Segment Replication - Remove redundant replica doc parsing on writes. (opensearch-project#7279)

This change removes unnecessary doc parsing currently performed on replicas by
updating applyIndexOperationOnReplicas to pass a doc id from the primary.

Signed-off-by: Marc Handalian <handalm@amazon.com>

Author: mch2

server/src/main/java/org/opensearch/action/bulk/TransportShardBulkAction.java

@@ -857,6 +857,7 @@ private static Engine.Result performOpOnReplica(
                     indexRequest.routing()
                 );
                 result = replica.applyIndexOperationOnReplica(
+                    primaryResponse.getId(),
                     primaryResponse.getSeqNo(),
                     primaryResponse.getPrimaryTerm(),
                     primaryResponse.getVersion(),

server/src/main/java/org/opensearch/index/shard/IndexShard.java

@@ -911,13 +911,31 @@ public Engine.IndexResult applyIndexOperationOnPrimary(
     }
 
     public Engine.IndexResult applyIndexOperationOnReplica(
+        String id,
         long seqNo,
         long opPrimaryTerm,
         long version,
         long autoGeneratedTimeStamp,
         boolean isRetry,
         SourceToParse sourceToParse
     ) throws IOException {
+        if (indexSettings.isSegRepEnabled()) {
+            Engine.Index index = new Engine.Index(
+                new Term(IdFieldMapper.NAME, Uid.encodeId(id)),
+                new ParsedDocument(null, null, id, null, null, sourceToParse.source(), sourceToParse.getXContentType(), null),
+                seqNo,
+                opPrimaryTerm,
+                version,
+                null,
+                Engine.Operation.Origin.REPLICA,
+                System.nanoTime(),
+                autoGeneratedTimeStamp,
+                isRetry,
+                UNASSIGNED_SEQ_NO,
+                0
+            );
+            return getEngine().index(index);
+        }
         return applyIndexOperation(
             getEngine(),
             seqNo,
@@ -1128,6 +1146,21 @@ public Engine.DeleteResult applyDeleteOperationOnPrimary(
     }
 
     public Engine.DeleteResult applyDeleteOperationOnReplica(long seqNo, long opPrimaryTerm, long version, String id) throws IOException {
+        if (indexSettings.isSegRepEnabled()) {
+            final Engine.Delete delete = new Engine.Delete(
+                id,
+                new Term(IdFieldMapper.NAME, Uid.encodeId(id)),
+                seqNo,
+                opPrimaryTerm,
+                version,
+                null,
+                Engine.Operation.Origin.REPLICA,
+                System.nanoTime(),
+                UNASSIGNED_SEQ_NO,
+                0
+            );
+            return getEngine().delete(delete);
+        }
         return applyDeleteOperation(
             getEngine(),
             seqNo,

server/src/test/java/org/opensearch/index/replication/RecoveryDuringReplicationTests.java

@@ -152,6 +152,7 @@ public void testRecoveryToReplicaThatReceivedExtraDocument() throws Exception {
             final IndexShard remainingReplica = shards.getReplicas().get(1);
             // slip the extra document into the replica
             remainingReplica.applyIndexOperationOnReplica(
+                "id",
                 remainingReplica.getLocalCheckpoint() + 1,
                 remainingReplica.getOperationPrimaryTerm(),
                 1,

