Enable storage retention for alpine

docs/configuration.md

@@ -88,7 +88,7 @@ A package rule wins over an ecosystem rule, which wins over `default`. Package r
 
 | Ecosystem key | Retention supported | Package rule form |
 |---------------|---------------------|-------------------|
-| `alpine` | no | - |
+| `alpine` | yes | `pkg:apk/curl` or `pkg:apk/alpine/curl`; case doesn't matter. A rule covers the package in every configured APK repository. Alpine mirrors keep only the current build of each package per release, so an evicted older one can't be fetched again. |
 | `cargo` | no | - |
 | `composer` | no | - |
 | `conan` | no | - |

internal/handler/apk.go

@@ -6,6 +6,9 @@ import (
 	"net/http"
 	"regexp"
 	"strings"
+
+	"github.com/git-pkgs/proxy/internal/retention"
+	"github.com/git-pkgs/purl"
 )
 
 const (
@@ -18,6 +21,24 @@ const (
 	apkMatchCount      = 3 // full match + name + version
 )
 
+// Packages are stored as pkg:apk/alpine/{name}, the namespace purl gives
+// every apk package, whichever configured repository they came from. Package
+// rules may leave the namespace out.
+func init() {
+	retention.Default.Register(retention.Spec{Key: apkEcosystem, CanonicalPackage: apkRetentionKey})
+}
+
+// apkRetentionKey maps pkg:apk/{name} and pkg:apk/alpine/{name} to the stored
+// package PURL. Another namespace, or a name with a slash, can never match a
+// stored package and is refused.
+func apkRetentionKey(p *purl.PURL) (string, bool) {
+	if p == nil || p.Type != "apk" || (p.Namespace != "" && p.Namespace != "alpine") || strings.Contains(p.Name, "/") {
+		return "", false
+	}
+	canonical := purl.MakePURLString(apkEcosystem, p.Name, "")
+	return canonical, canonical != ""
+}
+
 // APKHandler handles Alpine APK repository protocol requests. Each configured
 // upstream repository is mounted at /apk/{repository}/ and the remaining path
 // mirrors the upstream layout ({release}/{repo}/{arch}/{file}).

internal/handler/apk_test.go

@@ -8,6 +8,7 @@ import (
 	"testing"
 	"time"
 
+	"github.com/git-pkgs/purl"
 	"github.com/git-pkgs/registries/fetch"
 )
 
@@ -360,3 +361,32 @@ func serveAPKRequest(h *APKHandler, target string) *httptest.ResponseRecorder {
 	h.Routes().ServeHTTP(w, httptest.NewRequest(http.MethodGet, target, nil))
 	return w
 }
+
+func TestAPKRetentionKey(t *testing.T) {
+	tests := []struct {
+		key  string
+		want string
+		ok   bool
+	}{
+		{"pkg:apk/curl", "pkg:apk/alpine/curl", true},
+		{"pkg:apk/alpine/curl", "pkg:apk/alpine/curl", true},
+		{"pkg:apk/Alpine/Curl", "pkg:apk/alpine/curl", true},
+		{"pkg:apk/alpine/libstdc++", "pkg:apk/alpine/libstdc%2B%2B", true},
+		{"pkg:apk/wolfi/curl", "", false},
+		{"pkg:apk/alpine/main%2Fcurl", "", false},
+		{"pkg:deb/alpine/curl", "", false},
+	}
+	for _, tt := range tests {
+		p, err := purl.Parse(tt.key)
+		if err != nil {
+			t.Fatalf("Parse(%s): %v", tt.key, err)
+		}
+		got, ok := apkRetentionKey(p)
+		if ok != tt.ok || (ok && got != tt.want) {
+			t.Errorf("apkRetentionKey(%s) = %q, %v, want %q, %v", tt.key, got, ok, tt.want, tt.ok)
+		}
+	}
+	if _, ok := apkRetentionKey(nil); ok {
+		t.Error("apkRetentionKey(nil) accepted")
+	}
+}

