satellite/audit: add ContainmentSyncChore

We will be needing an infrequent chore to check which nodes are in the
reverify queue and synchronize that set with the 'contained' field in
the nodes db, since it is easily possible for them to get out of sync.
(We can't require that the reverification queue table be in the same
database as the nodes table, so maintaining consistency with SQL
transactions is out. Plus, even if they were in the same database, using
such SQL transactions to maintain consistency would be slow and
unwieldy.)

This commit adds the actual chore.

Refs: https://github.com/storj/storj/issues/5431
Change-Id: Id78b40bf69fae1ac39010e3b553315db8a1472bd

private/testplanet/satellite.go

@@ -148,6 +148,7 @@ type Satellite struct {
 		Verifier             *audit.Verifier
 		Reverifier           *audit.Reverifier
 		Reporter             audit.Reporter
+		ContainmentSyncChore *audit.ContainmentSyncChore
 	}
 
 	Reputation struct {
@@ -644,6 +645,7 @@ func createNewSystem(name string, log *zap.Logger, config satellite.Config, peer
 	system.Audit.Verifier = auditorPeer.Audit.Verifier
 	system.Audit.Reverifier = auditorPeer.Audit.Reverifier
 	system.Audit.Reporter = auditorPeer.Audit.Reporter
+	system.Audit.ContainmentSyncChore = peer.Audit.ContainmentSyncChore
 
 	system.GarbageCollection.Sender = gcPeer.GarbageCollection.Sender
 	system.GarbageCollection.BloomFilters = gcBFPeer.GarbageCollection.Service

satellite/audit/reverify_chore.go

@@ -0,0 +1,69 @@
+// Copyright (C) 2023 Storj Labs, Inc.
+// See LICENSE for copying information.
+
+package audit
+
+import (
+	"context"
+	"time"
+
+	"go.uber.org/zap"
+
+	"storj.io/common/sync2"
+	"storj.io/storj/satellite/overlay"
+)
+
+// ContainmentSyncChore is a chore to update the set of contained nodes in the
+// overlay cache. This is necessary because it is possible for the "contained"
+// field in the nodes table to disagree with whether a node appears in the
+// reverification queue. We make an effort to keep them in sync when making
+// changes to the reverification queue, but this infrequent chore will clean up
+// any inconsistencies that creep in (because we can't maintain perfect
+// consistency while the reverification queue and the nodes table may be in
+// separate databases). Fortunately, it is acceptable for a node's containment
+// status to be out of date for some amount of time.
+type ContainmentSyncChore struct {
+	log     *zap.Logger
+	queue   ReverifyQueue
+	overlay overlay.DB
+
+	Loop *sync2.Cycle
+}
+
+// NewContainmentSyncChore creates a new ContainmentSyncChore.
+func NewContainmentSyncChore(log *zap.Logger, queue ReverifyQueue, overlay overlay.DB, interval time.Duration) *ContainmentSyncChore {
+	return &ContainmentSyncChore{
+		log:     log,
+		queue:   queue,
+		overlay: overlay,
+		Loop:    sync2.NewCycle(interval),
+	}
+}
+
+// Run runs the reverify chore.
+func (rc *ContainmentSyncChore) Run(ctx context.Context) (err error) {
+	defer mon.Task()(&ctx)(&err)
+
+	return rc.Loop.Run(ctx, rc.syncContainedStatus)
+}
+
+// SyncContainedStatus updates the contained status of all nodes in the overlay cache
+// as necessary to match whether they currently appear in the reverification queue at
+// least once.
+func (rc *ContainmentSyncChore) syncContainedStatus(ctx context.Context) (err error) {
+	defer mon.Task()(&ctx)(&err)
+
+	containedSet, err := rc.queue.GetAllContainedNodes(ctx)
+	if err != nil {
+		rc.log.Error("failed to get set of contained nodes from reverify queue", zap.Error(err))
+		return nil
+	}
+	err = rc.overlay.SetAllContainedNodes(ctx, containedSet)
+	if err != nil {
+		rc.log.Error("failed to update the set of contained nodes in the overlay cache", zap.Error(err))
+		return nil
+	}
+	rc.log.Info("updated containment status of all nodes as necessary",
+		zap.Int("num contained nodes", len(containedSet)))
+	return nil
+}

