RIS Live
// streaming API, parses BGP UPDATE messages from the JSON stream, and
// dispatches them to registered handlers.
package streamer
import (
@@ -25,50 +26,60 @@ import (
// countingReader wraps an io.Reader and counts bytes read
type countingReader struct {
reader io.Reader
count int64
count atomic.Int64
}
// Read implements io.Reader and counts bytes
func (c *countingReader) Read(p []byte) (int, error) {
n, err := c.reader.Read(p)
atomic.AddInt64(&c.count, int64(n))
c.count.Add(int64(n))
return n, err
}
// Count returns the total bytes read
func (c *countingReader) Count() int64 {
return atomic.LoadInt64(&c.count)
return c.count.Load()
}
// Configuration constants for the RIS Live streamer.
const (
risLiveURL = "https://ris-live.ripe.net/v1/stream/?format=json&" +
"client=https%3A%2F%2Fgit.eeqj.de%2Fsneak%2Froutewatch"
metricsWindowSize = 60 // seconds for rolling average
metricsUpdateRate = time.Second
minBackoffDelay = 5 * time.Second
maxBackoffDelay = 320 * time.Second
streamIdleTimeout = 60 * time.Second // Reconnect when no line arrives for this long
metricsLogInterval = 10 * time.Second
bytesPerKB = 1024
bytesPerMB = 1024 * 1024
maxLineSize = bytesPerMB // Well above an extended BGP message (RFC 8654) in hex plus its prefixes
maxConcurrentHandlers = 800 // Maximum number of concurrent message handlers
metricsWindowSize = 60 // seconds for rolling average
metricsUpdateRate = time.Second
minBackoffDelay = 5 * time.Second
maxBackoffDelay = 320 * time.Second
// Reconnect when no line arrives for this long
streamIdleTimeout = 60 * time.Second
metricsLogInterval = 10 * time.Second
bytesPerKB = 1024
bytesPerMB = 1024 * 1024
// Well above an extended BGP message (RFC 8654) in hex plus its prefixes
maxLineSize = bytesPerMB
maxConcurrentHandlers = 800 // Maximum number of concurrent message handlers
// Backpressure constants
backpressureThreshold = 0.5 // Start dropping at 50% queue utilization
backpressureSlope = 2.0 // Slope for linear drop probability increase
)
// errStreamEnded is returned by stream when RIS Live ends the response.
var errStreamEnded = errors.New("RIS Live ended the stream")
var (
errAlreadyRunning = errors.New("streamer already running")
errUnexpectedStatus = errors.New("unexpected status code")
// errStreamEnded is returned by stream when RIS Live ends the response.
errStreamEnded = errors.New("RIS Live ended the stream")
// errStreamIdle is returned by stream when no line arrived for the idle
// timeout.
errStreamIdle = errors.New("no line received")
)
// MessageHandler defines the interface for processing RIS messages.
// Implementations must specify which message types they want to receive,
// how to process messages, and their desired queue capacity.
type MessageHandler interface {
// WantsMessage returns true if this handler wants to process messages of the given type.
// WantsMessage returns true if this handler wants to process messages of
// the given type.
WantsMessage(messageType string) bool
// HandleMessage processes a RIS message. This method is called from a dedicated
@@ -114,18 +125,20 @@ type handlerInfo struct {
// backoff, dispatches messages to registered handlers via per-handler queues, and
// implements backpressure to prevent queue overflow during high traffic periods.
type Streamer struct {
logger *logger.Logger
client *http.Client
url string
idleTimeout time.Duration // Drop the connection when no line arrives for this long
minBackoff time.Duration // First wait before reconnecting; doubles up to maxBackoffDelay
logger *logger.Logger
client *http.Client
url string
// Drop the connection when no line arrives for this long
idleTimeout time.Duration
// First wait before reconnecting; doubles up to maxBackoffDelay
minBackoff time.Duration
handlers []*handlerInfo
rawHandler RawMessageHandler
mu sync.RWMutex
cancel context.CancelFunc
running bool
metrics *metrics.Tracker
totalDropped uint64 // Total dropped messages across all handlers
totalDropped atomic.Uint64 // Total dropped messages across all handlers
random *rand.Rand // Random number generator for backpressure drops
bgpPeers map[string]bool // Track active BGP peers by peer IP
bgpPeersMu sync.RWMutex // Protects bgpPeers map
@@ -133,7 +146,8 @@ type Streamer struct {
// New creates a new Streamer instance configured to connect to the RIS Live API.
// The logger is used for structured logging of connection events and errors.
// The metrics tracker is used to record message counts, bytes received, and connection status.
// The metrics tracker is used to record message counts, bytes received, and
// connection status.
func New(logger *logger.Logger, metrics *metrics.Tracker) *Streamer {
return &Streamer{
logger: logger,
@@ -167,7 +181,8 @@ func (s *Streamer) RegisterHandler(handler MessageHandler) {
handler: handler,
queue: make(chan *ristypes.RISMessage, handler.QueueCapacity()),
metrics: handlerMetrics{
minTime: time.Duration(math.MaxInt64), // Initialize to max so first value sets the floor
// Initialize to max so first value sets the floor
minTime: time.Duration(math.MaxInt64),
},
}
@@ -185,22 +200,24 @@ func (s *Streamer) RegisterHandler(handler MessageHandler) {
func (s *Streamer) RegisterRawHandler(handler RawMessageHandler) {
s.mu.Lock()
defer s.mu.Unlock()
s.rawHandler = handler
}
// Start begins streaming BGP updates from the RIS Live API in a background goroutine.
// It starts worker goroutines for each registered handler and manages automatic
// reconnection with exponential backoff on connection failures.
// Returns an error if the streamer is already running.
func (s *Streamer) Start() error {
// Returns an error if the streamer is already running. It streams until Stop:
// the end of ctx does not stop it.
func (s *Streamer) Start(ctx context.Context) error {
s.mu.Lock()
defer s.mu.Unlock()
if s.running {
return fmt.Errorf("streamer already running")
return errAlreadyRunning
}
ctx, cancel := context.WithCancel(context.Background())
ctx, cancel := context.WithCancel(context.WithoutCancel(ctx))
s.cancel = cancel
s.running = true
@@ -230,44 +247,21 @@ func (s *Streamer) Stop() {
return
}
s.cancel()
s.cancel = nil
// Close all handler queues to signal workers to stop
for _, info := range s.handlers {
close(info.queue)
}
s.running = false
s.mu.Unlock()
s.metrics.SetConnected(false)
}
// runHandlerWorker processes messages for a specific handler
func (s *Streamer) runHandlerWorker(info *handlerInfo) {
for msg := range info.queue {
start := time.Now()
info.handler.HandleMessage(msg)
elapsed := time.Since(start)
// Update metrics
info.metrics.mu.Lock()
info.metrics.processedCount++
info.metrics.totalTime += elapsed
// Update min time
if elapsed < info.metrics.minTime {
info.metrics.minTime = elapsed
}
// Update max time
if elapsed > info.metrics.maxTime {
info.metrics.maxTime = elapsed
}
info.metrics.mu.Unlock()
}
}
// IsRunning reports whether the streamer is currently connected and processing messages.
