HBASE-8465 Auto-drop rollback snapshot for snapshot restore

git-svn-id: https://svn.apache.org/repos/asf/hbase/trunk@1545360 13f79535-47bb-0310-9956-ffa450edef68

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

@@ -51,6 +51,7 @@ import org.apache.hadoop.hbase.NotServingRegionException;
 import org.apache.hadoop.hbase.RegionException;
 import org.apache.hadoop.hbase.ServerName;
 import org.apache.hadoop.hbase.TableExistsException;
+import org.apache.hadoop.hbase.TableNotDisabledException;
 import org.apache.hadoop.hbase.TableNotEnabledException;
 import org.apache.hadoop.hbase.TableNotFoundException;
 import org.apache.hadoop.hbase.UnknownRegionException;
@@ -2773,8 +2774,11 @@ public class HBaseAdmin implements Abortable, Closeable {
 
   /**
    * Restore the specified snapshot on the original table. (The table must be disabled)
-   * Before restoring the table, a new snapshot with the current table state is created.
-   * In case of failure, the table will be rolled back to its original state.
+   * If the "hbase.snapshot.restore.take.failsafe.snapshot" configuration property
+   * is set to true, a snapshot of the current table is taken
+   * before executing the restore operation.
+   * In case of restore failure, the failsafe snapshot will be restored.
+   * If the restore completes without problem the failsafe snapshot is deleted.
    *
    * @param snapshotName name of the snapshot to restore
    * @throws IOException if a remote or network exception occurs
@@ -2788,8 +2792,11 @@ public class HBaseAdmin implements Abortable, Closeable {
 
   /**
    * Restore the specified snapshot on the original table. (The table must be disabled)
-   * Before restoring the table, a new snapshot with the current table state is created.
-   * In case of failure, the table will be rolled back to the its original state.
+   * If the "hbase.snapshot.restore.take.failsafe.snapshot" configuration property
+   * is set to true, a snapshot of the current table is taken
+   * before executing the restore operation.
+   * In case of restore failure, the failsafe snapshot will be restored.
+   * If the restore completes without problem the failsafe snapshot is deleted.
    *
    * @param snapshotName name of the snapshot to restore
    * @throws IOException if a remote or network exception occurs
@@ -2798,8 +2805,50 @@ public class HBaseAdmin implements Abortable, Closeable {
    */
   public void restoreSnapshot(final String snapshotName)
       throws IOException, RestoreSnapshotException {
-    String rollbackSnapshot = snapshotName + "-" + EnvironmentEdgeManager.currentTimeMillis();
+    boolean takeFailSafeSnapshot =
+      conf.getBoolean("hbase.snapshot.restore.take.failsafe.snapshot", false);
+    restoreSnapshot(snapshotName, takeFailSafeSnapshot);
+  }
 
+  /**
+   * Restore the specified snapshot on the original table. (The table must be disabled)
+   * If 'takeFailSafeSnapshot' is set to true, a snapshot of the current table is taken
+   * before executing the restore operation.
+   * In case of restore failure, the failsafe snapshot will be restored.
+   * If the restore completes without problem the failsafe snapshot is deleted.
+   *
+   * The failsafe snapshot name is configurable by using the property
+   * "hbase.snapshot.restore.failsafe.name".
+   *
+   * @param snapshotName name of the snapshot to restore
+   * @param takeFailSafeSnapshot true if the failsafe snapshot should be taken
+   * @throws IOException if a remote or network exception occurs
+   * @throws RestoreSnapshotException if snapshot failed to be restored
+   * @throws IllegalArgumentException if the restore request is formatted incorrectly
+   */
+  public void restoreSnapshot(final byte[] snapshotName, final boolean takeFailSafeSnapshot)
+      throws IOException, RestoreSnapshotException {
+    restoreSnapshot(Bytes.toString(snapshotName), takeFailSafeSnapshot);
+  }
+
+  /**
+   * Restore the specified snapshot on the original table. (The table must be disabled)
+   * If 'takeFailSafeSnapshot' is set to true, a snapshot of the current table is taken
+   * before executing the restore operation.
+   * In case of restore failure, the failsafe snapshot will be restored.
+   * If the restore completes without problem the failsafe snapshot is deleted.
+   *
+   * The failsafe snapshot name is configurable by using the property
+   * "hbase.snapshot.restore.failsafe.name".
+   *
+   * @param snapshotName name of the snapshot to restore
+   * @param takeFailSafeSnapshot true if the failsafe snapshot should be taken
+   * @throws IOException if a remote or network exception occurs
+   * @throws RestoreSnapshotException if snapshot failed to be restored
+   * @throws IllegalArgumentException if the restore request is formatted incorrectly
+   */
+  public void restoreSnapshot(final String snapshotName, boolean takeFailSafeSnapshot)
+      throws IOException, RestoreSnapshotException {
     TableName tableName = null;
     for (SnapshotDescription snapshotInfo: listSnapshots()) {
       if (snapshotInfo.getName().equals(snapshotName)) {
@@ -2813,24 +2862,65 @@ public class HBaseAdmin implements Abortable, Closeable {
         "Unable to find the table name for snapshot=" + snapshotName);
     }
 