satellite/audit/reverify_chore_test.go

@@ -0,0 +1,114 @@
+// Copyright (C) 2023 Storj Labs, Inc.
+// See LICENSE for copying information.
+
+package audit_test
+
+import (
+	"context"
+	"sort"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+
+	"storj.io/common/storj"
+	"storj.io/common/testcontext"
+	"storj.io/common/testrand"
+	"storj.io/storj/private/testplanet"
+	"storj.io/storj/satellite/audit"
+	"storj.io/storj/satellite/overlay"
+)
+
+func TestContainmentSyncChore(t *testing.T) {
+	testplanet.Run(t, testplanet.Config{
+		SatelliteCount:   1,
+		StorageNodeCount: 3,
+	}, func(t *testing.T, ctx *testcontext.Context, planet *testplanet.Planet) {
+		satellite := planet.Satellites[0]
+		reverifyQueue := satellite.Audit.ReverifyQueue
+		cache := satellite.Overlay.DB
+		syncChore := satellite.Audit.ContainmentSyncChore
+		syncChore.Loop.Pause()
+
+		node1 := planet.StorageNodes[0].ID()
+		node2 := planet.StorageNodes[1].ID()
+		node3 := planet.StorageNodes[2].ID()
+
+		// no nodes should be in the reverify queue
+		requireInReverifyQueue(ctx, t, reverifyQueue)
+		// and none should be marked contained in the overlay
+		requireContainedStatus(ctx, t, cache, node1, false, node2, false, node3, false)
+
+		// set node1 contained in the overlay, but node2 contained in the reverify queue
+		err := cache.SetNodeContained(ctx, node1, true)
+		require.NoError(t, err)
+		node2Piece := &audit.PieceLocator{StreamID: testrand.UUID(), NodeID: node2}
+		err = reverifyQueue.Insert(ctx, node2Piece)
+		require.NoError(t, err)
+		requireInReverifyQueue(ctx, t, reverifyQueue, node2)
+		requireContainedStatus(ctx, t, cache, node1, true, node2, false, node3, false)
+
+		// run the chore to synchronize
+		syncChore.Loop.TriggerWait()
+
+		// there should only be node2 in both places now
+		requireInReverifyQueue(ctx, t, reverifyQueue, node2)
+		requireContainedStatus(ctx, t, cache, node1, false, node2, true, node3, false)
+
+		// now get node3 in the reverify queue as well
+		node3Piece := &audit.PieceLocator{StreamID: testrand.UUID(), NodeID: node3}
+		err = reverifyQueue.Insert(ctx, node3Piece)
+		require.NoError(t, err)
+		requireInReverifyQueue(ctx, t, reverifyQueue, node2, node3)
+		requireContainedStatus(ctx, t, cache, node1, false, node2, true, node3, false)
+
+		// run the chore to synchronize
+		syncChore.Loop.TriggerWait()
+
+		// nodes 2 and 3 should both be contained in both places now
+		requireInReverifyQueue(ctx, t, reverifyQueue, node2, node3)
+		requireContainedStatus(ctx, t, cache, node1, false, node2, true, node3, true)
+
+		// remove both node2 and node3 from reverify queue
+		wasDeleted, err := reverifyQueue.Remove(ctx, node2Piece)
+		require.NoError(t, err)
+		require.True(t, wasDeleted)
+		wasDeleted, err = reverifyQueue.Remove(ctx, node3Piece)
+		require.NoError(t, err)
+		require.True(t, wasDeleted)
+		requireInReverifyQueue(ctx, t, reverifyQueue)
+		requireContainedStatus(ctx, t, cache, node1, false, node2, true, node3, true)
+
+		// run the chore to synchronize
+		syncChore.Loop.TriggerWait()
+
+		// nothing should be contained in either place now
+		requireInReverifyQueue(ctx, t, reverifyQueue)
+		requireContainedStatus(ctx, t, cache, node1, false, node2, false, node3, false)
+	})
+}
+
+func requireInReverifyQueue(ctx context.Context, t testing.TB, reverifyQueue audit.ReverifyQueue, expectedNodes ...storj.NodeID) {
+	nodesInReverifyQueue, err := reverifyQueue.GetAllContainedNodes(ctx)
+	require.NoError(t, err)
+
+	sort.Slice(nodesInReverifyQueue, func(i, j int) bool {
+		return nodesInReverifyQueue[i].Compare(nodesInReverifyQueue[j]) < 0
+	})
+	sort.Slice(nodesInReverifyQueue, func(i, j int) bool {
+		return expectedNodes[i].Compare(expectedNodes[j]) < 0
+	})
+	require.Equal(t, expectedNodes, nodesInReverifyQueue)
+}
+
+func requireContainedStatus(ctx context.Context, t testing.TB, cache overlay.DB, args ...interface{}) {
+	require.Equal(t, 0, len(args)%2, "must be given an even number of args")
+	for n := 0; n < len(args); n += 2 {
+		nodeID := args[n].(storj.NodeID)
+		expectedContainment := args[n+1].(bool)
+		nodeInDB, err := cache.Get(ctx, nodeID)
+		require.NoError(t, err)
+		require.Equalf(t, expectedContainment, nodeInDB.Contained,
+			"Expected nodeID %v (args[%d]) contained = %v, but got %v",
+			nodeID, n, expectedContainment, nodeInDB.Contained)
+	}
+}