// This is safe to call concurrently from multiple goroutines.
// IsRunning reports whether the streamer is currently connected and processing
// messages. This is safe to call concurrently from multiple goroutines.
func (s *Streamer) IsRunning() bool {
s.mu.RLock()
defer s.mu.RUnlock()
@@ -328,12 +322,15 @@ func (s *Streamer) GetHandlerStats() []HandlerStats {
// Calculate average time
if info.metrics.processedCount > 0 {
processedCount := info.metrics.processedCount
const maxInt64 = 1<<63 - 1
if processedCount > maxInt64 {
processedCount = maxInt64
}
//nolint:gosec // processedCount is explicitly bounded above
hs.AvgProcessTime = info.metrics.totalTime / time.Duration(processedCount)
hs.AvgProcessTime = info.metrics.totalTime /
time.Duration(processedCount)
}
info.metrics.mu.Unlock()
@@ -344,11 +341,38 @@ func (s *Streamer) GetHandlerStats() []HandlerStats {
return stats
}
// GetDroppedMessages returns the total number of messages dropped across all handlers
// due to queue overflow or backpressure. This counter is monotonically increasing
// and is safe to call concurrently.
// GetDroppedMessages returns the total number of messages dropped across all
// handlers due to queue overflow or backpressure. This counter is monotonically
// increasing and is safe to call concurrently.
func (s *Streamer) GetDroppedMessages() uint64 {
return atomic.LoadUint64(&s.totalDropped)
return s.totalDropped.Load()
}
// runHandlerWorker processes messages for a specific handler
func (s *Streamer) runHandlerWorker(info *handlerInfo) {
for msg := range info.queue {
start := time.Now()
info.handler.HandleMessage(msg)
elapsed := time.Since(start)
// Update metrics
info.metrics.mu.Lock()
info.metrics.processedCount++
info.metrics.totalTime += elapsed
// Update min time
if elapsed < info.metrics.minTime {
info.metrics.minTime = elapsed
}
// Update max time
if elapsed > info.metrics.maxTime {
info.metrics.maxTime = elapsed
}
info.metrics.mu.Unlock()
}
}
// logMetrics logs the current streaming statistics
@@ -357,7 +381,8 @@ func (s *Streamer) logMetrics() {
uptime := time.Since(metrics.ConnectedSince)
const bitsPerMegabit = 1000000
totalDropped := atomic.LoadUint64(&s.totalDropped)
totalDropped := s.totalDropped.Load()
s.logger.Info(
"Stream statistics",
@@ -383,16 +408,19 @@ func (s *Streamer) logMetrics() {
// Log per-handler statistics
s.mu.RLock()
for i, info := range s.handlers {
info.metrics.mu.Lock()
if info.metrics.processedCount > 0 {
// Safe conversion: processedCount is bounded by maxInt64
processedCount := info.metrics.processedCount
const maxInt64 = 1<<63 - 1
if processedCount > maxInt64 {
processedCount = maxInt64
}
//nolint:gosec // processedCount is explicitly bounded above
avgTime := info.metrics.totalTime / time.Duration(processedCount)
s.logger.Info(
"Handler statistics",
@@ -409,6 +437,7 @@ func (s *Streamer) logMetrics() {
}
info.metrics.mu.Unlock()
}
s.mu.RUnlock()
}
@@ -417,7 +446,8 @@ func (s *Streamer) updateMetrics(messageBytes int) {
s.metrics.RecordMessage(int64(messageBytes))
}
// streamWithReconnect handles streaming with automatic reconnection and exponential backoff
// streamWithReconnect handles streaming with automatic reconnection and
// exponential backoff
func (s *Streamer) streamWithReconnect(ctx context.Context) {
backoffDelay := s.minBackoff
consecutiveFailures := 0
@@ -455,6 +485,7 @@ func (s *Streamer) streamWithReconnect(ctx context.Context) {
if streamDuration > 30*time.Second {
s.logger.Info("Resetting backoff delay due to successful connection",
"stream_duration", streamDuration)
backoffDelay = s.minBackoff
consecutiveFailures = 0
} else {
@@ -494,7 +525,7 @@ func (s *Streamer) stream(ctx context.Context) error {
idleTimer := time.AfterFunc(s.idleTimeout, connCancel)
defer idleTimer.Stop()
req, err := http.NewRequestWithContext(connCtx, "GET", s.url, nil)
req, err := http.NewRequestWithContext(connCtx, http.MethodGet, s.url, nil)
if err != nil {
return fmt.Errorf("failed to create request: %w", err)
}
@@ -507,13 +538,14 @@ func (s *Streamer) stream(ctx context.Context) error {
return fmt.Errorf("failed to connect to RIS Live: %w", err)
}
defer func() {
if err := resp.Body.Close(); err != nil {
err := resp.Body.Close()
if err != nil {
s.logger.Error("Failed to close response body", "error", err)
}
}()
if resp.StatusCode != http.StatusOK {
return fmt.Errorf("unexpected status code: %d", resp.StatusCode)
return fmt.Errorf("%w: %d", errUnexpectedStatus, resp.StatusCode)
}
// Wrap body with counting reader to track actual wire bytes
@@ -527,7 +559,9 @@ func (s *Streamer) stream(ctx context.Context) error {
return fmt.Errorf("failed to create gzip reader: %w", err)
}
defer func() { _ = gzReader.Close() }()
reader = gzReader
s.logger.Info("Connected to RIS Live stream", "compression", "gzip")
} else {
s.logger.Info("Connected to RIS Live stream", "compression", "none")
@@ -535,9 +569,6 @@ func (s *Streamer) stream(ctx context.Context) error {
s.metrics.SetConnected(true)
// Track wire bytes for metrics updates
var lastWireBytes int64
// Start metrics logging goroutine
metricsTicker := time.NewTicker(metricsLogInterval)
defer metricsTicker.Stop()
@@ -554,25 +585,44 @@ func (s *Streamer) stream(ctx context.Context) error {
}()
// Wire byte update ticker - update metrics with actual wire bytes periodically
go s.recordWireBytes(connCtx, wireCounter)
return s.readStream(ctx, connCtx, reader, idleTimer)
}
// recordWireBytes adds the bytes wireCounter has counted to the metrics once a
// second until ctx is cancelled.
