HBASE-27125 The batch size of cleaning expired mob files should have an upper bound

Author: sunhelly

hbase-server/src/main/java/org/apache/hadoop/hbase/mob/MobConstants.java

@@ -46,6 +46,9 @@ public final class MobConstants {
   public static final String MOB_CACHE_EVICT_REMAIN_RATIO = "hbase.mob.cache.evict.remain.ratio";
   public static final Tag MOB_REF_TAG =
     new ArrayBackedTag(TagType.MOB_REFERENCE_TAG_TYPE, HConstants.EMPTY_BYTE_ARRAY);
+  public static final String MOB_CLEANER_BATCH_SIZE_UPPER_BOUND =
+    "hbase.master.mob.cleaner.batch.size.upper.bound";
+  public static final int DEFAULT_MOB_CLEANER_BATCH_SIZE_UPPER_BOUND = 10000;
 
   public static final float DEFAULT_EVICT_REMAIN_RATIO = 0.5f;
   public static final long DEFAULT_MOB_CACHE_EVICT_PERIOD = 3600L;

hbase-server/src/main/java/org/apache/hadoop/hbase/mob/MobUtils.java

@@ -17,6 +17,9 @@
  */
 package org.apache.hadoop.hbase.mob;
 
+import static org.apache.hadoop.hbase.mob.MobConstants.DEFAULT_MOB_CLEANER_BATCH_SIZE_UPPER_BOUND;
+import static org.apache.hadoop.hbase.mob.MobConstants.MOB_CLEANER_BATCH_SIZE_UPPER_BOUND;
+
 import java.io.FileNotFoundException;
 import java.io.IOException;
 import java.nio.ByteBuffer;
@@ -314,20 +317,24 @@ public static void cleanExpiredMobFiles(FileSystem fs, Configuration conf, Table
           }
           filesToClean
             .add(new HStoreFile(fs, file.getPath(), conf, cacheConfig, BloomType.NONE, true));
+          if (
+            filesToClean.size() > conf.getInt(MOB_CLEANER_BATCH_SIZE_UPPER_BOUND,
+              DEFAULT_MOB_CLEANER_BATCH_SIZE_UPPER_BOUND)
+          ) {
+            try {
+              removeMobFiles(conf, fs, tableName, mobTableDir, columnDescriptor.getName(),
+                filesToClean);
+              LOG.info("Table {} {} expired mob files are deleted", tableName, filesToClean.size());
+            } catch (IOException e) {
+              LOG.error("Failed to delete the mob files, table {}", tableName, e);
+            }
+            filesToClean.clear();
+          }
         }
       } catch (Exception e) {
         LOG.error("Cannot parse the fileName " + fileName, e);
       }
     }
-    if (!filesToClean.isEmpty()) {
-      try {
-        removeMobFiles(conf, fs, tableName, mobTableDir, columnDescriptor.getName(), filesToClean);
-        deletedFileCount = filesToClean.size();
-      } catch (IOException e) {
-        LOG.error("Failed to delete the mob files " + filesToClean, e);
-      }
-    }
-    LOG.info("{} expired mob files are deleted", deletedFileCount);
   }
 
   /**

hbase-server/src/test/java/org/apache/hadoop/hbase/mob/TestExpiredMobFileCleaner.java

@@ -17,6 +17,7 @@
  */
 package org.apache.hadoop.hbase.mob;
 
+import static org.apache.hadoop.hbase.mob.MobConstants.MOB_CLEANER_BATCH_SIZE_UPPER_BOUND;
 import static org.junit.Assert.assertEquals;
 
 import org.apache.hadoop.fs.FileStatus;
@@ -55,6 +56,7 @@ public class TestExpiredMobFileCleaner {
   private final static String family = "family";
   private final static byte[] row1 = Bytes.toBytes("row1");
   private final static byte[] row2 = Bytes.toBytes("row2");
+  private final static byte[] row3 = Bytes.toBytes("row3");
   private final static byte[] qf = Bytes.toBytes("qf");
 
   private static BufferedMutator table;
@@ -63,6 +65,7 @@ public class TestExpiredMobFileCleaner {
   @BeforeClass
   public static void setUpBeforeClass() throws Exception {
     TEST_UTIL.getConfiguration().setInt("hfile.format.version", 3);
+    TEST_UTIL.getConfiguration().setInt(MOB_CLEANER_BATCH_SIZE_UPPER_BOUND, 1);
   }
 
   @AfterClass
@@ -146,6 +149,12 @@ public void testCleaner() throws Exception {
     String f2 = secondFiles[1].getPath().getName();
     String secondFile = f1.equals(firstFile) ? f2 : f1;
 
+    ts = EnvironmentEdgeManager.currentTime() - 4 * secondsOfDay() * 1000; // 4 days before
+    putKVAndFlush(table, row3, dummyData, ts);
+    FileStatus[] thirdFiles = TEST_UTIL.getTestFileSystem().listStatus(mobDirPath);
+    // now there are 3 mob files
+    assertEquals("Before cleanup without delay 3", 3, thirdFiles.length);
+
     modifyColumnExpiryDays(2); // ttl = 2, make the first row expired
 
     // run the cleaner