Real fix for race in backup custom checksum checking (#7309)

Summary:
This is a "real" fix for the issue worked around in https://github.com/facebook/rocksdb/issues/7294.
To get DB checksum info for live files, we now read the manifest file
that will become part of the checkpoint/backup. This requires a little
extra handling in taking a custom checkpoint, including only reading the
manifest file up to the size prescribed by the checkpoint.

This moves GetFileChecksumsFromManifest from backup code to
file_checksum_helper.{h,cc} and removes apparently unnecessary checking
related to column families.

Updated HISTORY.md and warned potential future users of
DB::GetLiveFilesChecksumInfo()

Pull Request resolved: https://github.com/facebook/rocksdb/pull/7309

Test Plan: updated unit test, before and after

Reviewed By: ajkr

Differential Revision: D23311994

Pulled By: pdillinger

fbshipit-source-id: 741e30a2dc1830e8208f7648fcc8c5f000d4e2d5

HISTORY.md

@@ -5,7 +5,7 @@
 * Fix a possible corruption to the LSM state (overlapping files within a level) when a `CompactRange()` for refitting levels (`CompactRangeOptions::change_level == true`) and another manual compaction are executed in parallel.
 * Sanitize `recycle_log_file_num` to zero when the user attempts to enable it in combination with `WALRecoveryMode::kTolerateCorruptedTailRecords`. Previously the two features were allowed together, which compromised the user's configured crash-recovery guarantees.
 * Fix a bug where a level refitting in CompactRange() might race with an automatic compaction that puts the data to the target level of the refitting. The bug has been there for years.
-* BackupEngine::CreateNewBackup could fail intermittently with non-OK status when backing up a read-write DB configured with a DBOptions::file_checksum_gen_factory. This issue has been worked-around such that CreateNewBackup should succeed, but (until fully fixed) BackupEngine might not see all checksums available in the DB.
+* Fixed a bug in version 6.12 in which BackupEngine::CreateNewBackup could fail intermittently with non-OK status when backing up a read-write DB configured with a DBOptions::file_checksum_gen_factory.
 * Fix a bug where immutable flushed memtable is never destroyed because  a memtable is not added to delete list because of refernce hold by super version and super version doesn't switch because of empty delete list. So memory usage increases beyond write_buffer_size + max_write_buffer_size_to_maintain.
 * Fix useless no-op compactions scheduled upon snapshot release when options.disable-auto-compactions = true.
 

db/log_reader.cc

@@ -202,6 +202,10 @@ uint64_t Reader::LastRecordOffset() {
   return last_record_offset_;
 }
 
+uint64_t Reader::LastRecordEnd() {
+  return end_of_buffer_offset_ - buffer_.size();
+}
+
 void Reader::UnmarkEOF() {
   if (read_error_) {
     return;

db/log_reader.h

@@ -71,6 +71,11 @@ class Reader {
   // Undefined before the first call to ReadRecord.
   uint64_t LastRecordOffset();
 
+  // Returns the first physical offset after the last record returned by
+  // ReadRecord, or zero before first call to ReadRecord. This can also be
+  // thought of as the "current" position in processing the file bytes.
+  uint64_t LastRecordEnd();
+
   // returns true if the reader has encountered an eof condition.
   bool IsEOF() {
     return eof_;

include/rocksdb/db.h

@@ -1354,7 +1354,9 @@ class DB {
   virtual void GetLiveFilesMetaData(
       std::vector<LiveFileMetaData>* /*metadata*/) {}
 
-  // Return a list of all table checksum info
+  // Return a list of all table file checksum info.
+  // Note: This function might be of limited use because it cannot be
+  // synchronized with GetLiveFiles.
   virtual Status GetLiveFilesChecksumInfo(FileChecksumList* checksum_list) = 0;
 
   // Obtains the meta data of the specified column family of the DB.

util/file_checksum_helper.cc

@@ -9,6 +9,12 @@
 
 #include "util/file_checksum_helper.h"
 
+#include <unordered_set>
+
+#include "db/log_reader.h"
+#include "db/version_edit.h"
+#include "file/sequence_file_reader.h"
+
 namespace ROCKSDB_NAMESPACE {
 
 void FileChecksumListImpl::reset() { checksum_map_.clear(); }
@@ -83,4 +89,66 @@ std::shared_ptr<FileChecksumGenFactory> GetFileChecksumGenCrc32cFactory() {
   return default_crc32c_gen_factory;
 }
 