func (s *Streamer) recordWireBytes(
ctx context.Context, wireCounter *countingReader,
) {
wireUpdateTicker := time.NewTicker(time.Second)
defer wireUpdateTicker.Stop()
go func() {
for {
select {
case <-wireUpdateTicker.C:
currentBytes := wireCounter.Count()
delta := currentBytes - lastWireBytes
if delta > 0 {
s.metrics.RecordWireBytes(delta)
lastWireBytes = currentBytes
}
case <-connCtx.Done():
return
}
}
}()
var lastWireBytes int64
for {
select {
case <-wireUpdateTicker.C:
currentBytes := wireCounter.Count()
delta := currentBytes - lastWireBytes
if delta > 0 {
s.metrics.RecordWireBytes(delta)
lastWireBytes = currentBytes
}
case <-ctx.Done():
return
}
}
}
// readStream hands each line of the stream to processLine until the stream
// ends, a line fails, or ctx is cancelled. Each line restarts idleTimer, which
// cancels connCtx when it fires.
func (s *Streamer) readStream(
ctx, connCtx context.Context, reader io.Reader, idleTimer *time.Timer,
) error {
scanner := bufio.NewScanner(reader)
scanner.Buffer(nil, maxLineSize)
@@ -592,129 +642,10 @@ func (s *Streamer) stream(ctx context.Context) error {
continue
}
// Update metrics with decompressed message size
s.updateMetrics(len(line))
// Call raw handler if registered
s.mu.RLock()
rawHandler := s.rawHandler
s.mu.RUnlock()
if rawHandler != nil {
// Call raw handler synchronously to preserve order
rawHandler(string(line))
err := s.processLine(ctx, line)
if err != nil {
return err
}
// Parse the message first
var wrapper ristypes.RISLiveMessage
if err := json.Unmarshal(line, &wrapper); err != nil {
// Log the error and return to trigger reconnection
s.logger.Error("Failed to parse JSON",
"error", err,
"line", string(line),
"line_length", len(line))
return fmt.Errorf("JSON parse error: %w", err)
}
// Check if it's a ris_message wrapper
if wrapper.Type != "ris_message" {
s.logger.Error("Unexpected wrapper type",
"type", wrapper.Type,
"line", string(line),
)
continue
}
// Get the actual message
msg := wrapper.Data
// Parse the timestamp
msg.ParsedTimestamp = time.Unix(int64(msg.Timestamp), 0).UTC()
// Process based on message type
switch msg.Type {
case ristypes.MessageTypeUpdate:
// Process BGP UPDATE messages
// Will be dispatched to handlers
case "RIS_PEER_STATE":
// RIS peer state messages - silently ignore
continue
case "KEEPALIVE":
// BGP keepalive messages - silently process
continue
case "OPEN":
// BGP open messages - track peer as active
s.bgpPeersMu.Lock()
s.bgpPeers[msg.Peer] = true
peerCount := len(s.bgpPeers)
s.bgpPeersMu.Unlock()
s.metrics.SetBGPPeerCount(peerCount)
s.logger.Info("BGP session opened",
"peer", msg.Peer,
"peer_asn", msg.PeerASN,
"total_peers", peerCount,
)
continue
case "NOTIFICATION":
// BGP notification messages (session closed)
s.bgpPeersMu.Lock()
delete(s.bgpPeers, msg.Peer)
peerCount := len(s.bgpPeers)
s.bgpPeersMu.Unlock()
s.metrics.SetBGPPeerCount(peerCount)
s.logger.Warn("BGP notification",
"peer", msg.Peer,
"peer_asn", msg.PeerASN,
"total_peers", peerCount,
)
continue
case ristypes.MessageTypeState:
// The BGP session with the peer changed state, for example went
// down. Dispatched to the handlers that want it.
default:
s.logger.Warn("Unknown message type, skipping",
"type", msg.Type,
)
continue
}
// Dispatch to interested handlers. Stop cancels ctx and closes the
// queues under the write lock, so if ctx is cancelled here, under the
// read lock, the queues are closed and must not be sent to.
s.mu.RLock()
if ctx.Err() != nil {
s.mu.RUnlock()
return ctx.Err()
}
for _, info := range s.handlers {
if !info.handler.WantsMessage(msg.Type) {
continue
}
if !s.queueMessage(info, &msg) {
atomic.AddUint64(&info.metrics.droppedCount, 1)
atomic.AddUint64(&s.totalDropped, 1)
continue
}
// Update high water mark if needed
queueLen := len(info.queue)
info.metrics.mu.Lock()
if queueLen > info.metrics.queueHighWaterMark {
info.metrics.queueHighWaterMark = queueLen
}
info.metrics.mu.Unlock()
}
s.mu.RUnlock()
}
if ctx.Err() != nil {
@@ -722,16 +653,161 @@ func (s *Streamer) stream(ctx context.Context) error {
}
if connCtx.Err() != nil {
return fmt.Errorf("no line received for %s", s.idleTimeout)
return fmt.Errorf("%w for %s", errStreamIdle, s.idleTimeout)
}
if err := scanner.Err(); err != nil {
err := scanner.Err()
if err != nil {
return fmt.Errorf("scanner error: %w", err)
}
return errStreamEnded
}
// processLine passes one line of the stream to the raw handler, parses it,
// and dispatches an UPDATE or STATE message to the handlers. An error ends the
// connection.
func (s *Streamer) processLine(ctx context.Context, line []byte) error {
// Update metrics with decompressed message size
s.updateMetrics(len(line))
// Call raw handler if registered
s.mu.RLock()
rawHandler := s.rawHandler
s.mu.RUnlock()
if rawHandler != nil {
// Call raw handler synchronously to preserve order
rawHandler(string(line))
}
// Parse the message first
var wrapper ristypes.RISLiveMessage
err := json.Unmarshal(line, &wrapper)
if err != nil {
// Log the error and return to trigger reconnection
s.logger.Error("Failed to parse JSON",
"error", err,
"line", string(line),
"line_length", len(line))
return fmt.Errorf("JSON parse error: %w", err)
}
// Check if it's a ris_message wrapper
if wrapper.Type != "ris_message" {
s.logger.Error("Unexpected wrapper type",
"type", wrapper.Type,
"line", string(line),
)
return nil
}
// Get the actual message
msg := wrapper.Data
// Parse the timestamp
msg.ParsedTimestamp = time.Unix(int64(msg.Timestamp), 0).UTC()
// Process based on message type
switch msg.Type {
case ristypes.MessageTypeUpdate:
// Process BGP UPDATE messages
// Will be dispatched to handlers
case "RIS_PEER_STATE":
// RIS peer state messages - silently ignore
return nil
case "KEEPALIVE":
// BGP keepalive messages - silently process
return nil
case "OPEN":
// BGP open messages - track peer as active
s.bgpPeersMu.Lock()
s.bgpPeers[msg.Peer] = true
peerCount := len(s.bgpPeers)
s.bgpPeersMu.Unlock()
s.metrics.SetBGPPeerCount(peerCount)
s.logger.Info("BGP session opened",
"peer", msg.Peer,
"peer_asn", msg.PeerASN,
"total_peers", peerCount,
)
return nil
case "NOTIFICATION":
// BGP notification messages (session closed)
s.bgpPeersMu.Lock()
delete(s.bgpPeers, msg.Peer)
peerCount := len(s.bgpPeers)
s.bgpPeersMu.Unlock()
s.metrics.SetBGPPeerCount(peerCount)
s.logger.Warn("BGP notification",
"peer", msg.Peer,
"peer_asn", msg.PeerASN,
"total_peers", peerCount,
)
return nil
case ristypes.MessageTypeState:
// The BGP session with the peer changed state, for example went
// down. Dispatched to the handlers that want it.
