Blob DB: miscellaneous changes

Summary:
* Expose garbage collection related options
* Minor logging and counter name update
* Remove unused constants.
Closes facebook#3451

Differential Revision: D6867077

Pulled By: yiwu-arbug

fbshipit-source-id: 6c3272a9c9d78b125a0bd6b2e56d00d087cdd6c8

include/rocksdb/statistics.h

@@ -415,7 +415,7 @@ const std::vector<std::pair<Tickers, std::string>> TickersNameMap = {
     {BLOB_DB_WRITE_BLOB, "rocksdb.blobdb.write.blob"},
     {BLOB_DB_WRITE_BLOB_TTL, "rocksdb.blobdb.write.blob.ttl"},
     {BLOB_DB_BLOB_FILE_BYTES_WRITTEN, "rocksdb.blobdb.blob.file.bytes.written"},
-    {BLOB_DB_BLOB_FILE_BYTES_READ, "rocksdb.blobdb.blob.file,bytes.read"},
+    {BLOB_DB_BLOB_FILE_BYTES_READ, "rocksdb.blobdb.blob.file.bytes.read"},
     {BLOB_DB_BLOB_FILE_SYNCED, "rocksdb.blobdb.blob.file.synced"},
     {BLOB_DB_BLOB_INDEX_EXPIRED, "rocksdb.blobdb.blob.index.expired"},
     {BLOB_DB_GC_NUM_FILES, "rocksdb.blobdb.gc.num.files"},

utilities/blob_db/blob_db.h

@@ -78,6 +78,13 @@ struct BlobDBOptions {
   // blob files will be cleanup based on TTL.
   bool enable_garbage_collection = false;
 
+  // Time interval to trigger garbage collection, in seconds.
+  uint64_t garbage_collection_interval_secs = 60;
+
+  // If garbage collection is enabled, blob files with deleted size no less
+  // than this ratio will become candidates to be cleanup.
+  double garbage_collection_deletion_size_threshold = 0.75;
+
   // Disable all background job. Used for test only.
   bool disable_background_tasks = false;
 

utilities/blob_db/blob_db_impl.cc

@@ -211,7 +211,8 @@ void BlobDBImpl::StartBackgroundTasks() {
   tqueue_.add(
       kReclaimOpenFilesPeriodMillisecs,
       std::bind(&BlobDBImpl::ReclaimOpenFiles, this, std::placeholders::_1));
-  tqueue_.add(kGCCheckPeriodMillisecs,
+  tqueue_.add(static_cast<int64_t>(
+                  bdb_options_.garbage_collection_interval_secs * 1000),
               std::bind(&BlobDBImpl::RunGC, this, std::placeholders::_1));
   if (bdb_options_.enable_garbage_collection) {
     tqueue_.add(
@@ -1773,8 +1774,9 @@ Status BlobDBImpl::GCFileAndUpdateLSM(const std::shared_ptr<BlobFile>& bfptr,
   ROCKS_LOG_INFO(
       db_options_.info_log,
       "%s blob file %" PRIu64 ". Total blob records: %" PRIu64
-      ", Expired: %" PRIu64 " keys/%" PRIu64 " bytes, Overwritten: %" PRIu64
-      " keys/%" PRIu64 " bytes.",
+      ", expired: %" PRIu64 " keys/%" PRIu64
+      " bytes, updated or deleted by user: %" PRIu64 " keys/%" PRIu64
+      " bytes, rewrite to new file: %" PRIu64 " keys/%" PRIu64 " bytes.",
       s.ok() ? "Successfully garbage collected" : "Failed to garbage collect",
       bfptr->BlobFileNumber(), gc_stats->blob_count, gc_stats->num_keys_expired,
       gc_stats->bytes_expired, gc_stats->num_keys_overwritten,
@@ -1828,14 +1830,9 @@ bool BlobDBImpl::ShouldGCFile(std::shared_ptr<BlobFile> bfile, uint64_t now,
       return true;
     }
 
-    if (bdb_options_.ttl_range_secs < kPartialExpirationGCRangeSecs) {
-      *reason = "has ttl but partial expiration not turned on";
-      return false;
-    }
-
     ReadLock lockbfile_r(&bfile->mutex_);
-    bool ret = ((bfile->deleted_size_ * 100.0 / bfile->file_size_.load()) >
-                kPartialExpirationPercentage);
+    bool ret = ((bfile->deleted_size_ / bfile->file_size_.load()) >
+                bdb_options_.garbage_collection_deletion_size_threshold);
     if (ret) {
       *reason = "deleted blobs beyond threshold";
     } else {
@@ -1854,8 +1851,8 @@ bool BlobDBImpl::ShouldGCFile(std::shared_ptr<BlobFile> bfile, uint64_t now,
   ReadLock lockbfile_r(&bfile->mutex_);
 
   if (bdb_options_.enable_garbage_collection) {
-    if ((bfile->deleted_size_ * 100.0 / bfile->file_size_.load()) >
-        kPartialExpirationPercentage) {
+    if ((bfile->deleted_size_ / bfile->file_size_.load()) >
+        bdb_options_.garbage_collection_deletion_size_threshold) {
       *reason = "deleted simple blobs beyond threshold";
       return true;
     }

utilities/blob_db/blob_db_impl.h

@@ -144,20 +144,6 @@ class BlobDBImpl : public BlobDB {
   // how many random access open files can we tolerate
   static constexpr uint32_t kOpenFilesTrigger = 100;
 
-  // how many periods of stats do we keep.
-  static constexpr uint32_t kWriteAmplificationStatsPeriods = 24;
-
-  // we will garbage collect blob files in
-  // which entire files have expired. However if the
-  // ttl_range of files is very large say a day, we
-  // would have to wait for the entire day, before we
-  // recover most of the space.
-  static constexpr uint32_t kPartialExpirationGCRangeSecs = 4 * 3600;
-
-  // this should be based on allowed Write Amplification
-  // if 50% of the space of a blob file has been deleted/expired,
-  static constexpr uint32_t kPartialExpirationPercentage = 75;
-
   // how often to schedule reclaim open files.
   static constexpr uint32_t kReclaimOpenFilesPeriodMillisecs = 1 * 1000;