internal/server/retention_alpine_test.go

@@ -0,0 +1,195 @@
+package server
+
+import (
+	"context"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/git-pkgs/proxy/internal/config"
+	"github.com/git-pkgs/proxy/internal/database"
+	"github.com/git-pkgs/proxy/internal/handler"
+	"github.com/git-pkgs/proxy/internal/retention"
+	"github.com/git-pkgs/registries/fetch"
+)
+
+const (
+	alpineCurlPath    = "/v3.22/main/x86_64/curl-8.14.1-r1.apk"
+	alpineBusyboxPath = "/v3.22/main/x86_64/busybox-1.37.0-r12.apk"
+)
+
+// alpineRetentionUpstream serves package files the way an Alpine mirror
+// does, counting downloads.
+func alpineRetentionUpstream(t *testing.T) (*httptest.Server, *atomic.Int64) {
+	t.Helper()
+	var downloads atomic.Int64
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		switch r.URL.Path {
+		case alpineCurlPath, alpineBusyboxPath:
+			downloads.Add(1)
+			_, _ = io.WriteString(w, "apk bytes")
+		default:
+			http.NotFound(w, r)
+		}
+	}))
+	t.Cleanup(srv.Close)
+	return srv, &downloads
+}
+
+// newAlpineRetentionProxy mounts the upstream twice, as the default "alpine"
+// repository and as a named "edge" repository.
+func newAlpineRetentionProxy(t *testing.T, upstream *httptest.Server) (*httptest.Server, *database.DB) {
+	t.Helper()
+	db, store := setupEvictionTest(t)
+	// The upstream is a plain-HTTP loopback server; the default fetcher
+	// refuses loopback, a stock client reaches it.
+	fetcher := fetch.NewFetcher(fetch.WithHTTPClient(upstream.Client()), fetch.WithMaxRetries(0))
+	t.Cleanup(func() { _ = fetcher.Close() })
+	proxy := handler.NewProxy(db, store, fetcher, fetch.NewResolver(), discardLogger())
+	proxy.HTTPClient = upstream.Client()
+	h := handler.NewAPKHandler(proxy, "http://proxy.test", map[string]string{"alpine": upstream.URL, "edge": upstream.URL})
+	srv := httptest.NewServer(h.Routes())
+	t.Cleanup(srv.Close)
+	return srv, db
+}
+
+func alpineGet(t *testing.T, url string) {
+	t.Helper()
+	resp, err := http.Get(url)
+	if err != nil {
+		t.Fatalf("GET %s: %v", url, err)
+	}
+	body, _ := io.ReadAll(resp.Body)
+	_ = resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		t.Fatalf("GET %s = %d %s, want 200", url, resp.StatusCode, body)
+	}
+}
+
+// alpineArtifact returns the record of the file cached under filename,
+// which for apk is the repository-qualified request path.
+func alpineArtifact(t *testing.T, db *database.DB, filename string) database.Artifact {
+	t.Helper()
+	var art database.Artifact
+	if err := db.Get(&art, db.Rebind(`SELECT * FROM artifacts WHERE filename = ?`), filename); err != nil {
+		t.Fatalf("reading artifact %s: %v", filename, err)
+	}
+	return art
+}
+
+func alpineBackdate(t *testing.T, db *database.DB, filename string, fetched, accessed time.Time) {
+	t.Helper()
+	query := db.Rebind(`UPDATE artifacts SET fetched_at = ?, last_accessed_at = ? WHERE filename = ?`)
+	if _, err := db.Exec(query, fetched, accessed, filename); err != nil {
+		t.Fatalf("backdating %s: %v", filename, err)
+	}
+}
+
+func TestAlpineRegistersForRetention(t *testing.T) {
+	if _, ok := retention.Default.Lookup("alpine"); !ok {
+		t.Fatal("alpine is not registered for retention")
+	}
+	for _, value := range []string{"alpine", "apk"} {
+		if key, ok := retention.Default.KeyForDBEcosystem(value); !ok || key != "alpine" {
+			t.Errorf("KeyForDBEcosystem(%s) = %q, %v, want alpine, true", value, key, ok)
+		}
+	}
+}
+
+// A package served from the cache stays while it is used, is evicted once
+// its rule runs out, comes back from upstream on the next download, and