default:
s.logger.Warn("Unknown message type, skipping",
"type", msg.Type,
)
return nil
}
return s.dispatchMessage(ctx, &msg)
}
// dispatchMessage queues msg for each handler that wants it, unless
// queueMessage drops it for that handler.
func (s *Streamer) dispatchMessage(
ctx context.Context, msg *ristypes.RISMessage,
) error {
// Stop cancels ctx and closes the queues under the write lock, so if ctx
// is cancelled here, under the read lock, the queues are closed and must
// not be sent to.
s.mu.RLock()
if ctx.Err() != nil {
s.mu.RUnlock()
return ctx.Err()
}
for _, info := range s.handlers {
if !info.handler.WantsMessage(msg.Type) {
continue
}
if !s.queueMessage(info, msg) {
atomic.AddUint64(&info.metrics.droppedCount, 1)
s.totalDropped.Add(1)
continue
}
// Update high water mark if needed
queueLen := len(info.queue)
info.metrics.mu.Lock()
if queueLen > info.metrics.queueHighWaterMark {
info.metrics.queueHighWaterMark = queueLen
}
info.metrics.mu.Unlock()
}
s.mu.RUnlock()
return nil
}
// queueMessage puts msg on the handler's queue and reports false if it dropped
// msg instead. It never waits for room, so a full queue holds up neither the
// read loop nor Stop.
@@ -746,7 +822,9 @@ func (s *Streamer) stream(ctx context.Context) error {
//
// The held list stays small: a STATE message is a few hundred bytes and the
// feed sends about ten a second, so a queue full for a minute holds about 600.
func (s *Streamer) queueMessage(info *handlerInfo, msg *ristypes.RISMessage) bool {
func (s *Streamer) queueMessage(
info *handlerInfo, msg *ristypes.RISMessage,
) bool {
for len(info.heldState) > 0 && trySend(info.queue, info.heldState[0]) {
info.heldState = info.heldState[1:]
}
@@ -776,7 +854,8 @@ func trySend(queue chan *ristypes.RISMessage, msg *ristypes.RISMessage) bool {
}
}
// shouldDropForBackpressure determines if a message should be dropped based on queue utilization
// shouldDropForBackpressure determines if a message should be dropped based on
// queue utilization
func (s *Streamer) shouldDropForBackpressure(info *handlerInfo) bool {
// Calculate queue utilization
queueLen := len(info.queue)
internal/streamer/streamer_test.go → internal/streamer/streamer_internal_test.go
+177 -70
View File
@@ -18,6 +18,8 @@ import (
)
func TestNewStreamer(t *testing.T) {
t.Parallel()
logger := logger.New()
metricsTracker := metrics.New()
s := New(logger, metricsTracker)
@@ -56,10 +58,13 @@ func TestNewStreamer(t *testing.T) {
// goroutine count returns to its starting value. Each connection starts two
// ticker goroutines; before the fix they lived until the streamer's lifetime
// context was cancelled, so every reconnect leaked two.
//
//nolint:paralleltest // counts goroutines, so no other test may run alongside
func TestStreamDoesNotLeakTickersAcrossReconnects(t *testing.T) {
// The handler returns immediately, so the response body is empty and each
// stream call ends at once, standing in for a dropped connection.
srv := httptest.NewServer(http.HandlerFunc(func(_ http.ResponseWriter, _ *http.Request) {}))
srv := httptest.NewServer(http.HandlerFunc(
func(_ http.ResponseWriter, _ *http.Request) {}))
defer srv.Close()
s := New(logger.New(), metrics.New())
@@ -67,19 +72,23 @@ func TestStreamDoesNotLeakTickersAcrossReconnects(t *testing.T) {
// One warm-up connection so any persistent HTTP transport goroutine exists
// before we take the baseline.
if err := s.stream(context.Background()); !errors.Is(err, errStreamEnded) {
err := s.stream(context.Background())
if !errors.Is(err, errStreamEnded) {
t.Fatalf("warm-up stream returned %v, want %v", err, errStreamEnded)
}
s.client.CloseIdleConnections()
baseline := settledGoroutineCount()
const reconnects = 20
for range reconnects {
if err := s.stream(context.Background()); !errors.Is(err, errStreamEnded) {
err := s.stream(context.Background())
if !errors.Is(err, errStreamEnded) {
t.Fatalf("stream returned %v, want %v", err, errStreamEnded)
}
}
s.client.CloseIdleConnections()
if !waitForGoroutines(baseline) {
@@ -89,8 +98,8 @@ func TestStreamDoesNotLeakTickersAcrossReconnects(t *testing.T) {
}
// updateLine is one UPDATE message as RIS Live sends it.
const updateLine = `{"type":"ris_message","data":{"type":"UPDATE","peer":"192.0.2.1",` +
`"peer_asn":"64496","timestamp":1700000000}}` + "\n"
const updateLine = `{"type":"ris_message","data":{"type":"UPDATE",` +
`"peer":"192.0.2.1","peer_asn":"64496","timestamp":1700000000}}` + "\n"
// updateHandler wants UPDATE messages. For each one it handles it sends on
// handled if there is room, and otherwise does nothing; a nil handled is
@@ -99,8 +108,11 @@ type updateHandler struct {
handled chan struct{}
}
func (updateHandler) WantsMessage(messageType string) bool { return messageType == "UPDATE" }
func (updateHandler) QueueCapacity() int { return 10 }
func (updateHandler) WantsMessage(messageType string) bool {
return messageType == "UPDATE"
}
func (updateHandler) QueueCapacity() int { return 10 }
func (h updateHandler) HandleMessage(*ristypes.RISMessage) {
select {
@@ -109,11 +121,21 @@ func (h updateHandler) HandleMessage(*ristypes.RISMessage) {
}
}
// flush sends what w has buffered to the client. The writers httptest servers
// hand to handlers can always do that.
