Expose the start of the expiration range for TTL blob files through LiveFileMetaData (#7365)

Summary:
The patch adds support for exposing the start of the expiration range
for TTL blob files through the `GetLiveFilesMetaData` API. This can be
used for monitoring purposes, i.e. to make sure TTL blob files are
deleted in a timely manner. The patch also fixes a couple of uninitialized
variable issues.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/7365

Test Plan: `make check`

Reviewed By: pdillinger

Differential Revision: D23605465

Pulled By: ltamasi

fbshipit-source-id: 97a9612bf5f4b058423debdd3f28f576bb23a70f

include/rocksdb/metadata.h

@@ -62,7 +62,9 @@ struct SstFileMetaData {
         being_compacted(false),
         num_entries(0),
         num_deletions(0),
-        oldest_blob_file_number(0) {}
+        oldest_blob_file_number(0),
+        oldest_ancester_time(0),
+        file_creation_time(0) {}
 
   SstFileMetaData(const std::string& _file_name, uint64_t _file_number,
                   const std::string& _path, size_t _size,
@@ -117,6 +119,8 @@ struct SstFileMetaData {
   // oldest SST file that is the compaction ancester of this file.
   // The timestamp is provided Env::GetCurrentTime().
   // 0 if the information is not available.
+  //
+  // Note: for TTL blob files, it contains the start of the expiration range.
   uint64_t oldest_ancester_time;
   // Timestamp when the SST file is created, provided by Env::GetCurrentTime().
   // 0 if the information is not available.

utilities/blob_db/blob_db_impl_filesnapshot.cc

@@ -98,6 +98,9 @@ void BlobDBImpl::GetLiveFilesMetaData(std::vector<LiveFileMetaData>* metadata) {
     // Path should be relative to db_name, but begin with slash.
     filemetadata.name = BlobFileName("", bdb_options_.blob_dir, file_number);
     filemetadata.file_number = file_number;
+    if (blob_file->HasTTL()) {
+      filemetadata.oldest_ancester_time = blob_file->GetExpirationRange().first;
+    }
     auto cfh =
         static_cast_with_check<ColumnFamilyHandleImpl>(DefaultColumnFamily());
     filemetadata.column_family_name = cfh->GetName();

utilities/blob_db/blob_db_test.cc

@@ -956,29 +956,50 @@ TEST_F(BlobDBTest, ColumnFamilyNotSupported) {
 
 TEST_F(BlobDBTest, GetLiveFilesMetaData) {
   Random rnd(301);
+
   BlobDBOptions bdb_options;
   bdb_options.blob_dir = "blob_dir";
   bdb_options.path_relative = true;
+  bdb_options.ttl_range_secs = 10;
   bdb_options.min_blob_size = 0;
   bdb_options.disable_background_tasks = true;
-  Open(bdb_options);
+
+  Options options;
+  options.env = mock_env_.get();
+
+  Open(bdb_options, options);
+
   std::map<std::string, std::string> data;
   for (size_t i = 0; i < 100; i++) {
     PutRandom("key" + ToString(i), &rnd, &data);
   }
+
+  constexpr uint64_t expiration = 1000ULL;
+  PutRandomUntil("key100", expiration, &rnd, &data);
+
   std::vector<LiveFileMetaData> metadata;
   blob_db_->GetLiveFilesMetaData(&metadata);
-  ASSERT_EQ(1U, metadata.size());
+
+  ASSERT_EQ(2U, metadata.size());
   // Path should be relative to db_name, but begin with slash.
-  std::string filename = "/blob_dir/000001.blob";
+  const std::string filename1("/blob_dir/000001.blob");
-  ASSERT_EQ(filename, metadata[0].name);
+  ASSERT_EQ(filename1, metadata[0].name);
   ASSERT_EQ(1, metadata[0].file_number);
-  ASSERT_EQ("default", metadata[0].column_family_name);
+  ASSERT_EQ(0, metadata[0].oldest_ancester_time);
+  ASSERT_EQ(kDefaultColumnFamilyName, metadata[0].column_family_name);
+
+  const std::string filename2("/blob_dir/000002.blob");
+  ASSERT_EQ(filename2, metadata[1].name);
+  ASSERT_EQ(2, metadata[1].file_number);
+  ASSERT_EQ(expiration, metadata[1].oldest_ancester_time);
+  ASSERT_EQ(kDefaultColumnFamilyName, metadata[1].column_family_name);
+
   std::vector<std::string> livefile;
   uint64_t mfs;
   ASSERT_OK(blob_db_->GetLiveFiles(livefile, &mfs, false));
-  ASSERT_EQ(4U, livefile.size());
+  ASSERT_EQ(5U, livefile.size());
-  ASSERT_EQ(filename, livefile[3]);
+  ASSERT_EQ(filename1, livefile[3]);
+  ASSERT_EQ(filename2, livefile[4]);
   VerifyDB(data);
 }
 

utilities/blob_db/blob_file.h

@@ -189,7 +189,9 @@ class BlobFile {
 
   // All Get functions which are not atomic, will need ReadLock on the mutex
 
-  ExpirationRange GetExpirationRange() const { return expiration_range_; }
+  const ExpirationRange& GetExpirationRange() const {
+    return expiration_range_;
+  }
 
   void ExtendExpirationRange(uint64_t expiration) {
     expiration_range_.first = std::min(expiration_range_.first, expiration);