HBASE-27794: Tooling for parsing/reading the prefetch files list file (#5468) (#5491)

Signed-off-by: Wellington Chevreuil <[email protected]>

(cherry picked from commit fa4c896)

Author: Kota-SH

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

@@ -3317,4 +3317,9 @@ List<LogEntry> getLogEntries(Set<ServerName> serverNames, String logType, Server
    * @throws IOException if a remote or network exception occurs
    */
   Future<Void> truncateRegionAsync(byte[] regionName) throws IOException;
+
+  /**
+   * Get the list of cached files
+   */
+  List<String> getCachedFilesList(ServerName serverName) throws IOException;
 }

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

@@ -1702,4 +1702,9 @@ CompletableFuture<List<LogEntry>> getLogEntries(Set<ServerName> serverNames, Str
    * Flush master local region
    */
   CompletableFuture<Void> flushMasterStore();
+
+  /**
+   * Get the list of cached files
+   */
+  CompletableFuture<List<String>> getCachedFilesList(ServerName serverName);
 }

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

@@ -911,4 +911,9 @@ public CompletableFuture<List<LogEntry>> getLogEntries(Set<ServerName> serverNam
   public CompletableFuture<Void> flushMasterStore() {
     return wrap(rawAdmin.flushMasterStore());
   }
+
+  @Override
+  public CompletableFuture<List<String>> getCachedFilesList(ServerName serverName) {
+    return wrap(rawAdmin.getCachedFilesList(serverName));
+  }
 }

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

@@ -2078,14 +2078,23 @@ public Future<Void> truncateRegionAsync(byte[] regionName) throws IOException {
       throw new IllegalArgumentException("Invalid region: " + Bytes.toStringBinary(regionName));
     }
 
-    TableName tableName = hri.getTable();
+    TableName tableName = (hri != null) ? hri.getTable() : null;
 
     MasterProtos.TruncateRegionResponse response =
       executeCallable(getTruncateRegionCallable(tableName, hri));
 
     return new TruncateRegionFuture(this, tableName, response);
   }
 
+  /**
+   * Get the list of cached files
+   */
+  @Override
+  public List<String> getCachedFilesList(ServerName serverName) throws IOException {
+    return ProtobufUtil.getCachedFilesList(rpcControllerFactory.newController(),
+      this.connection.getAdmin(serverName));
+  }
+
   private MasterCallable<MasterProtos.TruncateRegionResponse>
     getTruncateRegionCallable(TableName tableName, RegionInfo hri) {
     return new MasterCallable<MasterProtos.TruncateRegionResponse>(getConnection(),

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

@@ -129,6 +129,8 @@
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.CompactionSwitchResponse;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.FlushRegionRequest;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.FlushRegionResponse;
+import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetCachedFilesListRequest;
+import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetCachedFilesListResponse;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetOnlineRegionRequest;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetOnlineRegionResponse;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetRegionInfoRequest;
@@ -4264,4 +4266,15 @@ Void> call(controller, stub, request.build(),
           (s, c, req, done) -> s.flushMasterStore(c, req, done), resp -> null))
       .call();
   }
+
+  @Override
+  public CompletableFuture<List<String>> getCachedFilesList(ServerName serverName) {
+    GetCachedFilesListRequest.Builder request = GetCachedFilesListRequest.newBuilder();
+    return this.<List<String>> newAdminCaller()
+      .action((controller, stub) -> this.<GetCachedFilesListRequest, GetCachedFilesListResponse,
+        List<String>> adminCall(controller, stub, request.build(),
+          (s, c, req, done) -> s.getCachedFilesList(c, req, done),
+          resp -> resp.getCachedFilesList()))
+      .serverName(serverName).call();
+  }
 }

hbase-client/src/main/java/org/apache/hadoop/hbase/shaded/protobuf/ProtobufUtil.java

@@ -149,6 +149,8 @@
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.AdminService;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.ClearSlowLogResponses;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.CloseRegionRequest;
+import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetCachedFilesListRequest;
+import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetCachedFilesListResponse;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetOnlineRegionRequest;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetOnlineRegionResponse;
 import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.GetRegionInfoRequest;
@@ -1746,6 +1748,21 @@ public static List<org.apache.hadoop.hbase.client.RegionInfo> getOnlineRegions(
     return getRegionInfos(response);
   }
 