func flush(w http.ResponseWriter) {
flusher, ok := w.(http.Flusher)
if ok {
flusher.Flush()
}
}
// stateDownLine is the STATE message RIS Live sends when the collector's BGP
// session with the peer stateDownPeer goes down.
const (
stateDownPeer = "192.0.2.1"
stateDownLine = `{"type":"ris_message","data":{"type":"STATE","peer":"` + stateDownPeer + `",` +
stateDownLine = `{"type":"ris_message","data":{"type":"STATE",` +
`"peer":"` + stateDownPeer + `",` +
`"peer_asn":"64496","timestamp":1700000000,"state":"down"}}` + "\n"
)
@@ -124,31 +146,39 @@ type stateHandler struct{}
func (stateHandler) WantsMessage(messageType string) bool {
return messageType == ristypes.MessageTypeState
}
func (stateHandler) QueueCapacity() int { return 10 }
func (stateHandler) QueueCapacity() int { return 10 }
func (stateHandler) HandleMessage(*ristypes.RISMessage) {}
// TestStreamHandsStateMessagesToHandlersThatWantThem has the server send one
// STATE message whose state is down. The streamer must queue it, with its
// peer and state, for a handler that wants STATE messages.
func TestStreamHandsStateMessagesToHandlersThatWantThem(t *testing.T) {
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, stateDownLine)
}))
t.Parallel()
srv := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, stateDownLine)
}))
defer srv.Close()
s := New(logger.New(), metrics.New())
s.url = srv.URL
s.RegisterHandler(stateHandler{})
if err := s.stream(context.Background()); !errors.Is(err, errStreamEnded) {
err := s.stream(context.Background())
if !errors.Is(err, errStreamEnded) {
t.Fatalf("stream returned %v, want %v", err, errStreamEnded)
}
select {
case msg := <-s.handlers[0].queue:
if msg.Type != ristypes.MessageTypeState || msg.Peer != stateDownPeer || msg.State != ristypes.StateDown {
t.Errorf("queued message has type %q, peer %q, state %q; want %q, %q, %q",
msg.Type, msg.Peer, msg.State, ristypes.MessageTypeState, stateDownPeer, ristypes.StateDown)
if msg.Type != ristypes.MessageTypeState || msg.Peer != stateDownPeer ||
msg.State != ristypes.StateDown {
t.Errorf("queued message has type %q, peer %q, state %q; "+
"want %q, %q, %q", msg.Type, msg.Peer, msg.State,
ristypes.MessageTypeState, stateDownPeer, ristypes.StateDown)
}
default:
t.Fatal("the STATE message was not queued for the handler that wants it")
@@ -161,26 +191,37 @@ func TestStreamHandsStateMessagesToHandlersThatWantThem(t *testing.T) {
// everything queued before it and before anything handed over after it. The
// update handed over while a STATE message is held is dropped.
func TestHeldStateMessagesKeepTheirPlaceInTheQueue(t *testing.T) {
t.Parallel()
s := New(logger.New(), metrics.New())
s.RegisterHandler(stateHandler{})
info := s.handlers[0]
before := &ristypes.RISMessage{Type: ristypes.MessageTypeUpdate}
for len(info.queue) < cap(info.queue) {
info.queue <- before
}
firstDown := &ristypes.RISMessage{Type: ristypes.MessageTypeState, State: ristypes.StateDown}
secondDown := &ristypes.RISMessage{Type: ristypes.MessageTypeState, State: ristypes.StateDown}
firstDown := &ristypes.RISMessage{
Type: ristypes.MessageTypeState, State: ristypes.StateDown,
}
secondDown := &ristypes.RISMessage{
Type: ristypes.MessageTypeState, State: ristypes.StateDown,
}
updateWhileHeld := &ristypes.RISMessage{Type: ristypes.MessageTypeUpdate}
after := &ristypes.RISMessage{Type: ristypes.MessageTypeUpdate}
if !s.queueMessage(info, firstDown) {
t.Fatal("the first STATE message was dropped on a full queue")
}
<-info.queue // the handler makes room for one message
if !s.queueMessage(info, secondDown) {
t.Fatal("the second STATE message was dropped")
}
if s.queueMessage(info, updateWhileHeld) {
t.Fatal("an update was queued while a STATE message was held")
}
@@ -191,6 +232,7 @@ func TestHeldStateMessagesKeepTheirPlaceInTheQueue(t *testing.T) {
t.Fatal("a message overtook those queued before the first STATE message")
}
}
if !s.queueMessage(info, after) {
t.Fatal("an update was dropped on a queue with room")
}
@@ -198,12 +240,16 @@ func TestHeldStateMessagesKeepTheirPlaceInTheQueue(t *testing.T) {
if <-info.queue != firstDown {
t.Fatal("the first STATE message did not come next")
}
if <-info.queue != secondDown {
t.Fatal("the second STATE message did not come after the first")
}
if <-info.queue != after {
t.Fatal("the update handed over after the STATE messages did not come after them")
t.Fatal("the update handed over after the STATE messages did not come " +
"after them")
}
if len(info.queue) != 0 {
t.Errorf("%d messages left on the queue, want none", len(info.queue))
}
@@ -215,6 +261,7 @@ func newStreamerWithFullStateQueue(url string) *Streamer {
s := New(logger.New(), metrics.New())
s.url = url
s.RegisterHandler(stateHandler{})
queue := s.handlers[0].queue
for len(queue) < cap(queue) {
queue <- &ristypes.RISMessage{Type: ristypes.MessageTypeState}
@@ -229,24 +276,33 @@ func newStreamerWithFullStateQueue(url string) *Streamer {
// for that handler, and the read loop must go on to queue the UPDATE for the
// handler that wants it.
func TestOtherHandlersKeepReceivingWhileAStateMessageIsHeld(t *testing.T) {
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, stateDownLine+updateLine)
}))
t.Parallel()
srv := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, stateDownLine+updateLine)
}))
defer srv.Close()
s := newStreamerWithFullStateQueue(srv.URL)
s.RegisterHandler(updateHandler{})
if err := s.stream(context.Background()); !errors.Is(err, errStreamEnded) {
err := s.stream(context.Background())
if !errors.Is(err, errStreamEnded) {
t.Fatalf("stream returned %v, want %v", err, errStreamEnded)
}
held := s.handlers[0].heldState
if len(held) != 1 || held[0].Peer != stateDownPeer || held[0].State != ristypes.StateDown {
t.Errorf("held %d STATE messages, want the down from %s alone", len(held), stateDownPeer)
if len(held) != 1 || held[0].Peer != stateDownPeer ||
held[0].State != ristypes.StateDown {
t.Errorf("held %d STATE messages, want the down from %s alone",
len(held), stateDownPeer)
}
if got := len(s.handlers[1].queue); got != 1 {
t.Errorf("%d messages queued for the handler that wants UPDATE messages, want 1", got)
got := len(s.handlers[1].queue)
if got != 1 {
t.Errorf("%d messages queued for the handler that wants UPDATE "+
"messages, want 1", got)
}
}
@@ -254,21 +310,27 @@ func TestOtherHandlersKeepReceivingWhileAStateMessageIsHeld(t *testing.T) {
// messages and hold the response open while the queue of the handler that
// wants them is full. Stop must return at once and end the stream.
