HDDS-2047. Datanodes fail to come up after 10 retries in a secure env… (#1379)

hadoop-hdds/common/src/main/java/org/apache/hadoop/hdds/HddsUtils.java

@@ -30,6 +30,7 @@
 import java.util.Map;
 import java.util.Optional;
 import java.util.TimeZone;
+import java.util.concurrent.TimeUnit;
 
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
@@ -43,6 +44,8 @@
 import org.apache.hadoop.hdds.protocol.SCMSecurityProtocol;
 import org.apache.hadoop.hdds.scm.protocolPB.ScmBlockLocationProtocolPB;
 import org.apache.hadoop.hdfs.DFSConfigKeys;
+import org.apache.hadoop.io.retry.RetryPolicies;
+import org.apache.hadoop.io.retry.RetryPolicy;
 import org.apache.hadoop.ipc.Client;
 import org.apache.hadoop.ipc.ProtobufRpcEngine;
 import org.apache.hadoop.ipc.RPC;
@@ -177,23 +180,27 @@ public static InetSocketAddress getScmAddressForBlockClients(
   /**
    * Create a scm security client.
    * @param conf    - Ozone configuration.
-   * @param address - inet socket address of scm.
    *
    * @return {@link SCMSecurityProtocol}
    * @throws IOException
    */
   public static SCMSecurityProtocolClientSideTranslatorPB getScmSecurityClient(
-      OzoneConfiguration conf, InetSocketAddress address) throws IOException {
+      OzoneConfiguration conf) throws IOException {
     RPC.setProtocolEngine(conf, SCMSecurityProtocolPB.class,
         ProtobufRpcEngine.class);
     long scmVersion =
         RPC.getProtocolVersion(ScmBlockLocationProtocolPB.class);
+    InetSocketAddress address =
+        getScmAddressForSecurityProtocol(conf);
+    RetryPolicy retryPolicy =
+        RetryPolicies.retryForeverWithFixedSleep(
+            1000, TimeUnit.MILLISECONDS);
     SCMSecurityProtocolClientSideTranslatorPB scmSecurityClient =
         new SCMSecurityProtocolClientSideTranslatorPB(
-            RPC.getProxy(SCMSecurityProtocolPB.class, scmVersion,
+            RPC.getProtocolProxy(SCMSecurityProtocolPB.class, scmVersion,
                 address, UserGroupInformation.getCurrentUser(),
                 conf, NetUtils.getDefaultSocketFactory(conf),
-                Client.getRpcTimeout(conf)));
+                Client.getRpcTimeout(conf), retryPolicy).getProxy());
     return scmSecurityClient;
   }
 

hadoop-hdds/container-service/src/main/java/org/apache/hadoop/ozone/HddsDatanodeService.java

@@ -272,8 +272,7 @@ private void getSCMSignedCert(OzoneConfiguration config) {
       PKCS10CertificationRequest csr = getCSR(config);
       // TODO: For SCM CA we should fetch certificate from multiple SCMs.
       SCMSecurityProtocolClientSideTranslatorPB secureScmClient =
-          HddsUtils.getScmSecurityClient(config,
+          HddsUtils.getScmSecurityClient(config);
-              HddsUtils.getScmAddressForSecurityProtocol(config));
       SCMGetCertResponseProto response = secureScmClient.
           getDataNodeCertificateChain(datanodeDetails.getProtoBufMessage(),
               getEncodedString(csr));

hadoop-ozone/ozone-manager/src/main/java/org/apache/hadoop/ozone/om/OzoneManager.java

@@ -48,10 +48,8 @@
 import org.apache.hadoop.hdds.conf.OzoneConfiguration;
 import org.apache.hadoop.hdds.protocol.DatanodeDetails;
 import org.apache.hadoop.hdds.protocol.proto.HddsProtos;
-import org.apache.hadoop.hdds.protocol.SCMSecurityProtocol;
 import org.apache.hadoop.hdds.protocol.proto.SCMSecurityProtocolProtos.SCMGetCertResponseProto;
 import org.apache.hadoop.hdds.protocolPB.SCMSecurityProtocolClientSideTranslatorPB;
-import org.apache.hadoop.hdds.protocolPB.SCMSecurityProtocolPB;
 import org.apache.hadoop.hdds.scm.ScmInfo;
 import org.apache.hadoop.hdds.scm.container.common.helpers.ExcludeList;
 import org.apache.hadoop.hdds.scm.protocol.ScmBlockLocationProtocol;
@@ -183,7 +181,6 @@
 import static org.apache.hadoop.hdds.HddsConfigKeys.HDDS_BLOCK_TOKEN_ENABLED_DEFAULT;
 import static org.apache.hadoop.hdds.HddsUtils.getScmAddressForBlockClients;
 import static org.apache.hadoop.hdds.HddsUtils.getScmAddressForClients;
-import static org.apache.hadoop.hdds.HddsUtils.getScmAddressForSecurityProtocol;
 import static org.apache.hadoop.hdds.protocol.proto.HddsProtos.NodeState.HEALTHY;
 import static org.apache.hadoop.hdds.security.x509.certificates.utils.CertificateSignRequest.getEncodedString;
 import static org.apache.hadoop.hdds.server.ServerUtils.updateRPCListenAddress;
@@ -951,29 +948,6 @@ private static ScmBlockLocationProtocol getScmBlockClient(
             conf);
   }
 
-  /**
-   * Create a scm security client, used to get SCM signed certificate.
-   *
-   * @return {@link SCMSecurityProtocol}
-   * @throws IOException
-   */
-  private static SCMSecurityProtocolClientSideTranslatorPB
-      getScmSecurityClient(OzoneConfiguration conf) throws IOException {
-    RPC.setProtocolEngine(conf, SCMSecurityProtocolPB.class,
-        ProtobufRpcEngine.class);
-    long scmVersion =
-        RPC.getProtocolVersion(ScmBlockLocationProtocolPB.class);
-    InetSocketAddress scmSecurityProtoAdd =
-        getScmAddressForSecurityProtocol(conf);
-    SCMSecurityProtocolClientSideTranslatorPB scmSecurityClient =
-        new SCMSecurityProtocolClientSideTranslatorPB(
-            RPC.getProxy(SCMSecurityProtocolPB.class, scmVersion,
-                scmSecurityProtoAdd, UserGroupInformation.getCurrentUser(),
-                conf, NetUtils.getDefaultSocketFactory(conf),
-                Client.getRpcTimeout(conf)));
-    return scmSecurityClient;
-  }
-
   /**
    * Returns a scm container client.
    *
@@ -1563,7 +1537,7 @@ private static void getSCMSignedCert(CertificateClient client,
         omDetailsProtoBuilder.build();
     LOG.info("OzoneManager ports added:{}", omDetailsProto.getPortsList());
     SCMSecurityProtocolClientSideTranslatorPB secureScmClient =
-        getScmSecurityClient(config);
+        HddsUtils.getScmSecurityClient(config);
 
     SCMGetCertResponseProto response = secureScmClient.
         getOMCertChain(omDetailsProto, getEncodedString(csr));