server/src/test/java/org/opensearch/index/shard/IndexShardTests.java

@@ -170,6 +170,7 @@
 import java.util.Locale;
 import java.util.Map;
 import java.util.Set;
+import java.util.UUID;
 import java.util.concurrent.BrokenBarrierException;
 import java.util.concurrent.CountDownLatch;
 import java.util.concurrent.CyclicBarrier;
@@ -2297,6 +2298,7 @@ public void testRecoverFromStoreWithOutOfOrderDelete() throws IOException {
         shard.applyDeleteOperationOnReplica(1, primaryTerm, 2, "id");
         shard.getEngine().translogManager().rollTranslogGeneration(); // isolate the delete in it's own generation
         shard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             0,
             primaryTerm,
             1,
@@ -2305,6 +2307,7 @@ public void testRecoverFromStoreWithOutOfOrderDelete() throws IOException {
             new SourceToParse(shard.shardId().getIndexName(), "id", new BytesArray("{}"), XContentType.JSON)
         );
         shard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             3,
             primaryTerm,
             3,
@@ -2315,6 +2318,7 @@ public void testRecoverFromStoreWithOutOfOrderDelete() throws IOException {
         // Flushing a new commit with local checkpoint=1 allows to skip the translog gen #1 in recovery.
         shard.flush(new FlushRequest().force(true).waitIfOngoing(true));
         shard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             2,
             primaryTerm,
             3,
@@ -2323,6 +2327,7 @@ public void testRecoverFromStoreWithOutOfOrderDelete() throws IOException {
             new SourceToParse(shard.shardId().getIndexName(), "id-2", new BytesArray("{}"), XContentType.JSON)
         );
         shard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             5,
             primaryTerm,
             1,
@@ -2470,6 +2475,7 @@ public void testRecoverFromStoreWithNoOps() throws IOException {
         updateMappings(otherShard, shard.indexSettings().getIndexMetadata());
         SourceToParse sourceToParse = new SourceToParse(shard.shardId().getIndexName(), "1", new BytesArray("{}"), XContentType.JSON);
         otherShard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             1,
             otherShard.getOperationPrimaryTerm(),
             1,
@@ -2597,6 +2603,7 @@ public void testRecoverFromStoreRemoveStaleOperations() throws Exception {
         final String indexName = shard.shardId().getIndexName();
         // Index #0, index #1
         shard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             0,
             primaryTerm,
             1,
@@ -2607,6 +2614,7 @@ public void testRecoverFromStoreRemoveStaleOperations() throws Exception {
         flushShard(shard);
         shard.updateGlobalCheckpointOnReplica(0, "test"); // stick the global checkpoint here.
         shard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             1,
             primaryTerm,
             1,
@@ -2619,6 +2627,7 @@ public void testRecoverFromStoreRemoveStaleOperations() throws Exception {
         shard.getEngine().translogManager().rollTranslogGeneration();
         shard.markSeqNoAsNoop(1, primaryTerm, "test");
         shard.applyIndexOperationOnReplica(
+            UUID.randomUUID().toString(),
             2,
             primaryTerm,
             1,
@@ -4009,6 +4018,7 @@ private Result indexOnReplicaWithGaps(final IndexShard indexShard, final int ope
                     XContentType.JSON
                 );
                 indexShard.applyIndexOperationOnReplica(
+                    UUID.randomUUID().toString(),
                     i,
                     indexShard.getOperationPrimaryTerm(),
                     1,
@@ -4633,6 +4643,7 @@ public void testDoNotTrimCommitsWhenOpenReadOnlyEngine() throws Exception {
                 seqNo++; // create gaps in sequence numbers
             }
             shard.applyIndexOperationOnReplica(
+                UUID.randomUUID().toString(),
                 seqNo,
                 shard.getOperationPrimaryTerm(),
                 1,

server/src/test/java/org/opensearch/index/shard/SegmentReplicationIndexShardTests.java

@@ -38,6 +38,8 @@
 import org.opensearch.index.replication.TestReplicationSource;
 import org.opensearch.index.store.Store;
 import org.opensearch.index.store.StoreFileMetadata;
+import org.opensearch.index.translog.SnapshotMatchers;
+import org.opensearch.index.translog.Translog;
 import org.opensearch.indices.recovery.RecoverySettings;
 import org.opensearch.indices.recovery.RecoveryTarget;
 import org.opensearch.indices.replication.CheckpointInfoResponse;
@@ -176,6 +178,7 @@ public void testSegmentReplication_Index_Update_Delete() throws Exception {
                 shards.index(new IndexRequest(index.getName()).id(String.valueOf(i)).source("{\"foo\": \"bar\"}", XContentType.JSON));
             }
 
+            assertEqualTranslogOperations(shards, primaryShard);
             primaryShard.refresh("Test");
             replicateSegments(primaryShard, shards.getReplicas());
 
@@ -189,7 +192,7 @@ public void testSegmentReplication_Index_Update_Delete() throws Exception {
                     );
                 }
             }
-
+            assertEqualTranslogOperations(shards, primaryShard);
             primaryShard.refresh("Test");
             replicateSegments(primaryShard, shards.getReplicas());
             shards.assertAllEqual(numDocs);
@@ -204,6 +207,7 @@ public void testSegmentReplication_Index_Update_Delete() throws Exception {
                     shards.delete(new DeleteRequest(index.getName()).id(String.valueOf(i)));
                 }
             }
+            assertEqualTranslogOperations(shards, primaryShard);
             primaryShard.refresh("Test");
             replicateSegments(primaryShard, shards.getReplicas());
             final List<DocIdSeqNoAndSource> docsAfterDelete = getDocIdAndSeqNos(primaryShard);
@@ -753,6 +757,7 @@ public void testNRTReplicaPromotedAsPrimary() throws Exception {
             for (IndexShard shard : shards.getReplicas()) {
                 assertDocCounts(shard, numDocs, numDocs);
             }
+            assertEqualTranslogOperations(shards, oldPrimary);
 