func TestStopReturnsAtOnceWithStateMessagesHeld(t *testing.T) {
t.Parallel()
const (
lines = 2
maxWait = 5 * time.Second
)
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
_, _ = io.WriteString(w, stateDownLine+stateDownLine)
w.(http.Flusher).Flush()
<-r.Context().Done()
}))
srv := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, r *http.Request) {
_, _ = io.WriteString(w, stateDownLine+stateDownLine)
flush(w)
<-r.Context().Done()
}))
defer srv.Close()
s := newStreamerWithFullStateQueue(srv.URL)
// The raw handler sees each line before the read loop hands it to the
// queues, so once it has seen the second line, the first is held.
linesSeen := make(chan struct{}, lines)
s.RegisterRawHandler(func(string) { linesSeen <- struct{}{} })
// As in TestStopBeforeMessageReachesQueues, setting cancel as Start does
@@ -276,25 +338,30 @@ func TestStopReturnsAtOnceWithStateMessagesHeld(t *testing.T) {
ctx, cancel := context.WithCancel(context.Background())
s.cancel = cancel
streamErr := make(chan error, 1)
go func() { streamErr <- s.stream(ctx) }()
for range lines {
select {
case <-linesSeen:
case <-time.After(maxWait):
t.Fatalf("the read loop did not read %d lines within %s", lines, maxWait)
t.Fatalf("the read loop did not read %d lines within %s",
lines, maxWait)
}
}
stopped := make(chan struct{})
go func() {
s.Stop()
close(stopped)
}()
select {
case <-stopped:
case <-time.After(maxWait):
t.Fatalf("Stop did not return within %s while a STATE message was held", maxWait)
t.Fatalf("Stop did not return within %s while a STATE message was held",
maxWait)
}
select {
@@ -313,9 +380,12 @@ func TestStopReturnsAtOnceWithStateMessagesHeld(t *testing.T) {
// used to end in "send on closed channel". The raw handler runs in that gap on
// the read loop itself, so calling Stop from it hits the gap every time.
func TestStopBeforeMessageReachesQueues(t *testing.T) {
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, updateLine)
}))
t.Parallel()
srv := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, updateLine)
}))
defer srv.Close()
s := New(logger.New(), metrics.New())
@@ -329,7 +399,8 @@ func TestStopBeforeMessageReachesQueues(t *testing.T) {
ctx, cancel := context.WithCancel(context.Background())
s.cancel = cancel
if err := s.stream(ctx); !errors.Is(err, context.Canceled) {
err := s.stream(ctx)
if !errors.Is(err, context.Canceled) {
t.Fatalf("stream returned %v, want %v", err, context.Canceled)
}
@@ -341,6 +412,8 @@ func TestStopBeforeMessageReachesQueues(t *testing.T) {
// normally, as when RIS Live restarts. The streamer must treat that as a
// disconnect and connect again.
func TestReconnectWhenStreamEnds(t *testing.T) {
t.Parallel()
endResponse := func(http.ResponseWriter, *http.Request) {}
checkReconnect(t, endResponse, streamIdleTimeout)
}
@@ -349,9 +422,12 @@ func TestReconnectWhenStreamEnds(t *testing.T) {
// open without sending anything. The streamer must drop it once the idle
// timeout passes and connect again.
func TestReconnectWhenStreamGoesIdle(t *testing.T) {
t.Parallel()
const idleTimeout = 50 * time.Millisecond
holdResponse := func(w http.ResponseWriter, r *http.Request) {
w.(http.Flusher).Flush()
flush(w)
<-r.Context().Done()
}
checkReconnect(t, holdResponse, idleTimeout)
@@ -361,7 +437,10 @@ func TestReconnectWhenStreamGoesIdle(t *testing.T) {
// request and never answer it, not even with headers. The streamer must give
// up on it once the idle timeout passes and connect again.
func TestReconnectWhenServerNeverAnswers(t *testing.T) {
t.Parallel()
const idleTimeout = 50 * time.Millisecond
neverAnswer := func(_ http.ResponseWriter, r *http.Request) {
<-r.Context().Done()
}
@@ -373,26 +452,32 @@ func TestReconnectWhenServerNeverAnswers(t *testing.T) {
// line must start the idle timeout over, so the stream lasts until the server
// ends it instead of being dropped as idle.
func TestStreamNotDroppedWhileLinesArrive(t *testing.T) {
t.Parallel()
const (
idleTimeout = 100 * time.Millisecond
lineInterval = idleTimeout / 10
sendFor = 5 * idleTimeout
)
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
deadline := time.Now().Add(sendFor)
for time.Now().Before(deadline) {
_, _ = io.WriteString(w, updateLine)
w.(http.Flusher).Flush()
time.Sleep(lineInterval)
}
}))
srv := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, _ *http.Request) {
deadline := time.Now().Add(sendFor)
for time.Now().Before(deadline) {
_, _ = io.WriteString(w, updateLine)
flush(w)
time.Sleep(lineInterval)
}
}))
defer srv.Close()
s := New(logger.New(), metrics.New())
s.url = srv.URL
s.idleTimeout = idleTimeout
if err := s.stream(context.Background()); !errors.Is(err, errStreamEnded) {
err := s.stream(context.Background())
if !errors.Is(err, errStreamEnded) {
t.Fatalf("stream returned %v, want %v", err, errStreamEnded)
}
}
@@ -404,24 +489,31 @@ func TestStreamNotDroppedWhileLinesArrive(t *testing.T) {
// handler's queue and read on until the server ends the stream, instead of
// dropping the connection at the long line.
func TestStreamReadsLongLine(t *testing.T) {
t.Parallel()
const rawHexLength = 200 * bytesPerKB
longLine := `{"type":"ris_message","data":{"type":"UPDATE","peer":"192.0.2.1",` +
`"peer_asn":"64496","timestamp":1700000000,"raw":"` +
strings.Repeat("0", rawHexLength) + `"}}` + "\n"
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, longLine)
}))
srv := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, _ *http.Request) {
_, _ = io.WriteString(w, longLine)
}))
defer srv.Close()
s := New(logger.New(), metrics.New())
s.url = srv.URL
s.RegisterHandler(updateHandler{})
if err := s.stream(context.Background()); !errors.Is(err, errStreamEnded) {
err := s.stream(context.Background())
if !errors.Is(err, errStreamEnded) {
t.Fatalf("stream returned %v, want %v", err, errStreamEnded)
}
if queued := len(s.handlers[0].queue); queued != 1 {
queued := len(s.handlers[0].queue)
if queued != 1 {
t.Errorf("handler queue holds %d messages, want 1", queued)
}
}
@@ -431,28 +523,36 @@ func TestStreamReadsLongLine(t *testing.T) {
// line and then stays open. The streamer must come back for a second
// response, must not be marked connected when it does, and must handle the
// UPDATE in it.