-    // Take a snapshot of the current state
-    snapshot(rollbackSnapshot, tableName);
-
-    // Restore snapshot
+    // The table does not exists, switch to clone.
+    if (!tableExists(tableName)) {
       try {
+        cloneSnapshot(snapshotName, tableName);
+      } catch (InterruptedException e) {
+        throw new InterruptedIOException("Interrupted when restoring a nonexistent table: " +
+          e.getMessage());
+      }
+      return;
+    }
+
+    // Check if the table is disabled
+    if (!isTableDisabled(tableName)) {
+      throw new TableNotDisabledException(tableName);
+    }
+
+    // Take a snapshot of the current state
+    String failSafeSnapshotSnapshotName = null;
+    if (takeFailSafeSnapshot) {
+      failSafeSnapshotSnapshotName = conf.get("hbase.snapshot.restore.failsafe.name",
+        "hbase-failsafe-{snapshot.name}-{restore.timestamp}");
+      failSafeSnapshotSnapshotName = failSafeSnapshotSnapshotName
+        .replace("{snapshot.name}", snapshotName)
+        .replace("{table.name}", tableName.toString().replace(TableName.NAMESPACE_DELIM, '.'))
+        .replace("{restore.timestamp}", String.valueOf(EnvironmentEdgeManager.currentTimeMillis()));
+      LOG.info("Taking restore-failsafe snapshot: " + failSafeSnapshotSnapshotName);
+      snapshot(failSafeSnapshotSnapshotName, tableName);
+    }
+
+    try {
+      // Restore snapshot
       internalRestoreSnapshot(snapshotName, tableName);
     } catch (IOException e) {
-      // Try to rollback
+      // Somthing went wrong during the restore...
+      // if the pre-restore snapshot is available try to rollback
+      if (takeFailSafeSnapshot) {
         try {
+          internalRestoreSnapshot(failSafeSnapshotSnapshotName, tableName);
           String msg = "Restore snapshot=" + snapshotName +
-          " failed. Rollback to snapshot=" + rollbackSnapshot + " succeeded.";
+            " failed. Rollback to snapshot=" + failSafeSnapshotSnapshotName + " succeeded.";
           LOG.error(msg, e);
-        internalRestoreSnapshot(rollbackSnapshot, tableName);
           throw new RestoreSnapshotException(msg, e);
         } catch (IOException ex) {
-        String msg = "Failed to restore and rollback to snapshot=" + rollbackSnapshot;
+          String msg = "Failed to restore and rollback to snapshot=" + failSafeSnapshotSnapshotName;
           LOG.error(msg, ex);
-        throw new RestoreSnapshotException(msg, ex);
+          throw new RestoreSnapshotException(msg, e);
+        }
+      } else {
+        throw new RestoreSnapshotException("Failed to restore snapshot=" + snapshotName, e);
+      }
+    }
+
+    // If the restore is succeeded, delete the pre-restore snapshot
+    if (takeFailSafeSnapshot) {
+      try {
+        LOG.info("Deleting restore-failsafe snapshot: " + failSafeSnapshotSnapshotName);
+        deleteSnapshot(failSafeSnapshotSnapshotName);
+      } catch (IOException e) {
+        LOG.error("Unable to remove the failsafe snapshot: " + failSafeSnapshotSnapshotName, e);
       }
     }
   }

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

@@ -953,6 +953,20 @@ possible configurations would overwhelm and obscure the important.
     <value>true</value>
     <description>Set to true to allow snapshots to be taken / restored / cloned.</description>
   </property>
+  <property>
+    <name>hbase.snapshot.restore.take.failsafe.snapshot</name>
+    <value>true</value>
+    <description>Set to true to take a snapshot before the restore operation.
+      The snapshot taken will be used in case of failure, to restore the previous state.
+      At the end of the restore operation this snapshot will be deleted</description>
+  </property>
+  <property>
+    <name>hbase.snapshot.restore.failsafe.name</name>
+    <value>hbase-failsafe-{snapshot.name}-{restore.timestamp}</value>
+    <description>Name of the failsafe snapshot taken by the restore operation.
+      You can use the {snapshot.name}, {table.name} and {restore.timestamp} variables
+      to create a name based on what you are restoring.</description>
+  </property>
   <property>
     <name>hbase.server.compactchecker.interval.multiplier</name>
     <value>1000</value>

hbase-server/src/test/java/org/apache/hadoop/hbase/client/TestRestoreSnapshotFromClient.java

@@ -163,6 +163,11 @@ public class TestRestoreSnapshotFromClient {
     admin.restoreSnapshot(snapshotName1);
     admin.enableTable(tableName);
     SnapshotTestingUtils.verifyRowCount(TEST_UTIL, tableName, snapshot1Rows);
+
+    // Restore from snapshot-1
+    TEST_UTIL.deleteTable(tableName);
+    admin.restoreSnapshot(snapshotName1);
+    SnapshotTestingUtils.verifyRowCount(TEST_UTIL, tableName, snapshot1Rows);
   }
 
   @Test