satellite/audit/worker.go

@@ -34,6 +34,8 @@ type Config struct {
 
 	ReverifyWorkerConcurrency   int           `help:"number of workers to run reverify audits on pieces" default:"2"`
 	ReverificationRetryInterval time.Duration `help:"how long a single reverification job can take before it may be taken over by another worker" releaseDefault:"6h" devDefault:"10m"`
+
+	ContainmentSyncChoreInterval time.Duration `help:"how often to run the containment-sync chore" releaseDefault:"2h" devDefault:"2m" testDefault:"$TESTINTERVAL"`
 }
 
 // Worker contains information for populating audit queue and processing audits.

satellite/core.go

@@ -117,7 +117,9 @@ type Core struct {
 
 	Audit struct {
 		VerifyQueue          audit.VerifyQueue
+		ReverifyQueue        audit.ReverifyQueue
 		Chore                *audit.Chore
+		ContainmentSyncChore *audit.ContainmentSyncChore
 	}
 
 	ExpiredDeletion struct {
@@ -401,6 +403,7 @@ func New(log *zap.Logger, full *identity.FullIdentity, db DB,
 		config := config.Audit
 
 		peer.Audit.VerifyQueue = db.VerifyQueue()
+		peer.Audit.ReverifyQueue = db.ReverifyQueue()
 
 		if config.UseRangedLoop {
 			peer.Log.Named("audit:chore").Info("using ranged loop")
@@ -417,6 +420,18 @@ func New(log *zap.Logger, full *identity.FullIdentity, db DB,
 			})
 			peer.Debug.Server.Panel.Add(
 				debug.Cycle("Audit Chore", peer.Audit.Chore.Loop))
+
+			peer.Audit.ContainmentSyncChore = audit.NewContainmentSyncChore(peer.Log.Named("audit:containment-sync-chore"),
+				peer.Audit.ReverifyQueue,
+				peer.Overlay.DB,
+				config.ContainmentSyncChoreInterval,
+			)
+			peer.Services.Add(lifecycle.Item{
+				Name: "audit:containment-sync-chore",
+				Run:  peer.Audit.ContainmentSyncChore.Run,
+			})
+			peer.Debug.Server.Panel.Add(
+				debug.Cycle("Audit Containment Sync Chore", peer.Audit.ContainmentSyncChore.Loop))
 		}
 	}
 

scripts/testdata/satellite-config.yaml.lock

@@ -37,6 +37,9 @@
 # how often to run the reservoir chore
 # audit.chore-interval: 24h0m0s
 
+# how often to run the containment-sync chore
+# audit.containment-sync-chore-interval: 2h0m0s
+
 # max number of times to attempt updating a statdb batch
 # audit.max-retries-stat-db: 3