func checkReconnect(t *testing.T, firstResponse http.HandlerFunc, idleTimeout time.Duration) {
func checkReconnect(
t *testing.T, firstResponse http.HandlerFunc, idleTimeout time.Duration,
) {
t.Helper()
const maxWait = 5 * time.Second
tracker := metrics.New()
connectedAtReconnect := make(chan bool, 1)
var requests atomic.Int32
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if requests.Add(1) == 1 {
firstResponse(w, r)
return
}
select {
case connectedAtReconnect <- tracker.IsConnected():
default:
}
_, _ = io.WriteString(w, updateLine)
w.(http.Flusher).Flush()
<-r.Context().Done()
}))
var requests atomic.Int32
srv := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, r *http.Request) {
if requests.Add(1) == 1 {
firstResponse(w, r)
return
}
select {
case connectedAtReconnect <- tracker.IsConnected():
default:
}
_, _ = io.WriteString(w, updateLine)
flush(w)
<-r.Context().Done()
}))
defer srv.Close()
s := New(logger.New(), tracker)
@@ -461,7 +561,9 @@ func checkReconnect(t *testing.T, firstResponse http.HandlerFunc, idleTimeout ti
s.minBackoff = time.Millisecond
handler := updateHandler{handled: make(chan struct{}, 1)}
s.RegisterHandler(handler)
if err := s.Start(); err != nil {
err := s.Start(t.Context())
if err != nil {
t.Fatalf("Start returned error: %v", err)
}
defer s.Stop()
@@ -473,7 +575,8 @@ func checkReconnect(t *testing.T, firstResponse http.HandlerFunc, idleTimeout ti
}
if <-connectedAtReconnect {
t.Error("streamer was still marked connected when it came back for a second response")
t.Error("streamer was still marked connected when it came back " +
"for a second response")
}
}
@@ -481,12 +584,15 @@ func checkReconnect(t *testing.T, firstResponse http.HandlerFunc, idleTimeout ti
// current count.
func settledGoroutineCount() int {
prev := runtime.NumGoroutine()
for range 20 {
time.Sleep(10 * time.Millisecond)
cur := runtime.NumGoroutine()
if cur == prev {
return cur
}
prev = cur
}
@@ -499,6 +605,7 @@ func waitForGoroutines(target int) bool {
if runtime.NumGoroutine() <= target {
return true
}
time.Sleep(10 * time.Millisecond)
}
internal/templates/templates.go
+11 -4
View File
@@ -60,6 +60,7 @@ func timeSince(t time.Time) string {
if duration < time.Minute {
return "just now"
}
if duration < time.Hour {
minutes := int(duration.Minutes())
if minutes == 1 {
@@ -68,6 +69,7 @@ func timeSince(t time.Time) string {
return duration.Truncate(time.Minute).String() + " ago"
}
if duration < hoursPerDay*time.Hour {
hours := int(duration.Hours())
if hours == 1 {
@@ -76,10 +78,12 @@ func timeSince(t time.Time) string {
return duration.Truncate(time.Hour).String() + " ago"
}
days := int(duration.Hours() / hoursPerDay)
if days == 1 {
return "1 day ago"
}
if days < daysPerMonth {
return duration.Truncate(hoursPerDay*time.Hour).String() + " ago"
}
@@ -118,7 +122,7 @@ func initTemplates() {
"appLicense": func() string { return version.License },
"appRepoURL": func() string { return version.RepoURL },
"appGitRevision": func() string { return version.GitRevisionShort },
"appGitCommitURL": func() string { return version.CommitURL() },
"appGitCommitURL": version.CommitURL,
}
// Parse index template
@@ -134,19 +138,22 @@ func initTemplates() {
}
// Parse AS detail template
defaultTemplates.ASDetail, err = template.New("asDetail").Funcs(funcs).Parse(asDetailHTML)
defaultTemplates.ASDetail, err = template.New("asDetail").
Funcs(funcs).Parse(asDetailHTML)
if err != nil {
panic("failed to parse AS detail template: " + err.Error())
}
// Parse prefix detail template
defaultTemplates.PrefixDetail, err = template.New("prefixDetail").Funcs(funcs).Parse(prefixDetailHTML)
defaultTemplates.PrefixDetail, err = template.New("prefixDetail").
Funcs(funcs).Parse(prefixDetailHTML)
if err != nil {
panic("failed to parse prefix detail template: " + err.Error())
}
// Parse prefix length template
defaultTemplates.PrefixLength, err = template.New("prefixLength").Funcs(funcs).Parse(prefixLengthHTML)
defaultTemplates.PrefixLength, err = template.New("prefixLength").
Funcs(funcs).Parse(prefixLengthHTML)
if err != nil {
panic("failed to parse prefix length template: " + err.Error())
}
internal/version/version.go
+1 -1
View File
@@ -3,7 +3,7 @@ package version
// Build-time variables set via ldflags
//
//nolint:gochecknoglobals // These must be variables to allow ldflags injection at build time
//nolint:gochecknoglobals // variables so that ldflags can set them at build time
var (
// GitRevision is the git commit hash
GitRevision = "unknown"
internal/whois/whois.go
+106 -78
View File
@@ -98,7 +98,8 @@ func (c *Client) LookupASN(ctx context.Context, asn int) (*ASNInfo, error) {
if err != nil {
// Return partial data from IANA if RIR query fails
info := c.parseResponse(asn, rir, ianaResp)
info.RawResponse = ianaResp + "\n--- RIR query failed: " + err.Error() + " ---\n"
info.RawResponse = fmt.Sprintf("%s\n--- RIR query failed: %v ---\n",
ianaResp, err)
return info, nil
}
@@ -120,28 +121,33 @@ func (c *Client) query(ctx context.Context, server, query string) (string, error
defer func() { _ = conn.Close() }()
// Set deadlines
if err := conn.SetWriteDeadline(time.Now().Add(writeTimeout)); err != nil {
err = conn.SetWriteDeadline(time.Now().Add(writeTimeout))
if err != nil {
return "", fmt.Errorf("set write deadline: %w", err)
}
// Send query
if _, err := fmt.Fprintf(conn, "%s\r\n", query); err != nil {
_, err = fmt.Fprintf(conn, "%s\r\n", query)
if err != nil {
return "", fmt.Errorf("write query: %w", err)
}
// Read response
if err := conn.SetReadDeadline(time.Now().Add(readTimeout)); err != nil {
err = conn.SetReadDeadline(time.Now().Add(readTimeout))
if err != nil {
return "", fmt.Errorf("set read deadline: %w", err)
}
var sb strings.Builder
scanner := bufio.NewScanner(conn)
for scanner.Scan() {
sb.WriteString(scanner.Text())
sb.WriteString("\n")
}
if err := scanner.Err(); err != nil {
err = scanner.Err()
if err != nil {
return sb.String(), fmt.Errorf("read response: %w", err)
}
@@ -149,10 +155,10 @@ func (c *Client) query(ctx context.Context, server, query string) (string, error
}
// parseIANAReferral extracts the RIR and WHOIS server from an IANA response.
