From 75789048c13d925d9ee9b67ad8b8dc356e7547ff Mon Sep 17 00:00:00 2001
From: Esteban Gutierrez <esteban@apache.org>
Date: Thu, 27 Jul 2017 15:58:16 -0500
Subject: [PATCH] HBASE-18362 hbck should not report split replica parent
 region from meta as errors (Huaxiang Sun)

Signed-off-by: Esteban Gutierrez <esteban@apache.org>
---
 .../apache/hadoop/hbase/util/HBaseFsck.java   | 10 ++++++++
 .../hbase/util/TestHBaseFsckReplicas.java     | 25 +++++++++++++++++++
 2 files changed, 35 insertions(+)

diff --git a/hbase-server/src/main/java/org/apache/hadoop/hbase/util/HBaseFsck.java b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/HBaseFsck.java
index d4f8e4585e6..ff5d4827990 100644
--- a/hbase-server/src/main/java/org/apache/hadoop/hbase/util/HBaseFsck.java
+++ b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/HBaseFsck.java
@@ -2484,6 +2484,16 @@ public class HBaseFsck extends Configured implements Closeable {
           return;
         }
       }
+
+      // For Replica region, we need to do a similar check. If replica is not split successfully,
+      // error is going to be reported against primary daughter region.
+      if (hbi.getReplicaId() != HRegionInfo.DEFAULT_REPLICA_ID) {
+        LOG.info("Region " + descriptiveName + " is a split parent in META, in HDFS, "
+            + "and not deployed on any region server. This may be transient.");
+        hbi.setSkipChecks(true);
+        return;
+      }
+
       errors.reportError(ERROR_CODE.LINGERING_SPLIT_PARENT, "Region "
           + descriptiveName + " is a split parent in META, in HDFS, "
           + "and not deployed on any region server. This could be transient, "
diff --git a/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestHBaseFsckReplicas.java b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestHBaseFsckReplicas.java
index 3d0647e5e78..6e49b81f0c6 100644
--- a/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestHBaseFsckReplicas.java
+++ b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestHBaseFsckReplicas.java
@@ -110,6 +110,31 @@ public class TestHBaseFsckReplicas extends BaseTestHBaseFsck {
     EnvironmentEdgeManager.reset();
   }
 
+  /*
+ * This creates a table with region_replica > 1, do a split, check
+ * that hbck will not report split replica parent as lingering split parent
+ */
+  @Test public void testHbckReportReplicaLingeringSplitParent() throws Exception {
+    TableName table = TableName.valueOf("testHbckReportReplicaLingeringSplitParent");
+
+    try {
+      setupTableWithRegionReplica(table, 2);
+      admin.flush(table);
+
+      // disable catalog janitor
+      admin.enableCatalogJanitor(false);
+      admin.split(table, Bytes.toBytes("A1"));
+
+      Thread.sleep(1000);
+      // run hbck again to make sure we don't see any errors
+      assertNoErrors(doFsck(conf, false));
+    } finally {
+      cleanupTable(table);
+      // enable catalog janitor
+      admin.enableCatalogJanitor(true);
+    }
+  }
+
   /*
  * This creates a table with region_replica > 1 and verifies hbck runs
  * successfully