Add (Live)FileStorageInfo API (#8968)

Summary:
New classes FileStorageInfo and LiveFileStorageInfo and
'experimental' function DB::GetLiveFilesStorageInfo, which is intended
to largely replace several fragmented DB functions needed to create
checkpoints and backups.

This function is now used to create checkpoints and backups, because
it fixes many (probably not all) of the prior complexities of checkpoint
not having atomic access to DB metadata. This also ensures strong
functional test coverage of the new API. Specifically, much of the old
CheckpointImpl::CreateCustomCheckpoint has been migrated to and
updated in DBImpl::GetLiveFilesStorageInfo, with the former now
calling the latter.

Also, the class FileStorageInfo in metadata.h compatibly replaces
BackupFileInfo and serves as a new base class for SstFileMetaData.
Some old fields of SstFileMetaData are still provided (for now) but
deprecated.

Although FileStorageInfo::directory is accurate when using db_paths
and/or cf_paths, these have never been supported by Checkpoint
nor BackupEngine and still are not. This change does now detect
these cases and return NotSupported when appropriate. (More work
needed for support.)

Somehow this change broke ProgressCallbackDuringBackup, but
the progress_callback logic was dubious to begin with because it
would call the callback based on copy buffer size, not size actually
copied. Logic and test updated to track size actually copied
per-thread.

Pull Request resolved: #8968

Test Plan:
tests updated.
DB::GetLiveFilesStorageInfo mostly tested by use in CheckpointImpl.
DBTest.SnapshotFiles updated to also test GetLiveFilesStorageInfo,
including reading the data after DB close.
Added CheckpointTest.CheckpointWithDbPath (NotSupported).

Reviewed By: siying

Differential Revision: D31242045

Pulled By: pdillinger

fbshipit-source-id: b183d1ce9799e220daaefd6b3b5365d98de676c0

HISTORY.md

@@ -12,6 +12,7 @@
 ### New Features
 * Print information about blob files when using "ldb list_live_files_metadata"
 * Provided support for SingleDelete with user defined timestamp.
+* Experimental new function DB::GetLiveFilesStorageInfo offers essentially a unified version of other functions like GetLiveFiles, GetLiveFilesChecksumInfo, and GetSortedWalFiles. Checkpoints and backups could show small behavioral changes and/or improved performance as they now use this new API.
 * Add remote compaction read/write bytes statistics: `REMOTE_COMPACT_READ_BYTES`, `REMOTE_COMPACT_WRITE_BYTES`.
 * Introduce an experimental feature to dump out the blocks from block cache and insert them to the secondary cache to reduce the cache warmup time (e.g., used while migrating DB instance). More information are in `class CacheDumper` and `CacheDumpedLoader` at `rocksdb/utilities/cache_dump_load.h` Note that, this feature is subject to the potential change in the future, it is still experimental.
 * Introduced a new BlobDB configuration option `blob_garbage_collection_force_threshold`, which can be used to trigger compactions targeting the SST files which reference the oldest blob files when the ratio of garbage in those blob files meets or exceeds the specified threshold. This can reduce space amplification with skewed workloads where the affected SST files might not otherwise get picked up for compaction.
@@ -20,6 +21,7 @@
 * Made SystemClock extend the Customizable class and added a CreateFromString method.  Implementations need to be registered with the ObjectRegistry and to implement a Name() method in order to be created via this method.
 * Made SliceTransform extend the Customizable class and added a CreateFromString method.  Implementations need to be registered with the ObjectRegistry and to implement a Name() method in order to be created via this method.  The Capped and Prefixed transform classes return a short name (no length); use GetId for the fully qualified name.
 * Made FileChecksumGenFactory, SstPartitionerFactory, TablePropertiesCollectorFactory, and WalFilter extend the Customizable class and added a CreateFromString method.
+* Some fields of SstFileMetaData are deprecated for compatibility with new base class FileStorageInfo.
 * Add `file_temperature` to `IngestExternalFileArg` such that when ingesting SST files, we are able to indicate the temperature of the this batch of files.
 
