refactor(agent): stop embedding GeoIP MMDB in agent binary

Agent ships without City/Country MMDB in the binary; Docker images COPY
databases into data_dir, bare installs seed via download on first start.
Server keeps Country-only embed for optional MaxMind control-plane use.
Also harden fetch script nonempty check and reject non-file MMDB paths.
This commit is contained in:
ryan
2026-08-06 16:24:57 +08:00
parent fb08002e99
commit 3328d3d121
17 changed files with 226 additions and 180 deletions
Binary file not shown.

Before

Width:  |  Height:  |  Size: 63 MiB

Binary file not shown.
+7 -10
View File
@@ -1,16 +1,13 @@
// Package geoipdata embeds the default MaxMind GeoLite2 databases.
// Package geoipdata holds shared GeoIP database filename constants.
//
// MaxMind MMDB files are NOT embedded into the agent binary. Docker images
// COPY them onto the default data paths; bare binary installs seed via download
// on first start (see geoipupdate).
package geoipdata
import "embed"
// FS holds the embedded GeoLite2 Country and City databases.
//
//go:embed GeoLite2-Country.mmdb GeoLite2-City.mmdb
var FS embed.FS
const (
// DefaultMMDBName is the filename of the embedded MaxMind Country database.
// DefaultMMDBName is the default Country database filename.
DefaultMMDBName = "GeoLite2-Country.mmdb"
// DefaultCityMMDBName is the filename of the embedded MaxMind City database.
// DefaultCityMMDBName is the default City database filename.
DefaultCityMMDBName = "GeoLite2-City.mmdb"
)
@@ -1,38 +0,0 @@
package geoipdata
import (
"io/fs"
"strings"
"testing"
"github.com/oschwald/maxminddb-golang"
)
func TestEmbeddedDatabasesAreValid(t *testing.T) {
tests := []struct {
name string
filename string
databaseTypePart string
}{
{name: "Country", filename: DefaultMMDBName, databaseTypePart: "Country"},
{name: "City", filename: DefaultCityMMDBName, databaseTypePart: "City"},
}
for _, test := range tests {
t.Run(test.name, func(t *testing.T) {
data, err := fs.ReadFile(FS, test.filename)
if err != nil {
t.Fatalf("read embedded database: %v", err)
}
reader, err := maxminddb.FromBytes(data)
if err != nil {
t.Fatalf("open embedded database: %v", err)
}
defer reader.Close()
if !strings.Contains(reader.Metadata.DatabaseType, test.databaseTypePart) {
t.Fatalf("unexpected database type %q", reader.Metadata.DatabaseType)
}
})
}
}
+56 -46
View File
@@ -5,23 +5,16 @@ import (
"context"
"errors"
"fmt"
"io/fs"
"log/slog"
"os"
"path/filepath"
"time"
"github.com/Rain-kl/Wavelet/internal/apps/agent/geoipdata"
"github.com/Rain-kl/Wavelet/pkg/geoip"
)
const (
mmdbDirPerm = 0o750
mmdbFilePerm = 0o600
)
// Updater periodically downloads a fresh GeoIP MMDB file and seeds the
// initial embedded database when none is present on disk.
// Updater periodically downloads a fresh GeoIP MMDB file and seeds missing
// databases via download (or relies on image-provided files under data_dir).
type Updater struct {
MMDBPath string
DownloadURL string
@@ -31,48 +24,58 @@ type Updater struct {
downloadDatabase func(context.Context, string, string) error
}
// EnsureInitialDatabase seeds the Country MMDB file from the embedded database if it does not exist on disk.
func (u *Updater) EnsureInitialDatabase() error {
return ensureEmbeddedDatabase(u.MMDBPath, geoipdata.DefaultMMDBName, "Country")
}
func ensureEmbeddedDatabase(targetPath string, embeddedName string, databaseName string) error {
path := filepath.Clean(targetPath)
if path == "" || path == "." {
// EnsureInitialDatabases downloads any missing Country/City MMDB once.
// When files already exist (e.g. Docker image COPY), this is a no-op.