+// stays after that even with the evicted copy's old access time on the
+// record.
+func TestAlpineRetentionEvictsAndRefetches(t *testing.T) {
+	upstream, downloads := alpineRetentionUpstream(t)
+	srv, db := newAlpineRetentionProxy(t, upstream)
+	const file = "alpine" + alpineCurlPath
+
+	alpineGet(t, srv.URL+"/alpine"+alpineCurlPath)
+	alpineGet(t, srv.URL+"/alpine"+alpineCurlPath)
+	if got := downloads.Load(); got != 1 {
+		t.Fatalf("upstream downloads = %d, want 1", got)
+	}
+	if !alpineArtifact(t, db, file).LastAccessedAt.Valid {
+		t.Fatal("cache hit did not record an access")
+	}
+
+	rules := retention.Rules{Ecosystems: map[string]time.Duration{"alpine": retentionDay}}
+	now := time.Now()
+	sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, now)
+	if !alpineArtifact(t, db, file).StoragePath.Valid {
+		t.Fatal("package used today was evicted")
+	}
+
+	old := now.Add(-2 * retentionDay)
+	alpineBackdate(t, db, file, old, old)
+	sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, now)
+	if alpineArtifact(t, db, file).StoragePath.Valid {
+		t.Fatal("package unused for two days kept under a one-day rule")
+	}
+
+	alpineGet(t, srv.URL+"/alpine"+alpineCurlPath)
+	if got := downloads.Load(); got != 2 {
+		t.Errorf("upstream downloads = %d, want 2 after the eviction", got)
+	}
+	art := alpineArtifact(t, db, file)
+	alpineBackdate(t, db, file, art.FetchedAt.Time, now.Add(-30*retentionDay))
+	sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, time.Now())
+	if !alpineArtifact(t, db, file).StoragePath.Valid {
+		t.Error("refetched package evicted again by the next sweep")
+	}
+}
+
+// A key in another namespace parses as an apk PURL but can never match a
+// stored package, so it is refused at validation. Both documented forms are
+// accepted.
+func TestAlpineRetentionRefusesOtherNamespaces(t *testing.T) {
+	cfg := config.Default()
+	cfg.Storage.Retention.Packages = map[string]string{"pkg:apk/curl": "0", "pkg:apk/alpine/busybox": "0"}
+	if _, err := cfg.RetentionRules(retention.Default); err != nil {
+		t.Fatalf("RetentionRules: %v", err)
+	}
+	cfg.Storage.Retention.Packages = map[string]string{"pkg:apk/wolfi/curl": "0"}
+	_, err := cfg.RetentionRules(retention.Default)
+	if err == nil || !strings.Contains(err.Error(), "does not match how alpine packages are stored") {
+		t.Errorf("RetentionRules(pkg:apk/wolfi/curl) error = %v, want the key refused", err)
+	}
+}
+
+// A package rule reaches the package in every configured repository, in
+// either documented form.
+func TestAlpineRetentionPackageRules(t *testing.T) {
+	upstream, _ := alpineRetentionUpstream(t)
+	srv, db := newAlpineRetentionProxy(t, upstream)
+	files := map[string]bool{
+		"alpine" + alpineCurlPath:    true,
+		"edge" + alpineCurlPath:      true,
+		"alpine" + alpineBusyboxPath: false,
+	}
+
+	for _, key := range []string{"pkg:apk/curl", "pkg:apk/alpine/curl"} {
+		old := time.Now().Add(-30 * retentionDay)
+		for file := range files {
+			alpineGet(t, srv.URL+"/"+file)
+			alpineBackdate(t, db, file, old, old)
+		}
+		cfg := config.Default()
+		cfg.Storage.Retention.Ecosystems = map[string]string{"alpine": "1d"}
+		cfg.Storage.Retention.Packages = map[string]string{key: "0"}
+		rules, err := cfg.RetentionRules(retention.Default)
+		if err != nil {
+			t.Fatalf("RetentionRules(%s): %v", key, err)
+		}
+		sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, time.Now())
+
+		for file, want := range files {
+			if got := alpineArtifact(t, db, file).StoragePath.Valid; got != want {
+				t.Errorf("with %s: %s cached = %v, want %v", key, file, got, want)
+			}
+		}
+	}
+}