HBASE-19359 Revisit the default config of hbase client retries number

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

@@ -2628,4 +2628,4 @@ public interface Admin extends Abortable, Closeable {
    * @return List of servers that are not cleared
    */
   List<ServerName> clearDeadServers(final List<ServerName> servers) throws IOException;
-}
+}

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

@@ -116,7 +116,8 @@ public final class ConnectionUtils {
       HConstants.DEFAULT_HBASE_CLIENT_RETRIES_NUMBER);
     // Go big. Multiply by 10. If we can't get to meta after this many retries
     // then something seriously wrong.
-    int serversideMultiplier = c.getInt("hbase.client.serverside.retries.multiplier", 10);
+    int serversideMultiplier = c.getInt(HConstants.HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER,
+      HConstants.DEFAULT_HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER);
     int retries = hcRetries * serversideMultiplier;
     c.setInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, retries);
     log.info(sn + " server-side Connection retries=" + retries);

hbase-common/src/main/java/org/apache/hadoop/hbase/HConstants.java

@@ -790,7 +790,12 @@ public final class HConstants {
   /**
    * Default value of {@link #HBASE_CLIENT_RETRIES_NUMBER}.
    */
-  public static final int DEFAULT_HBASE_CLIENT_RETRIES_NUMBER = 35;
+  public static final int DEFAULT_HBASE_CLIENT_RETRIES_NUMBER = 10;
+
+  public static final String HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER =
+      "hbase.client.serverside.retries.multiplier";
+
+  public static final int DEFAULT_HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER = 3;
 
   /**
    * Parameter name to set the default scanner caching for all clients.

hbase-common/src/main/resources/hbase-default.xml

@@ -488,7 +488,7 @@ possible configurations would overwhelm and obscure the important.
   </property>
   <property>
     <name>hbase.client.retries.number</name>
-    <value>35</value>
+    <value>10</value>
     <description>Maximum retries.  Used as maximum for all retryable
     operations such as the getting of a cell's value, starting a row update,
     etc.  Retry interval is a rough function based on hbase.client.pause.  At

hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java

@@ -3533,4 +3533,4 @@ public class HMaster extends HRegionServer implements MasterServices {
   public SpaceQuotaSnapshotNotifier getSpaceQuotaSnapshotNotifier() {
     return this.spaceQuotaSnapshotNotifier;
   }
-}
+}

hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java

@@ -1957,7 +1957,7 @@ public class HRegionServer extends HasThread implements
         conf.getInt("hbase.log.replay.retries.number", 8)); // 8 retries take about 23 seconds
     sinkConf.setInt(HConstants.HBASE_RPC_TIMEOUT_KEY,
         conf.getInt("hbase.log.replay.rpc.timeout", 30000)); // default 30 seconds
-    sinkConf.setInt("hbase.client.serverside.retries.multiplier", 1);
+    sinkConf.setInt(HConstants.HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER, 1);
     if (this.csm != null) {
       // SplitLogWorker needs csm. If none, don't start this.
       this.splitLogWorker = new SplitLogWorker(this, sinkConf, this,

hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/handler/RegionReplicaFlushHandler.java

@@ -96,7 +96,8 @@ public class RegionReplicaFlushHandler extends EventHandler {
     int numRetries = conf.getInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER,
       HConstants.DEFAULT_HBASE_CLIENT_RETRIES_NUMBER);
     if (numRetries > 10) {
-      int mult = conf.getInt("hbase.client.serverside.retries.multiplier", 10);
+      int mult = conf.getInt(HConstants.HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER,
+        HConstants.DEFAULT_HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER);
       numRetries = numRetries / mult; // reset if HRS has multiplied this already
     }
     return numRetries;

hbase-server/src/main/java/org/apache/hadoop/hbase/replication/regionserver/RegionReplicaReplicationEndpoint.java

@@ -119,11 +119,12 @@ public class RegionReplicaReplicationEndpoint extends HBaseReplicationEndpoint {
     int defaultNumRetries = conf.getInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER,
       HConstants.DEFAULT_HBASE_CLIENT_RETRIES_NUMBER);
     if (defaultNumRetries > 10) {
-      int mult = conf.getInt("hbase.client.serverside.retries.multiplier", 10);
+      int mult = conf.getInt(HConstants.HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER,
+        HConstants.DEFAULT_HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER);
       defaultNumRetries = defaultNumRetries / mult; // reset if HRS has multiplied this already
     }
 
-    conf.setInt("hbase.client.serverside.retries.multiplier", 1);
+    conf.setInt(HConstants.HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER, 1);
     int numRetries = conf.getInt(CLIENT_RETRIES_NUMBER, defaultNumRetries);
     conf.setInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, numRetries);
 

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

@@ -101,7 +101,7 @@ public class TestRegionReplicaReplicationEndpoint {
     conf.setInt("replication.stats.thread.period.seconds", 5);
     conf.setBoolean("hbase.tests.use.shortcircuit.reads", false);
     conf.setInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, 5); // less number of retries is needed
-    conf.setInt("hbase.client.serverside.retries.multiplier", 1);
+    conf.setInt(HConstants.HBASE_CLIENT_SERVERSIDE_RETRIES_MULTIPLIER, 1);
 
     HTU.startMiniCluster(NB_SERVERS);
   }