diff --git a/hbase-client/src/main/java/org/apache/hadoop/hbase/HTableDescriptor.java b/hbase-client/src/main/java/org/apache/hadoop/hbase/HTableDescriptor.java
index 17956ff7590..00578dedead 100644
--- a/hbase-client/src/main/java/org/apache/hadoop/hbase/HTableDescriptor.java
+++ b/hbase-client/src/main/java/org/apache/hadoop/hbase/HTableDescriptor.java
@@ -42,6 +42,7 @@ import org.apache.hadoop.hbase.classification.InterfaceStability;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.Path;
 import org.apache.hadoop.hbase.client.Durability;
+import org.apache.hadoop.hbase.client.RegionReplicaUtil;
 import org.apache.hadoop.hbase.exceptions.DeserializationException;
 import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
 import org.apache.hadoop.hbase.protobuf.ProtobufUtil;
@@ -1195,6 +1196,10 @@ public class HTableDescriptor implements WritableComparable<HTableDescriptor> {
    */
   public HTableDescriptor setRegionMemstoreReplication(boolean memstoreReplication) {
     setValue(REGION_MEMSTORE_REPLICATION_KEY, memstoreReplication ? TRUE : FALSE);
+    // If the memstore replication is setup, we do not have to wait for observing a flush event
+    // from primary before starting to serve reads, because gaps from replication is not applicable
+    setConfiguration(RegionReplicaUtil.REGION_REPLICA_WAIT_FOR_PRIMARY_FLUSH_CONF_KEY,
+      Boolean.toString(memstoreReplication));
     return this;
   }
 
diff --git a/hbase-client/src/main/java/org/apache/hadoop/hbase/client/RegionReplicaUtil.java b/hbase-client/src/main/java/org/apache/hadoop/hbase/client/RegionReplicaUtil.java
index bc7c0ecef1d..70a7a307874 100644
--- a/hbase-client/src/main/java/org/apache/hadoop/hbase/client/RegionReplicaUtil.java
+++ b/hbase-client/src/main/java/org/apache/hadoop/hbase/client/RegionReplicaUtil.java
@@ -31,6 +31,19 @@ import org.apache.hadoop.hbase.util.Bytes;
 @InterfaceAudience.Private
 public class RegionReplicaUtil {
 
+  /**
+   * Whether or not the secondary region will wait for observing a flush / region open event
+   * from the primary region via async wal replication before enabling read requests. Since replayed
+   * edits from async wal replication from primary is not persisted in WAL, the memstore of the
+   * secondary region might be non-empty at the time of close or crash. For ensuring seqId's not
+   * "going back in time" in the secondary region replica, this should be enabled. However, in some
+   * cases the above semantics might be ok for some application classes.
+   * See HBASE-11580 for more context.
+   */
+  public static final String REGION_REPLICA_WAIT_FOR_PRIMARY_FLUSH_CONF_KEY
+    = "hbase.region.replica.wait.for.primary.flush";
+  protected static final boolean DEFAULT_REGION_REPLICA_WAIT_FOR_PRIMARY_FLUSH = true;
+
   /**
    * The default replicaId for the region
    */
diff --git a/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java b/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
index d5f694a8312..6ee60b01ef6 100644
--- a/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
+++ b/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
@@ -1915,9 +1915,9 @@ public class HRegionServer extends HasThread implements
     if (ServerRegionReplicaUtil.isDefaultReplica(region.getRegionInfo())) {
       return;
     }
-    if (!ServerRegionReplicaUtil.isRegionReplicaReplicationEnabled(getConfiguration()) ||
+    if (!ServerRegionReplicaUtil.isRegionReplicaReplicationEnabled(region.conf) ||
         !ServerRegionReplicaUtil.isRegionReplicaWaitForPrimaryFlushEnabled(
-          getConfiguration())) {
+          region.conf)) {
       region.setReadsEnabled(true);
       return;
     }
diff --git a/hbase-server/src/main/java/org/apache/hadoop/hbase/util/ServerRegionReplicaUtil.java b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/ServerRegionReplicaUtil.java
index 45517220575..67f8e84f24c 100644
--- a/hbase-server/src/main/java/org/apache/hadoop/hbase/util/ServerRegionReplicaUtil.java
+++ b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/ServerRegionReplicaUtil.java
@@ -55,19 +55,6 @@ public class ServerRegionReplicaUtil extends RegionReplicaUtil {
   private static final boolean DEFAULT_REGION_REPLICA_REPLICATION = false;
   private static final String REGION_REPLICA_REPLICATION_PEER = "region_replica_replication";
 
-  /**
-   * Whether or not the secondary region will wait for observing a flush / region open event
-   * from the primary region via async wal replication before enabling read requests. Since replayed
-   * edits from async wal replication from primary is not persisted in WAL, the memstore of the
-   * secondary region might be non-empty at the time of close or crash. For ensuring seqId's not
-   * "going back in time" in the secondary region replica, this should be enabled. However, in some
-   * cases the above semantics might be ok for some application classes.
-   * See HBASE-11580 for more context.
-   */
-  public static final String REGION_REPLICA_WAIT_FOR_PRIMARY_FLUSH_CONF_KEY
-    = "hbase.region.replica.wait.for.primary.flush";
-  private static final boolean DEFAULT_REGION_REPLICA_WAIT_FOR_PRIMARY_FLUSH = true;
-
   /**
    * Enables or disables refreshing store files of secondary region replicas when the memory is
    * above the global memstore lower limit. Refreshing the store files means that we will do a file