 ## 6.25.0 (2021-09-20)

db/db_filesnapshot.cc

@@ -6,11 +6,11 @@
 
 #ifndef ROCKSDB_LITE
 
-#include <stdint.h>
-
 #include <algorithm>
-#include <cinttypes>
+#include <cstdint>
+#include <memory>
 #include <string>
+#include <vector>
 
 #include "db/db_impl/db_impl.h"
 #include "db/job_context.h"
@@ -21,11 +21,52 @@
 #include "port/port.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
+#include "rocksdb/metadata.h"
+#include "rocksdb/types.h"
 #include "test_util/sync_point.h"
+#include "util/file_checksum_helper.h"
 #include "util/mutexlock.h"
 
 namespace ROCKSDB_NAMESPACE {
 
+Status DBImpl::FlushForGetLiveFiles() {
+  mutex_.AssertHeld();
+
+  // flush all dirty data to disk.
+  Status status;
+  if (immutable_db_options_.atomic_flush) {
+    autovector<ColumnFamilyData*> cfds;
+    SelectColumnFamiliesForAtomicFlush(&cfds);
+    mutex_.Unlock();
+    status =
+        AtomicFlushMemTables(cfds, FlushOptions(), FlushReason::kGetLiveFiles);
+    if (status.IsColumnFamilyDropped()) {
+      status = Status::OK();
+    }
+    mutex_.Lock();
+  } else {
+    for (auto cfd : *versions_->GetColumnFamilySet()) {
+      if (cfd->IsDropped()) {
+        continue;
+      }
+      cfd->Ref();
+      mutex_.Unlock();
+      status = FlushMemTable(cfd, FlushOptions(), FlushReason::kGetLiveFiles);
+      TEST_SYNC_POINT("DBImpl::GetLiveFiles:1");
+      TEST_SYNC_POINT("DBImpl::GetLiveFiles:2");
+      mutex_.Lock();
+      cfd->UnrefAndTryDelete();
+      if (!status.ok() && !status.IsColumnFamilyDropped()) {
+        break;
+      } else if (status.IsColumnFamilyDropped()) {
+        status = Status::OK();
+      }
+    }
+  }
+  versions_->GetColumnFamilySet()->FreeDeadColumnFamilies();
+  return status;
+}
+
 Status DBImpl::GetLiveFiles(std::vector<std::string>& ret,
                             uint64_t* manifest_file_size,
                             bool flush_memtable) {
@@ -34,39 +75,7 @@ Status DBImpl::GetLiveFiles(std::vector<std::string>& ret,
   mutex_.Lock();
 
   if (flush_memtable) {
-    // flush all dirty data to disk.
-    Status status;
-    if (immutable_db_options_.atomic_flush) {
-      autovector<ColumnFamilyData*> cfds;
-      SelectColumnFamiliesForAtomicFlush(&cfds);
-      mutex_.Unlock();
-      status = AtomicFlushMemTables(cfds, FlushOptions(),
-                                    FlushReason::kGetLiveFiles);
-      if (status.IsColumnFamilyDropped()) {
-        status = Status::OK();
-      }
-      mutex_.Lock();
-    } else {
-      for (auto cfd : *versions_->GetColumnFamilySet()) {
-        if (cfd->IsDropped()) {
-          continue;
-        }
-        cfd->Ref();
-        mutex_.Unlock();
-        status = FlushMemTable(cfd, FlushOptions(), FlushReason::kGetLiveFiles);
-        TEST_SYNC_POINT("DBImpl::GetLiveFiles:1");
-        TEST_SYNC_POINT("DBImpl::GetLiveFiles:2");
-        mutex_.Lock();
-        cfd->UnrefAndTryDelete();
-        if (!status.ok() && !status.IsColumnFamilyDropped()) {
-          break;
-        } else if (status.IsColumnFamilyDropped()) {
-          status = Status::OK();
-        }
-      }
-    }
-    versions_->GetColumnFamilySet()->FreeDeadColumnFamilies();
-
+    Status status = FlushForGetLiveFiles();
     if (!status.ok()) {
       mutex_.Unlock();
       ROCKS_LOG_ERROR(immutable_db_options_.info_log, "Cannot Flush data %s\n",
@@ -157,6 +166,245 @@ Status DBImpl::GetCurrentWalFile(std::unique_ptr<LogFile>* current_log_file) {
 