+Status GetFileChecksumsFromManifest(Env* src_env, const std::string& abs_path,
+                                    uint64_t manifest_file_size,
+                                    FileChecksumList* checksum_list) {
+  if (checksum_list == nullptr) {
+    return Status::InvalidArgument("checksum_list is nullptr");
+  }
+
+  checksum_list->reset();
+  Status s;
+
+  std::unique_ptr<SequentialFileReader> file_reader;
+  {
+    std::unique_ptr<FSSequentialFile> file;
+    const std::shared_ptr<FileSystem>& fs = src_env->GetFileSystem();
+    s = fs->NewSequentialFile(abs_path,
+                              fs->OptimizeForManifestRead(FileOptions()), &file,
+                              nullptr /* dbg */);
+    if (!s.ok()) {
+      return s;
+    }
+    file_reader.reset(new SequentialFileReader(std::move(file), abs_path));
+  }
+
+  struct LogReporter : public log::Reader::Reporter {
+    Status* status_ptr;
+    virtual void Corruption(size_t /*bytes*/, const Status& st) override {
+      if (status_ptr->ok()) {
+        *status_ptr = st;
+      }
+    }
+  } reporter;
+  reporter.status_ptr = &s;
+  log::Reader reader(nullptr, std::move(file_reader), &reporter,
+                     true /* checksum */, 0 /* log_number */);
+  Slice record;
+  std::string scratch;
+  while (reader.LastRecordEnd() < manifest_file_size &&
+         reader.ReadRecord(&record, &scratch) && s.ok()) {
+    VersionEdit edit;
+    s = edit.DecodeFrom(record);
+    if (!s.ok()) {
+      break;
+    }
+
+    // Remove the deleted files from the checksum_list
+    for (const auto& deleted_file : edit.GetDeletedFiles()) {
+      checksum_list->RemoveOneFileChecksum(deleted_file.second);
+    }
+
+    // Add the new files to the checksum_list
+    for (const auto& new_file : edit.GetNewFiles()) {
+      checksum_list->InsertOneFileChecksum(
+          new_file.second.fd.GetNumber(), new_file.second.file_checksum,
+          new_file.second.file_checksum_func_name);
+    }
+  }
+  assert(!s.ok() ||
+         manifest_file_size == std::numeric_limits<uint64_t>::max() ||
+         reader.LastRecordEnd() == manifest_file_size);
+  return s;
+}
+
 }  // namespace ROCKSDB_NAMESPACE

util/file_checksum_helper.h

@@ -89,4 +89,10 @@ class FileChecksumListImpl : public FileChecksumList {
       checksum_map_;
 };
 
+// If manifest_file_size < std::numeric_limits<uint64_t>::max(), only use
+// that length prefix of the manifest file.
+Status GetFileChecksumsFromManifest(Env* src_env, const std::string& abs_path,
+                                    uint64_t manifest_file_size,
+                                    FileChecksumList* checksum_list);
+
 }  // namespace ROCKSDB_NAMESPACE

utilities/backupable/backupable_db.cc

@@ -42,6 +42,7 @@
 #include "util/channel.h"
 #include "util/coding.h"
 #include "util/crc32c.h"
+#include "util/file_checksum_helper.h"
 #include "util/string_util.h"
 #include "utilities/checkpoint/checkpoint_impl.h"
 
@@ -485,9 +486,6 @@ class BackupEngineImpl : public BackupEngine {
                                               const BackupMeta* backup,
                                               FileChecksumList* checksum_list);
 
-  Status GetFileChecksumsFromManifest(Env* src_env, const std::string& abs_path,
-                                      FileChecksumList* checksum_list);
-
   Status VerifyFileWithCrc32c(Env* src_env, const BackupMeta* backup,
                               const std::string& rel_path);
 
@@ -2159,6 +2157,7 @@ Status BackupEngineImpl::GetFileDbIdentities(Env* src_env,
     return s;
   }
 }
+
 Status BackupEngineImpl::GetFileChecksumsFromManifestInBackup(
     Env* src_env, const BackupID& backup_id, const BackupMeta* backup,
     FileChecksumList* checksum_list) {
@@ -2197,87 +2196,11 @@ Status BackupEngineImpl::GetFileChecksumsFromManifestInBackup(
     return s;
   }
 