             // 2. Create ops that are in the replica's xlog, not in the index.
             // index some more into both but don't replicate. replica will have only numDocs searchable, but should have totalDocs
@@ -761,6 +766,7 @@ public void testNRTReplicaPromotedAsPrimary() throws Exception {
             final int totalDocs = numDocs + additonalDocs;
 
             assertDocCounts(oldPrimary, totalDocs, totalDocs);
+            assertEqualTranslogOperations(shards, oldPrimary);
             for (IndexShard shard : shards.getReplicas()) {
                 assertDocCounts(shard, totalDocs, numDocs);
             }
@@ -1083,4 +1089,20 @@ private void assertEqualCommittedSegments(IndexShard primary, IndexShard... repl
             assertTrue(diff.missing.isEmpty());
         }
     }
+
+    private void assertEqualTranslogOperations(ReplicationGroup shards, IndexShard primaryShard) throws IOException {
+        try (final Translog.Snapshot snapshot = getTranslog(primaryShard).newSnapshot()) {
+            List<Translog.Operation> operations = new ArrayList<>();
+            Translog.Operation op;
+            while ((op = snapshot.next()) != null) {
+                final Translog.Operation newOp = op;
+                operations.add(newOp);
+            }
+            for (IndexShard replica : shards.getReplicas()) {
+                try (final Translog.Snapshot replicaSnapshot = getTranslog(replica).newSnapshot()) {
+                    assertThat(replicaSnapshot, SnapshotMatchers.containsOperationsInAnyOrder(operations));
+                }
+            }
+        }
+    }
 }

server/src/test/java/org/opensearch/indices/recovery/PeerRecoveryTargetServiceTests.java

@@ -65,6 +65,7 @@
 import java.util.Collections;
 import java.util.List;
 import java.util.Optional;
+import java.util.UUID;
 import java.util.concurrent.ArrayBlockingQueue;
 import java.util.concurrent.BlockingQueue;
 import java.util.concurrent.CyclicBarrier;
@@ -182,6 +183,7 @@ private SeqNoStats populateRandomData(IndexShard shard) throws IOException {
         Randomness.shuffle(seqNos);
         for (long seqNo : seqNos) {
             shard.applyIndexOperationOnReplica(
+                UUID.randomUUID().toString(),
                 seqNo,
                 1,
                 shard.getOperationPrimaryTerm(),

server/src/test/java/org/opensearch/indices/recovery/RecoveryTests.java

@@ -78,6 +78,7 @@
 import java.util.HashMap;
 import java.util.List;
 import java.util.Map;
+import java.util.UUID;
 import java.util.concurrent.CountDownLatch;
 import java.util.concurrent.Future;
 import java.util.concurrent.atomic.AtomicReference;
@@ -181,6 +182,7 @@ public void testRecoveryWithOutOfOrderDeleteWithSoftDeletes() throws Exception {
             orgReplica.flush(new FlushRequest().force(true)); // isolate delete#1 in its own translog generation and lucene segment
             // index #0
             orgReplica.applyIndexOperationOnReplica(
+                UUID.randomUUID().toString(),
                 0,
                 primaryTerm,
                 1,
@@ -190,6 +192,7 @@ public void testRecoveryWithOutOfOrderDeleteWithSoftDeletes() throws Exception {
             );
             // index #3
             orgReplica.applyIndexOperationOnReplica(
+                UUID.randomUUID().toString(),
                 3,
                 primaryTerm,
                 1,
@@ -201,6 +204,7 @@ public void testRecoveryWithOutOfOrderDeleteWithSoftDeletes() throws Exception {
             orgReplica.flush(new FlushRequest().force(true).waitIfOngoing(true));
             // index #2
             orgReplica.applyIndexOperationOnReplica(
+                UUID.randomUUID().toString(),
                 2,
                 primaryTerm,
                 1,
@@ -212,6 +216,7 @@ public void testRecoveryWithOutOfOrderDeleteWithSoftDeletes() throws Exception {
             orgReplica.updateGlobalCheckpointOnReplica(3L, "test");
             // index #5 -> force NoOp #4.
             orgReplica.applyIndexOperationOnReplica(
+                UUID.randomUUID().toString(),
                 5,
                 primaryTerm,
                 1,

test/framework/src/main/java/org/opensearch/index/shard/IndexShardTestCase.java

@@ -151,6 +151,7 @@
 import java.util.List;
 import java.util.Map;
 import java.util.Set;
+import java.util.UUID;
 import java.util.concurrent.CountDownLatch;
 import java.util.concurrent.ExecutionException;
 import java.util.concurrent.TimeUnit;
@@ -1122,6 +1123,7 @@ protected Engine.IndexResult indexDoc(IndexShard shard, String id, String source
             final long seqNo = shard.seqNoStats().getMaxSeqNo() + 1;
             shard.advanceMaxSeqNoOfUpdatesOrDeletes(seqNo); // manually replicate max_seq_no_of_updates
             result = shard.applyIndexOperationOnReplica(
+                UUID.randomUUID().toString(),
                 seqNo,
                 shard.getOperationPrimaryTerm(),
                 0,