func (c *Client) parseIANAReferral(response string) (rir, whoisServer string) {
lines := strings.Split(response, "\n")
func (c *Client) parseIANAReferral(response string) (string, string) {
var rir string
for _, line := range lines {
for line := range strings.SplitSeq(response, "\n") {
line = strings.TrimSpace(line)
// Look for "refer:" line
@@ -183,18 +189,9 @@ func (c *Client) parseIANAReferral(response string) (rir, whoisServer string) {
// Also check organisation line for RIR hints
if strings.HasPrefix(strings.ToLower(line), "organisation:") {
org := strings.ToLower(line)
switch {
case strings.Contains(org, "arin"):
rir = RIRARIN
case strings.Contains(org, "ripe"):
rir = RIRRIPE
case strings.Contains(org, "apnic"):
rir = RIRAPNIC
case strings.Contains(org, "lacnic"):
rir = RIRLACNIC
case strings.Contains(org, "afrinic"):
rir = RIRAFRNIC
orgRIR := rirInOrganisationLine(strings.ToLower(line))
if orgRIR != "" {
rir = orgRIR
}
}
}
@@ -202,6 +199,25 @@ func (c *Client) parseIANAReferral(response string) (rir, whoisServer string) {
return rir, ""
}
// rirInOrganisationLine returns the RIR a lowercased IANA "organisation:"
// line names, or "" when it names none.
func rirInOrganisationLine(org string) string {
switch {
case strings.Contains(org, "arin"):
return RIRARIN
case strings.Contains(org, "ripe"):
return RIRRIPE
case strings.Contains(org, "apnic"):
return RIRAPNIC
case strings.Contains(org, "lacnic"):
return RIRLACNIC
case strings.Contains(org, "afrinic"):
return RIRAFRNIC
default:
return ""
}
}
// parseResponse extracts ASN information from a WHOIS response.
func (c *Client) parseResponse(asn int, rir, response string) *ASNInfo {
info := &ASNInfo{
@@ -211,11 +227,13 @@ func (c *Client) parseResponse(asn int, rir, response string) *ASNInfo {
}
lines := strings.Split(response, "\n")
var addressLines []string
for _, line := range lines {
line = strings.TrimSpace(line)
if line == "" || strings.HasPrefix(line, "%") || strings.HasPrefix(line, "#") {
if line == "" || strings.HasPrefix(line, "%") ||
strings.HasPrefix(line, "#") {
continue
}
@@ -232,63 +250,14 @@ func (c *Client) parseResponse(asn int, rir, response string) *ASNInfo {
continue
}
switch key {
// AS Name (varies by RIR)
case "asname", "as-name":
if info.ASName == "" {
info.ASName = value
}
// Organization
case "orgname", "org-name", "owner":
if info.OrgName == "" {
info.OrgName = value
}
case "orgid", "org-id", "org":
if info.OrgID == "" {
info.OrgID = value
}
// Address (collect multiple lines)
case "address":
if key == "address" {
addressLines = append(addressLines, value)
// Country
case "country":
if info.CountryCode == "" && len(value) == 2 {
info.CountryCode = strings.ToUpper(value)
}
// Abuse contact
case "orgabuseemail", "abuse-mailbox":
if info.AbuseEmail == "" {
info.AbuseEmail = value
}
case "orgabusephone":
if info.AbusePhone == "" {
info.AbusePhone = value
}
// Tech contact
case "orgtechemail":
if info.TechEmail == "" {
info.TechEmail = value
}
case "orgtechphone":
if info.TechPhone == "" {
info.TechPhone = value
}
// Registration dates
case "regdate", "created":
if info.RegDate == nil {
info.RegDate = c.parseDate(value)
}
case "updated", "last-modified", "changed":
if info.LastMod == nil {
info.LastMod = c.parseDate(value)
}
continue
}
c.setResponseField(info, key, value)
}
// Combine address lines
@@ -304,6 +273,65 @@ func (c *Client) parseResponse(asn int, rir, response string) *ASNInfo {
return info
}
// setResponseField stores the value of one "key: value" line of a WHOIS
// response in info, keeping the first value seen for each field.
func (c *Client) setResponseField(info *ASNInfo, key, value string) {
switch key {
// Country
case "country":
if info.CountryCode == "" && len(value) == 2 {
info.CountryCode = strings.ToUpper(value)
}
// Registration dates
case "regdate", "created":
if info.RegDate == nil {
info.RegDate = c.parseDate(value)
}
case "updated", "last-modified", "changed":
if info.LastMod == nil {
info.LastMod = c.parseDate(value)
}
default:
field := textFieldForKey(info, key)
if field != nil && *field == "" {
*field = value
}
}
}
// textFieldForKey returns the text field of info that a WHOIS response key
// fills, or nil when the key fills none.
func textFieldForKey(info *ASNInfo, key string) *string {
switch key {
// AS Name (varies by RIR)
case "asname", "as-name":
return &info.ASName
// Organization
case "orgname", "org-name", "owner":
return &info.OrgName
case "orgid", "org-id", "org":
return &info.OrgID
// Abuse contact
case "orgabuseemail", "abuse-mailbox":
return &info.AbuseEmail
case "orgabusephone":
return &info.AbusePhone
// Tech contact
case "orgtechemail":
return &info.TechEmail
case "orgtechphone":
return &info.TechPhone
default:
return nil
}
}
// parseDate attempts to parse various date formats used in WHOIS responses.
func (c *Client) parseDate(value string) *time.Time {
// Common formats
@@ -319,15 +347,15 @@ func (c *Client) parseDate(value string) *time.Time {
value = strings.TrimSpace(value)
// Handle "YYYYMMDD" format from LACNIC
if len(value) == lacnicDateFormatLen {
if _, err := time.Parse("20060102", value); err == nil {
t, _ := time.Parse("20060102", value)
t, err := time.Parse("20060102", value)
if err == nil {
return &t
}
}
for _, format := range formats {
if t, err := time.Parse(format, value); err == nil {
t, err := time.Parse(format, value)
if err == nil {
return &t
}
}