HBASE-26649 Support meta replica LoadBalance mode for RegionLocator#getAllRegionLocations() (#4442)

Signed-off-by: Duo Zhang <zhangduo@apache.org>

hbase-client/src/main/java/org/apache/hadoop/hbase/CatalogReplicaMode.java

@@ -15,7 +15,7 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
-package org.apache.hadoop.hbase.client;
+package org.apache.hadoop.hbase;
 
 import org.apache.yetus.audience.InterfaceAudience;
 
@@ -33,7 +33,7 @@ import org.apache.yetus.audience.InterfaceAudience;
  * </ol>
  */
 @InterfaceAudience.Private
-enum CatalogReplicaMode {
+public enum CatalogReplicaMode {
   NONE {
     @Override
     public String toString() {

hbase-client/src/main/java/org/apache/hadoop/hbase/ClientMetaTableAccessor.java

@@ -17,6 +17,7 @@
  */
 package org.apache.hadoop.hbase;
 
+import static org.apache.hadoop.hbase.client.RegionLocator.LOCATOR_META_REPLICAS_MODE;
 import static org.apache.hadoop.hbase.util.FutureUtils.addListener;
 
 import java.io.Closeable;
@@ -26,6 +27,7 @@ import java.util.Collections;
 import java.util.List;
 import java.util.Optional;
 import java.util.concurrent.CompletableFuture;
+import java.util.concurrent.ThreadLocalRandom;
 import java.util.stream.Collectors;
 import org.apache.hadoop.hbase.client.AdvancedScanResultConsumer;
 import org.apache.hadoop.hbase.client.AsyncTable;
@@ -257,7 +259,37 @@ public final class ClientMetaTableAccessor {
     }
 
     CompletableFuture<Void> future = new CompletableFuture<Void>();
-    metaTable.scan(scan, new MetaTableScanResultConsumer(rowUpperLimit, visitor, future));
+    // Get the region locator's meta replica mode.
+    CatalogReplicaMode metaReplicaMode = CatalogReplicaMode.fromString(metaTable.getConfiguration()
+      .get(LOCATOR_META_REPLICAS_MODE, CatalogReplicaMode.NONE.toString()));
+
+    if (metaReplicaMode == CatalogReplicaMode.LOAD_BALANCE) {
+      addListener(metaTable.getDescriptor(), (desc, error) -> {
+        if (error != null) {
+          LOG.error("Failed to get meta table descriptor, error: ", error);
+          future.completeExceptionally(error);
+          return;
+        }
+
+        int numOfReplicas = desc.getRegionReplication();
+        if (numOfReplicas > 1) {
+          int replicaId = ThreadLocalRandom.current().nextInt(numOfReplicas);
+
+          // When the replicaId is 0, do not set to Consistency.TIMELINE
+          if (replicaId > 0) {
+            scan.setReplicaId(replicaId);
+            scan.setConsistency(Consistency.TIMELINE);
+          }
+        }
+        metaTable.scan(scan, new MetaTableScanResultConsumer(rowUpperLimit, visitor, future));
+      });
+    } else {
+      if (metaReplicaMode == CatalogReplicaMode.HEDGED_READ) {
+        scan.setConsistency(Consistency.TIMELINE);
+      }
+      metaTable.scan(scan, new MetaTableScanResultConsumer(rowUpperLimit, visitor, future));
+    }
+
     return future;
   }
 

hbase-client/src/main/java/org/apache/hadoop/hbase/client/AsyncNonMetaRegionLocator.java

@@ -50,6 +50,7 @@ import java.util.concurrent.ConcurrentSkipListMap;
 import java.util.concurrent.TimeUnit;
 import org.apache.commons.lang3.ObjectUtils;
 import org.apache.hadoop.hbase.CatalogFamilyFormat;
+import org.apache.hadoop.hbase.CatalogReplicaMode;
 import org.apache.hadoop.hbase.HBaseIOException;
 import org.apache.hadoop.hbase.HConstants;
 import org.apache.hadoop.hbase.HRegionLocation;

hbase-server/src/test/java/org/apache/hadoop/hbase/client/TestAsyncNonMetaRegionLocator.java