   return wal_manager_.GetLiveWalFile(current_logfile_number, current_log_file);
 }
+
+Status DBImpl::GetLiveFilesStorageInfo(
+    const LiveFilesStorageInfoOptions& opts,
+    std::vector<LiveFileStorageInfo>* files) {
+  // To avoid returning partial results, only move to ouput on success
+  assert(files);
+  files->clear();
+  std::vector<LiveFileStorageInfo> results;
+
+  // NOTE: This implementation was largely migrated from Checkpoint.
+
+  Status s;
+  VectorLogPtr live_wal_files;
+  bool flush_memtable = true;
+  if (!immutable_db_options_.allow_2pc) {
+    if (opts.wal_size_for_flush == port::kMaxUint64) {
+      flush_memtable = false;
+    } else if (opts.wal_size_for_flush > 0) {
+      // If out standing log files are small, we skip the flush.
+      s = GetSortedWalFiles(live_wal_files);
+
+      if (!s.ok()) {
+        return s;
+      }
+
+      // Don't flush column families if total log size is smaller than
+      // log_size_for_flush. We copy the log files instead.
+      // We may be able to cover 2PC case too.
+      uint64_t total_wal_size = 0;
+      for (auto& wal : live_wal_files) {
+        total_wal_size += wal->SizeFileBytes();
+      }
+      if (total_wal_size < opts.wal_size_for_flush) {
+        flush_memtable = false;
+      }
+      live_wal_files.clear();
+    }
+  }
+
+  // This is a modified version of GetLiveFiles, to get access to more
+  // metadata.
+  mutex_.Lock();
+  if (flush_memtable) {
+    Status status = FlushForGetLiveFiles();
+    if (!status.ok()) {
+      mutex_.Unlock();
+      ROCKS_LOG_ERROR(immutable_db_options_.info_log, "Cannot Flush data %s\n",
+                      status.ToString().c_str());
+      return status;
+    }
+  }
+
+  // Make a set of all of the live table and blob files
+  for (auto cfd : *versions_->GetColumnFamilySet()) {
+    if (cfd->IsDropped()) {
+      continue;
+    }
+    VersionStorageInfo& vsi = *cfd->current()->storage_info();
+    auto& cf_paths = cfd->ioptions()->cf_paths;
+
+    auto GetDir = [&](size_t path_id) {
+      // Matching TableFileName() behavior
+      if (path_id >= cf_paths.size()) {
+        assert(false);
+        return cf_paths.back().path;
+      } else {
+        return cf_paths[path_id].path;
+      }
+    };
+
+    for (int level = 0; level < vsi.num_levels(); ++level) {
+      const auto& level_files = vsi.LevelFiles(level);
+      for (const auto& meta : level_files) {
+        assert(meta);
+
+        results.emplace_back();
+        LiveFileStorageInfo& info = results.back();
+
+        info.relative_filename = MakeTableFileName(meta->fd.GetNumber());
+        info.directory = GetDir(meta->fd.GetPathId());
+        info.file_number = meta->fd.GetNumber();
+        info.file_type = kTableFile;
+        info.size = meta->fd.GetFileSize();
+        if (opts.include_checksum_info) {
+          info.file_checksum_func_name = meta->file_checksum_func_name;
+          info.file_checksum = meta->file_checksum;
+          if (info.file_checksum_func_name.empty()) {
+            info.file_checksum_func_name = kUnknownFileChecksumFuncName;
+            info.file_checksum = kUnknownFileChecksum;
+          }
+        }
+        info.temperature = meta->temperature;
+      }
+    }
+    const auto& blob_files = vsi.GetBlobFiles();
+    for (const auto& pair : blob_files) {
+      const auto& meta = pair.second;
+      assert(meta);
+
+      results.emplace_back();
+      LiveFileStorageInfo& info = results.back();
+
+      info.relative_filename = BlobFileName(meta->GetBlobFileNumber());
+      info.directory = GetName();  // TODO?: support db_paths/cf_paths
+      info.file_number = meta->GetBlobFileNumber();
+      info.file_type = kBlobFile;
+      info.size = meta->GetBlobFileSize();
+      if (opts.include_checksum_info) {
+        info.file_checksum_func_name = meta->GetChecksumMethod();
+        info.file_checksum = meta->GetChecksumValue();
+        if (info.file_checksum_func_name.empty()) {
+          info.file_checksum_func_name = kUnknownFileChecksumFuncName;
+          info.file_checksum = kUnknownFileChecksum;
+        }
+      }
+      // TODO?: info.temperature
+    }
+  }
+
+  // Capture some final info before releasing mutex
+  const uint64_t manifest_number = versions_->manifest_file_number();
+  const uint64_t manifest_size = versions_->manifest_file_size();
+  const uint64_t options_number = versions_->options_file_number();
+  const uint64_t options_size = versions_->options_file_size_;
+  const uint64_t min_log_num = MinLogNumberToKeep();
+
+  mutex_.Unlock();
+
+  std::string manifest_fname = DescriptorFileName(manifest_number);
+  {  // MANIFEST
+    results.emplace_back();
+    LiveFileStorageInfo& info = results.back();
+