// Network is used only when a managed path is absent — not for binary embeds.
func (u *Updater) EnsureInitialDatabases(ctx context.Context) error {
if u == nil {
return nil
}
if _, err := os.Stat(path); err == nil {
return nil
} else if !os.IsNotExist(err) {
return fmt.Errorf("stat mmdb file failed: %w", err)
}
data, err := fs.ReadFile(geoipdata.FS, embeddedName)
if err != nil {
return fmt.Errorf("read embedded %s mmdb failed: %w", databaseName, err)
}
if err := os.MkdirAll(filepath.Dir(path), mmdbDirPerm); err != nil {
return fmt.Errorf("create mmdb directory failed: %w", err)
}
if err := os.WriteFile(path, data, mmdbFilePerm); err != nil {
return fmt.Errorf("write initial mmdb failed: %w", err)
}
slog.Info("initialized GeoIP mmdb from embedded database", "database", databaseName, "path", path, "size", len(data))
return nil
return u.ensureMissingDatabases(ctx)
}
// EnsureInitialDatabases seeds both Country and City from embedded databases
// when either managed file is absent. Network downloads are reserved for the periodic updater.
func (u *Updater) EnsureInitialDatabases(_ context.Context) error {
func (u *Updater) ensureMissingDatabases(ctx context.Context) error {
databases := u.managedDatabases()
var errs []error
if err := u.EnsureInitialDatabase(); err != nil {
errs = append(errs, err)
}
if err := ensureEmbeddedDatabase(u.CityMMDBPath, geoipdata.DefaultCityMMDBName, "City"); err != nil {
errs = append(errs, err)
for _, database := range databases {
if database.path == "" || database.downloadURL == "" {
continue
}
exists, err := fileExists(database.path)
if err != nil {
errs = append(errs, fmt.Errorf("stat GeoIP %s mmdb failed: %w", database.name, err))
continue
}
if exists {
continue
}
if err := u.download(ctx, database.path, database.downloadURL); err != nil {
errs = append(errs, fmt.Errorf("seed GeoIP %s mmdb failed: %w", database.name, err))
continue
}
slog.Info("seeded GeoIP mmdb via download", "database", database.name, "path", database.path)
}
return errors.Join(errs...)
}
func fileExists(path string) (bool, error) {
path = filepath.Clean(path)
if path == "" || path == "." {
return false, nil
}
info, err := os.Stat(path)
if err == nil {
if !info.Mode().IsRegular() {
return false, fmt.Errorf("GeoIP MMDB path is not a regular file: %s", path)
}
return true, nil
}
if os.IsNotExist(err) {
return false, nil
}
return false, err
}
func (u *Updater) download(ctx context.Context, path string, downloadURL string) error {
if u.downloadDatabase != nil {
return u.downloadDatabase(ctx, path, downloadURL)
@@ -80,8 +83,12 @@ func (u *Updater) download(ctx context.Context, path string, downloadURL string)
return geoip.DownloadMaxMindDatabase(ctx, path, downloadURL)
}
func (u *Updater) updateDatabases(ctx context.Context) error {
databases := []struct {
func (u *Updater) managedDatabases() []struct {
name string
path string
downloadURL string
} {
return []struct {
name string
path string
downloadURL string
@@ -89,9 +96,12 @@ func (u *Updater) updateDatabases(ctx context.Context) error {
{name: "Country", path: u.MMDBPath, downloadURL: u.DownloadURL},
{name: "City", path: u.CityMMDBPath, downloadURL: u.CityDownloadURL},
}
}
func (u *Updater) updateDatabases(ctx context.Context) error {
var errs []error
for _, database := range databases {
if database.path == "" || (database.name == "City" && database.downloadURL == "") {
for _, database := range u.managedDatabases() {
if database.path == "" || database.downloadURL == "" {
continue
}
if err := u.download(ctx, database.path, database.downloadURL); err != nil {
+61 -51
View File
@@ -6,77 +6,66 @@ import (
"os"
"path/filepath"
"slices"
"strings"
"testing"
)
func TestEnsureInitialDatabaseCopiesEmbeddedMMDB(t *testing.T) {
tempDir := t.TempDir()
path := filepath.Join(tempDir, "GeoLite2-Country.mmdb")
updater := &Updater{MMDBPath: path}
if err := updater.EnsureInitialDatabase(); err != nil {
t.Fatalf("EnsureInitialDatabase failed: %v", err)
}
info, err := os.Stat(path)
if err != nil {
t.Fatalf("expected mmdb to exist: %v", err)
}
if info.Size() == 0 {
t.Fatal("expected copied mmdb to be non-empty")
}
}
func TestEnsureInitialDatabasesCopiesEmbeddedCityWithoutDownload(t *testing.T) {
tempDir := t.TempDir()