@@ -37,6 +37,7 @@ import java.util.concurrent.ExecutionException;
 import java.util.concurrent.ThreadLocalRandom;
 import java.util.stream.IntStream;
 import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.hbase.CatalogReplicaMode;
 import org.apache.hadoop.hbase.HBaseClassTestRule;
 import org.apache.hadoop.hbase.HBaseTestingUtil;
 import org.apache.hadoop.hbase.HRegionLocation;

hbase-server/src/test/java/org/apache/hadoop/hbase/replication/regionserver/TestMetaRegionReplicaReplication.java

@@ -43,6 +43,7 @@ import org.apache.hadoop.hbase.client.Connection;
 import org.apache.hadoop.hbase.client.ConnectionFactory;
 import org.apache.hadoop.hbase.client.Get;
 import org.apache.hadoop.hbase.client.RegionInfo;
+import org.apache.hadoop.hbase.client.RegionLocator;
 import org.apache.hadoop.hbase.client.Result;
 import org.apache.hadoop.hbase.client.Scan;
 import org.apache.hadoop.hbase.client.Table;
@@ -396,11 +397,8 @@ public class TestMetaRegionReplicaReplication {
   }
 
   private void primaryIncreaseReplicaIncrease(final long[] before, final long[] after) {
-    // There are read requests increase for primary meta replica.
+    // There are read requests increase for all meta replica regions,
-    assertTrue(after[RegionInfo.DEFAULT_REPLICA_ID] > before[RegionInfo.DEFAULT_REPLICA_ID]);
+    for (int i = 0; i < after.length; i++) {
-
-    // There are read requests incrase for meta replica regions.
-    for (int i = 1; i < after.length; i++) {
       assertTrue(after[i] > before[i]);
     }
   }
@@ -420,6 +418,7 @@ public class TestMetaRegionReplicaReplication {
     final Region[] metaRegions = getAllRegions(TableName.META_TABLE_NAME, numOfMetaReplica);
     long[] readReqsForMetaReplicas = new long[numOfMetaReplica];
     long[] readReqsForMetaReplicasAfterGet = new long[numOfMetaReplica];
+    long[] readReqsForMetaReplicasAfterGetAllLocations = new long[numOfMetaReplica];
     long[] readReqsForMetaReplicasAfterMove = new long[numOfMetaReplica];
     long[] readReqsForMetaReplicasAfterSecondMove = new long[numOfMetaReplica];
     long[] readReqsForMetaReplicasAfterThirdGet = new long[numOfMetaReplica];
@@ -468,6 +467,16 @@ public class TestMetaRegionReplicaReplication {
       // There are more reads against all meta replica regions, including the primary region.
       primaryIncreaseReplicaIncrease(readReqsForMetaReplicas, readReqsForMetaReplicasAfterGet);
 
+      RegionLocator locator = tableForGet.getRegionLocator();
+
+      for (int j = 0; j < numOfMetaReplica * 3; j ++) {
+        locator.getAllRegionLocations();
+      }
+
+      getMetaReplicaReadRequests(metaRegions, readReqsForMetaReplicasAfterGetAllLocations);
+      primaryIncreaseReplicaIncrease(readReqsForMetaReplicasAfterGet,
+        readReqsForMetaReplicasAfterGetAllLocations);
+
       // move one of regions so it meta cache may be invalid.
       HTU.moveRegionAndWait(userRegion.getRegionInfo(), destRs.getServerName());
 
@@ -477,7 +486,7 @@ public class TestMetaRegionReplicaReplication {
 
       // There are read requests increase for primary meta replica.
       // For rest of meta replicas, there is no change as regionMove will tell the new location
-      primaryIncreaseReplicaNoChange(readReqsForMetaReplicasAfterGet,
+      primaryIncreaseReplicaNoChange(readReqsForMetaReplicasAfterGetAllLocations,
         readReqsForMetaReplicasAfterMove);
       // Move region again.
       HTU.moveRegionAndWait(userRegion.getRegionInfo(), srcRs.getServerName());