+  /**
+   * Get the list of cached files
+   */
+  public static List<String> getCachedFilesList(final RpcController controller,
+    final AdminService.BlockingInterface admin) throws IOException {
+    GetCachedFilesListRequest request = GetCachedFilesListRequest.newBuilder().build();
+    GetCachedFilesListResponse response = null;
+    try {
+      response = admin.getCachedFilesList(controller, request);
+    } catch (ServiceException se) {
+      throw getRemoteException(se);
+    }
+    return new ArrayList<>(response.getCachedFilesList());
+  }
+
   /**
    * Get the list of region info from a GetOnlineRegionResponse
    * @param proto the GetOnlineRegionResponse

hbase-protocol-shaded/src/main/protobuf/Admin.proto

@@ -283,6 +283,13 @@ message ExecuteProceduresRequest {
 message ExecuteProceduresResponse {
 }
 
+message GetCachedFilesListRequest {
+}
+
+message GetCachedFilesListResponse {
+  repeated string cached_files = 1;
+}
+
 /**
  * Slow/Large log (LogRequest) use-case specific RPC request. This request payload will be
  * converted in bytes and sent to generic RPC API: GetLogEntries
@@ -406,4 +413,7 @@ service AdminService {
   rpc GetLogEntries(LogRequest)
     returns(LogEntry);
 
+  rpc GetCachedFilesList(GetCachedFilesListRequest)
+    returns(GetCachedFilesListResponse);
+
 }

hbase-server/src/main/java/org/apache/hadoop/hbase/io/hfile/BlockCache.java

@@ -18,6 +18,8 @@
 package org.apache.hadoop.hbase.io.hfile;
 
 import java.util.Iterator;
+import java.util.Map;
+import java.util.Optional;
 import org.apache.yetus.audience.InterfaceAudience;
 
 /**
@@ -146,4 +148,11 @@ Cacheable getBlock(BlockCacheKey cacheKey, boolean caching, boolean repeat,
   default boolean isMetaBlock(BlockType blockType) {
     return blockType != null && blockType.getCategory() != BlockType.BlockCategory.DATA;
   }
+
+  /**
+   * Returns the list of fully cached files
+   */
+  default Optional<Map<String, Boolean>> getFullyCachedFiles() {
+    return Optional.empty();
+  }
 }

hbase-server/src/main/java/org/apache/hadoop/hbase/io/hfile/CombinedBlockCache.java

@@ -18,6 +18,8 @@
 package org.apache.hadoop.hbase.io.hfile;
 
 import java.util.Iterator;
+import java.util.Map;
+import java.util.Optional;
 import org.apache.hadoop.hbase.io.HeapSize;
 import org.apache.hadoop.hbase.io.hfile.bucket.BucketCache;
 import org.apache.yetus.audience.InterfaceAudience;
@@ -369,6 +371,14 @@ public BlockCache[] getBlockCaches() {
     return new BlockCache[] { this.l1Cache, this.l2Cache };
   }
 
+  /**
+   * Returns the list of fully cached files
+   */
+  @Override
+  public Optional<Map<String, Boolean>> getFullyCachedFiles() {
+    return this.l2Cache.getFullyCachedFiles();
+  }
+
   @Override
   public void setMaxSize(long size) {
     this.l1Cache.setMaxSize(size);

hbase-server/src/main/java/org/apache/hadoop/hbase/io/hfile/HFilePreadReader.java

@@ -42,8 +42,9 @@ public HFilePreadReader(ReaderContext context, HFileInfo fileInfo, CacheConfig c
     final MutableBoolean fileAlreadyCached = new MutableBoolean(false);
     Optional<BucketCache> bucketCacheOptional =
       BucketCache.getBucketCacheFromCacheConfig(cacheConf);
-    bucketCacheOptional.ifPresent(bc -> fileAlreadyCached
-      .setValue(bc.getFullyCachedFiles().get(path.getName()) == null ? false : true));
+    bucketCacheOptional.flatMap(BucketCache::getFullyCachedFiles).ifPresent(fcf -> {
+      fileAlreadyCached.setValue(fcf.get(path.getName()) == null ? false : true);
+    });
     // Prefetch file blocks upon open if requested
     if (
       cacheConf.shouldPrefetchOnOpen() && cacheIfCompactionsOff()