HADOOP-12825. Log slow name resolutions. (Sidharta Seethana via stevel)

hadoop-common-project/hadoop-common/CHANGES.txt

@@ -1160,6 +1160,9 @@ Release 2.8.0 - UNRELEASED
     HADOOP-12824. Collect network and disk usage on the node running Windows.
     (Inigo Goiri via xyao)
 
+    HADOOP-12825. Log slow name resolutions.
+    (Sidharta Seethana via stevel)
+
   OPTIMIZATIONS
 
     HADOOP-11785. Reduce the number of listStatus operation in distcp

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/CommonConfigurationKeys.java

@@ -204,6 +204,18 @@ public class CommonConfigurationKeys extends CommonConfigurationKeysPublic {
   public static final boolean HADOOP_SECURITY_TOKEN_SERVICE_USE_IP_DEFAULT =
       true;
 
+  /** See <a href="{@docRoot}/../core-default.html">core-default.xml .</a> */
+  public static final String HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_ENABLED_KEY =
+      "hadoop.security.dns.log-slow-lookups.enabled";
+  public static final boolean
+      HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_ENABLED_DEFAULT = false;
+  /** See <a href="{@docRoot}/../core-default.html">core-default.xml .</a> */
+  public static final String
+      HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_THRESHOLD_MS_KEY =
+      "hadoop.security.dns.log-slow-lookups.threshold.ms";
+  public static final int
+      HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_THRESHOLD_MS_DEFAULT = 1000;
+
   /**
    * HA health monitor and failover controller.
    */

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/security/SecurityUtil.java

@@ -30,6 +30,7 @@ import java.security.PrivilegedExceptionAction;
 import java.util.Arrays;
 import java.util.List;
 import java.util.ServiceLoader;
+import java.util.concurrent.TimeUnit;
 
 import javax.annotation.Nullable;
 import javax.security.auth.kerberos.KerberosPrincipal;
@@ -47,6 +48,7 @@ import org.apache.hadoop.net.NetUtils;
 import org.apache.hadoop.security.UserGroupInformation.AuthenticationMethod;
 import org.apache.hadoop.security.token.Token;
 import org.apache.hadoop.security.token.TokenInfo;
+import org.apache.hadoop.util.StopWatch;
 import org.apache.hadoop.util.StringUtils;
 
 
@@ -79,6 +81,9 @@ public class SecurityUtil {
     setTokenServiceUseIp(useIp);
   }
 
+  private static boolean logSlowLookups = getLogSlowLookupsEnabled();
+  private static int slowLookupThresholdMs = getSlowLookupThresholdMs();
+
   /**
    * For use only by tests and initialization
    */
@@ -480,9 +485,27 @@ public class SecurityUtil {
     }
   }
 
+  private static boolean getLogSlowLookupsEnabled() {
+    Configuration conf = new Configuration();
+
+    return conf.getBoolean(CommonConfigurationKeys
+            .HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_ENABLED_KEY,
+        CommonConfigurationKeys
+            .HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_ENABLED_DEFAULT);
+  }
+
+  private static int getSlowLookupThresholdMs() {
+    Configuration conf = new Configuration();
+
+    return conf.getInt(CommonConfigurationKeys
+            .HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_THRESHOLD_MS_KEY,
+        CommonConfigurationKeys
+            .HADOOP_SECURITY_DNS_LOG_SLOW_LOOKUPS_THRESHOLD_MS_DEFAULT);
+  }
+
   /**
    * Resolves a host subject to the security requirements determined by
-   * hadoop.security.token.service.use_ip.
+   * hadoop.security.token.service.use_ip. Optionally logs slow resolutions.
    * 
    * @param hostname host or ip to resolve
    * @return a resolved host
@@ -491,8 +514,23 @@ public class SecurityUtil {
   @InterfaceAudience.Private
   public static
   InetAddress getByName(String hostname) throws UnknownHostException {
+    if (logSlowLookups || LOG.isTraceEnabled()) {
+      StopWatch lookupTimer = new StopWatch().start();
+      InetAddress result = hostResolver.getByName(hostname);
+      long elapsedMs = lookupTimer.stop().now(TimeUnit.MILLISECONDS);
+
+      if (elapsedMs >= slowLookupThresholdMs) {
+        LOG.warn("Slow name lookup for " + hostname + ". Took " + elapsedMs +
+            " ms.");
+      } else if (LOG.isTraceEnabled()) {
+        LOG.trace("Name lookup for " + hostname + " took " + elapsedMs +
+            " ms.");
+      }
+      return result;
+    } else {
       return hostResolver.getByName(hostname);
     }
+  }
   
   interface HostResolver {
     InetAddress getByName(String host) throws UnknownHostException;    

hadoop-common-project/hadoop-common/src/main/resources/core-default.xml

@@ -113,6 +113,24 @@
   </description>
 </property>
 
+<property>
+  <name>hadoop.security.dns.log-slow-lookups.enabled</name>
+  <value>false</value>
+  <description>
+    Time name lookups (via SecurityUtil) and log them if they exceed the
+    configured threshold.
+  </description>
+</property>
+
+<property>
+  <name>hadoop.security.dns.log-slow-lookups.threshold.ms</name>
+  <value>1000</value>
+  <description>
+    If slow lookup logging is enabled, this threshold is used to decide if a
+    lookup is considered slow enough to be logged.
+  </description>
+</property>
+
 <!-- 
 === Multiple group mapping providers configuration sample === 
   This sample illustrates a typical use case for CompositeGroupsMapping where