+    info.relative_filename = manifest_fname;
+    info.directory = GetName();
+    info.file_number = manifest_number;
+    info.file_type = kDescriptorFile;
+    info.size = manifest_size;
+    info.trim_to_size = true;
+    if (opts.include_checksum_info) {
+      info.file_checksum_func_name = kUnknownFileChecksumFuncName;
+      info.file_checksum = kUnknownFileChecksum;
+    }
+  }
+
+  {  // CURRENT
+    results.emplace_back();
+    LiveFileStorageInfo& info = results.back();
+
+    info.relative_filename = kCurrentFileName;
+    info.directory = GetName();
+    info.file_type = kCurrentFile;
+    // CURRENT could be replaced so we have to record the contents we want
+    // for it
+    info.replacement_contents = manifest_fname + "\n";
+    info.size = manifest_fname.size() + 1;
+    if (opts.include_checksum_info) {
+      info.file_checksum_func_name = kUnknownFileChecksumFuncName;
+      info.file_checksum = kUnknownFileChecksum;
+    }
+  }
+
+  // The OPTIONS file number is zero in read-write mode when OPTIONS file
+  // writing failed and the DB was configured with
+  // `fail_if_options_file_error == false`. In read-only mode the OPTIONS file
+  // number is zero when no OPTIONS file exist at all. In those cases we do not
+  // record any OPTIONS file in the live file list.
+  if (options_number != 0) {
+    results.emplace_back();
+    LiveFileStorageInfo& info = results.back();
+
+    info.relative_filename = OptionsFileName(options_number);
+    info.directory = GetName();
+    info.file_number = options_number;
+    info.file_type = kOptionsFile;
+    info.size = options_size;
+    if (opts.include_checksum_info) {
+      info.file_checksum_func_name = kUnknownFileChecksumFuncName;
+      info.file_checksum = kUnknownFileChecksum;
+    }
+  }
+
+  // Some legacy testing stuff  TODO: carefully clean up obsolete parts
+  TEST_SYNC_POINT("CheckpointImpl::CreateCheckpoint:FlushDone");
+
+  TEST_SYNC_POINT("CheckpointImpl::CreateCheckpoint:SavedLiveFiles1");
+  TEST_SYNC_POINT("CheckpointImpl::CreateCheckpoint:SavedLiveFiles2");
+
+  if (s.ok()) {
+    s = FlushWAL(false /* sync */);
+  }
+
+  TEST_SYNC_POINT("CheckpointImpl::CreateCustomCheckpoint:AfterGetLive1");
+  TEST_SYNC_POINT("CheckpointImpl::CreateCustomCheckpoint:AfterGetLive2");
+
+  // if we have more than one column family, we need to also get WAL files
+  if (s.ok()) {
+    s = GetSortedWalFiles(live_wal_files);
+  }
+  if (!s.ok()) {
+    return s;
+  }
+
+  size_t wal_size = live_wal_files.size();
+
+  ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                 "Number of log files %" ROCKSDB_PRIszt, live_wal_files.size());
+
+  // Link WAL files. Copy exact size of last one because it is the only one
+  // that has changes after the last flush.
+  auto wal_dir = immutable_db_options_.GetWalDir();
+  for (size_t i = 0; s.ok() && i < wal_size; ++i) {
+    if ((live_wal_files[i]->Type() == kAliveLogFile) &&
+        (!flush_memtable || live_wal_files[i]->LogNumber() >= min_log_num)) {
+      results.emplace_back();
+      LiveFileStorageInfo& info = results.back();
+      auto f = live_wal_files[i]->PathName();
+      assert(!f.empty() && f[0] == '/');
+      info.relative_filename = f.substr(1);
+      info.directory = wal_dir;
+      info.file_number = live_wal_files[i]->LogNumber();
+      info.file_type = kWalFile;
+      info.size = live_wal_files[i]->SizeFileBytes();
+      // Only last should need to be trimmed
+      info.trim_to_size = (i + 1 == wal_size);
+      if (opts.include_checksum_info) {
+        info.file_checksum_func_name = kUnknownFileChecksumFuncName;
+        info.file_checksum = kUnknownFileChecksum;
+      }
+    }
+  }
+
+  if (s.ok()) {
+    // Only move output on success
+    *files = std::move(results);
+  }
+  return s;
+}
+
 }  // namespace ROCKSDB_NAMESPACE
 
 #endif  // ROCKSDB_LITE

db/db_impl/db_impl.cc

@@ -4284,11 +4284,16 @@ Status DBImpl::RenameTempFileToOptionsFile(const std::string& file_name) {
   uint64_t options_file_number = versions_->NewFileNumber();
   std::string options_file_name =
       OptionsFileName(GetName(), options_file_number);
-  // Retry if the file name happen to conflict with an existing one.
-  s = GetEnv()->RenameFile(file_name, options_file_name);
+  uint64_t options_file_size = 0;
+  s = GetEnv()->GetFileSize(file_name, &options_file_size);
+  if (s.ok()) {
+    // Retry if the file name happen to conflict with an existing one.
+    s = GetEnv()->RenameFile(file_name, options_file_name);
+  }
   if (s.ok()) {
     InstrumentedMutexLock l(&mutex_);
     versions_->options_file_number_ = options_file_number;
+    versions_->options_file_size_ = options_file_size;
   }
 
   if (0 == disable_delete_obsolete_files_) {