[GitHub] [hbase] Apache9 commented on a change in pull request #3617: HBASE-26079 Use StoreFileTracker when splitting and merging

[GitBox]

Apache9 commented on a change in pull request #3617:
URL: https://github.com/apache/hbase/pull/3617#discussion_r699262490



##########
File path: 
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/storefiletracker/StoreFileTrackerFactory.java
##########
@@ -17,23 +17,61 @@
  */
 package org.apache.hadoop.hbase.regionserver.storefiletracker;
 
+import org.apache.commons.lang3.StringUtils;
 import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.hbase.client.ColumnFamilyDescriptor;
+import org.apache.hadoop.hbase.client.ColumnFamilyDescriptorBuilder;
+import org.apache.hadoop.hbase.client.TableDescriptor;
+import org.apache.hadoop.hbase.regionserver.HRegionFileSystem;
 import org.apache.hadoop.hbase.regionserver.StoreContext;
+import org.apache.hadoop.hbase.util.Bytes;
 import org.apache.hadoop.hbase.util.ReflectionUtils;
 import org.apache.yetus.audience.InterfaceAudience;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 
 /**
  * Factory method for creating store file tracker.
  */
 @InterfaceAudience.Private
 public final class StoreFileTrackerFactory {
-
+  
   public static final String TRACK_IMPL = "hbase.store.file-tracker.impl";
+  
+  private static final Logger LOG = 
LoggerFactory.getLogger(StoreFileTrackerFactory.class);
 
   public static StoreFileTracker create(Configuration conf, boolean 
isPrimaryReplica,
-    StoreContext ctx) {
-    Class<? extends StoreFileTracker> tracker =
-      conf.getClass(TRACK_IMPL, DefaultStoreFileTracker.class, 
StoreFileTracker.class);
-    return ReflectionUtils.newInstance(tracker, conf, isPrimaryReplica, ctx);
+      StoreContext ctx) {
+    String className = conf.get(TRACK_IMPL, 
DefaultStoreFileTracker.class.getName());

Review comment:
       What is the purpose of changing the getClass to get class name? With 
getClass call we could get a Class<? extends StoreFileTracker> directly withtou 
casting.

##########
File path: 
hbase-server/src/main/java/org/apache/hadoop/hbase/master/assignment/MergeTableRegionsProcedure.java
##########
@@ -584,40 +588,46 @@ private void createMergedRegion(final MasterProcedureEnv 
env) throws IOException
     final MasterFileSystem mfs = env.getMasterServices().getMasterFileSystem();
     final Path tableDir = CommonFSUtils.getTableDir(mfs.getRootDir(), 
regionsToMerge[0].getTable());
     final FileSystem fs = mfs.getFileSystem();
-
+    List<Path> mergedFiles = new ArrayList<>();
     HRegionFileSystem mergeRegionFs = 
HRegionFileSystem.createRegionOnFileSystem(
       env.getMasterConfiguration(), fs, tableDir, mergedRegion);
 
     for (RegionInfo ri: this.regionsToMerge) {
       HRegionFileSystem regionFs = HRegionFileSystem.openRegionFromFileSystem(
           env.getMasterConfiguration(), fs, tableDir, ri, false);
-      mergeStoreFiles(env, regionFs, mergeRegionFs, mergedRegion);
+      mergedFiles.addAll(mergeStoreFiles(env, regionFs, mergeRegionFs, 
mergedRegion));
     }
     assert mergeRegionFs != null;
-    mergeRegionFs.commitMergedRegion();
+    mergeRegionFs.commitMergedRegion(mergedFiles, env);
 
     // Prepare to create merged regions
     env.getAssignmentManager().getRegionStates().
         getOrCreateRegionStateNode(mergedRegion).setState(State.MERGING_NEW);
   }
 
-  private void mergeStoreFiles(MasterProcedureEnv env, HRegionFileSystem 
regionFs,
+  private List<Path> mergeStoreFiles(MasterProcedureEnv env, HRegionFileSystem 
regionFs,
     HRegionFileSystem mergeRegionFs, RegionInfo mergedRegion) throws 
IOException {
     final TableDescriptor htd = env.getMasterServices().getTableDescriptors()
       .get(mergedRegion.getTable());
+    List<Path> mergedFiles = new ArrayList<>();
     for (ColumnFamilyDescriptor hcd : htd.getColumnFamilies()) {
       String family = hcd.getNameAsString();
-      final Collection<StoreFileInfo> storeFiles = 
regionFs.getStoreFiles(family);
+      Configuration trackerConfig =
+        
StoreFileTrackerFactory.mergeConfigurations(env.getMasterConfiguration(), htd, 
hcd);

Review comment:
       Good.

##########
File path: 
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionFileSystem.java
##########
@@ -592,19 +598,46 @@ void cleanupDaughterRegion(final RegionInfo regionInfo) 
throws IOException {
    * @param regionInfo daughter {@link 
org.apache.hadoop.hbase.client.RegionInfo}
    * @throws IOException
    */
-  public Path commitDaughterRegion(final RegionInfo regionInfo)
-      throws IOException {
+  public Path commitDaughterRegion(final RegionInfo regionInfo, List<Path> 
allRegionFiles,
+      MasterProcedureEnv env) throws IOException {
     Path regionDir = this.getSplitsDir(regionInfo);
     if (fs.exists(regionDir)) {
       // Write HRI to a file in case we need to recover hbase:meta
       Path regionInfoFile = new Path(regionDir, REGION_INFO_FILE);
       byte[] regionInfoContent = getRegionInfoFileContent(regionInfo);
       writeRegionInfoFileContent(conf, fs, regionInfoFile, regionInfoContent);
+      HRegionFileSystem regionFs = HRegionFileSystem.openRegionFromFileSystem(
+        env.getMasterConfiguration(), fs, getTableDir(), regionInfo, false);
+      loadRegionFilesIntoStoreTracker(allRegionFiles, env, regionFs);
     }
-
     return regionDir;
   }
 
+  private void loadRegionFilesIntoStoreTracker(List<Path> allFiles, 
MasterProcedureEnv env,

Review comment:
       The word 'load' is a bit confusing here, as we have a load method in 
StoreFileTracker. Just call it 'insert' or something else?




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@hbase.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org