countryPath := filepath.Join(tempDir, "GeoLite2-Country.mmdb")
cityPath := filepath.Join(tempDir, "GeoLite2-City.mmdb")
updater := &Updater{
MMDBPath: countryPath,
CityMMDBPath: cityPath,
CityDownloadURL: "https://geo.example/GeoLite2-City.mmdb",
downloadDatabase: func(_ context.Context, path, downloadURL string) error {
t.Fatalf("initial embedded seed must not download %s from %s", path, downloadURL)
return nil
},
}
if err := updater.EnsureInitialDatabases(context.Background()); err != nil {
t.Fatalf("EnsureInitialDatabases failed: %v", err)
}
if _, err := os.Stat(countryPath); err != nil {
t.Fatalf("expected embedded Country database: %v", err)
}
data, err := os.ReadFile(cityPath)
if err != nil || len(data) == 0 {
t.Fatalf("expected embedded City database, size=%d err=%v", len(data), err)
}
}
func TestEnsureInitialDatabasesKeepsExistingCityWithoutDownload(t *testing.T) {
func TestEnsureInitialDatabasesDownloadsMissingOnly(t *testing.T) {
tempDir := t.TempDir()
countryPath := filepath.Join(tempDir, "GeoLite2-Country.mmdb")
cityPath := filepath.Join(tempDir, "GeoLite2-City.mmdb")
if err := os.WriteFile(cityPath, []byte("existing-city"), 0o600); err != nil {
t.Fatal(err)
}
var downloaded []string
updater := &Updater{
MMDBPath: countryPath,
DownloadURL: "https://geo.example/GeoLite2-Country.mmdb",
CityMMDBPath: cityPath,
CityDownloadURL: "https://geo.example/GeoLite2-City.mmdb",
downloadDatabase: func(_ context.Context, _, _ string) error {
return errors.New("city unavailable")
downloadDatabase: func(_ context.Context, path, _ string) error {
downloaded = append(downloaded, path)
return os.WriteFile(path, []byte("downloaded"), 0o600)
},
}
if err := updater.EnsureInitialDatabases(context.Background()); err != nil {
t.Fatalf("EnsureInitialDatabases failed: %v", err)
}
if _, err := os.Stat(countryPath); err != nil {
t.Fatalf("expected Country fallback to remain available: %v", err)
if !slices.Equal(downloaded, []string{countryPath}) {
t.Fatalf("expected only missing Country download, got %#v", downloaded)
}
data, err := os.ReadFile(cityPath)
if err != nil || string(data) != "existing-city" {
t.Fatalf("expected existing City database to remain untouched, data=%q err=%v", data, err)
if data, err := os.ReadFile(cityPath); err != nil || string(data) != "existing-city" {
t.Fatalf("existing City must stay untouched, data=%q err=%v", data, err)
}
if data, err := os.ReadFile(countryPath); err != nil || string(data) != "downloaded" {
t.Fatalf("Country should be seeded via download, data=%q err=%v", data, err)
}
}
func TestEnsureInitialDatabasesNoOpWhenPresent(t *testing.T) {
tempDir := t.TempDir()
countryPath := filepath.Join(tempDir, "GeoLite2-Country.mmdb")
cityPath := filepath.Join(tempDir, "GeoLite2-City.mmdb")
if err := os.WriteFile(countryPath, []byte("c"), 0o600); err != nil {
t.Fatal(err)
}
if err := os.WriteFile(cityPath, []byte("city"), 0o600); err != nil {
t.Fatal(err)
}
updater := &Updater{
MMDBPath: countryPath,
DownloadURL: "https://geo.example/GeoLite2-Country.mmdb",
CityMMDBPath: cityPath,
CityDownloadURL: "https://geo.example/GeoLite2-City.mmdb",
downloadDatabase: func(_ context.Context, path, downloadURL string) error {
t.Fatalf("must not download when files exist: %s %s", path, downloadURL)
return nil
},
}
if err := updater.EnsureInitialDatabases(context.Background()); err != nil {
t.Fatalf("EnsureInitialDatabases failed: %v", err)
}
}
@@ -101,3 +90,24 @@ func TestUpdateDatabasesAttemptsCityAfterCountryFailure(t *testing.T) {
t.Fatalf("expected independent Country then City attempts, paths=%#v err=%v", paths, err)
}
}
func TestEnsureInitialDatabasesRejectsDirectoryPath(t *testing.T) {
tempDir := t.TempDir()
// Point Country path at a directory so fileExists must not treat it as seeded.
updater := &Updater{
MMDBPath: tempDir,
DownloadURL: "https://geo.example/GeoLite2-Country.mmdb",
downloadDatabase: func(_ context.Context, path, downloadURL string) error {
t.Fatalf("must not download when path is a directory: %s %s", path, downloadURL)
return nil
},
}
err := updater.EnsureInitialDatabases(context.Background())
if err == nil {
t.Fatal("expected error when MMDB path is a directory")
}
if !strings.Contains(err.Error(), "not a regular file") {
t.Fatalf("expected regular-file error, got %v", err)
}
}