-  s = GetFileChecksumsFromManifest(src_env, GetAbsolutePath(manifest_rel_path),
-                                   checksum_list);
-  return s;
-}
-
-Status BackupEngineImpl::GetFileChecksumsFromManifest(
-    Env* src_env, const std::string& abs_path,
-    FileChecksumList* checksum_list) {
-  if (checksum_list == nullptr) {
-    return Status::InvalidArgument("checksum_list is nullptr");
-  }
-
-  checksum_list->reset();
-  Status s;
-
-  std::unique_ptr<SequentialFileReader> file_reader;
-  {
-    std::unique_ptr<FSSequentialFile> file;
-    const std::shared_ptr<FileSystem>& fs = src_env->GetFileSystem();
-    s = fs->NewSequentialFile(abs_path,
-                              fs->OptimizeForManifestRead(FileOptions()), &file,
-                              nullptr /* dbg */);
-    if (!s.ok()) {
-      return s;
-    }
-    file_reader.reset(new SequentialFileReader(std::move(file), abs_path));
-  }
-
-  LogReporter reporter;
-  reporter.status = &s;
-  log::Reader reader(nullptr, std::move(file_reader), &reporter,
-                     true /* checksum */, 0 /* log_number */);
-  Slice record;
-  std::string scratch;
-  // Set of column families initialized with default CF
-  std::unordered_set<uint32_t> cf_set = {0};
-  while (reader.ReadRecord(&record, &scratch) && s.ok()) {
-    VersionEdit edit;
-    s = edit.DecodeFrom(record);
-    if (!s.ok()) {
-      break;
-    }
-    // Check current CF status
-    uint32_t column_family = edit.GetColumnFamily();
-    auto cf_set_itr = cf_set.find(column_family);
-    bool cf_exist = (cf_set_itr != cf_set.end());
-    if (edit.IsColumnFamilyAdd()) {
-      if (cf_exist) {
-        s = Status::Corruption("Manifest adding the same column family twice");
-        break;
-      }
-      cf_set.insert(column_family);
-    } else if (edit.IsColumnFamilyDrop()) {
-      if (!cf_exist) {
-        s = Status::Corruption(
-            "Manifest dropping non-existing column family: " +
-            ToString(column_family));
-        break;
-      }
-      cf_set.erase(cf_set_itr);
-    } else {
-      if (!cf_exist) {
-        s = Status::Corruption("Manifest referencing unknown column family: " +
-                               ToString(column_family));
-        break;
-      }
-      assert(cf_set.find(column_family) != cf_set.end());
-
-      // Remove the deleted files from the checksum_list
-      for (const auto& deleted_file : edit.GetDeletedFiles()) {
-        checksum_list->RemoveOneFileChecksum(deleted_file.second);
-      }
-
-      // Add the new files to the checksum_list
-      for (const auto& new_file : edit.GetNewFiles()) {
-        checksum_list->InsertOneFileChecksum(
-            new_file.second.fd.GetNumber(), new_file.second.file_checksum,
-            new_file.second.file_checksum_func_name);
-      }
-    }
-  }
+  // Read whole manifest file in backup
+  s = GetFileChecksumsFromManifest(
+      src_env, GetAbsolutePath(manifest_rel_path),
+      std::numeric_limits<uint64_t>::max() /*manifest_file_size*/,
+      checksum_list);
   return s;
 }
 

utilities/backupable/backupable_db_test.cc

@@ -13,6 +13,7 @@
 
 #include <algorithm>
 #include <limits>
+#include <regex>
 #include <string>
 #include <utility>
 
@@ -1823,13 +1824,13 @@ TEST_F(BackupableDBTest, FlushCompactDuringBackupCheckpoint) {
     // That FillDB leaves a mix of flushed and unflushed data
     SyncPoint::GetInstance()->LoadDependency(
         {{"CheckpointImpl::CreateCustomCheckpoint:AfterGetLive1",
-          "BackupableDBTest::FlushDuringBackupCheckpoint:BeforeFlush"},
-         {"BackupableDBTest::FlushDuringBackupCheckpoint:AfterFlush",
+          "BackupableDBTest::FlushCompactDuringBackupCheckpoint:Before"},
+         {"BackupableDBTest::FlushCompactDuringBackupCheckpoint:After",
           "CheckpointImpl::CreateCustomCheckpoint:AfterGetLive2"}});
     SyncPoint::GetInstance()->EnableProcessing();
     ROCKSDB_NAMESPACE::port::Thread flush_thread{[this]() {
       TEST_SYNC_POINT(
-          "BackupableDBTest::FlushDuringBackupCheckpoint:BeforeFlush");
+          "BackupableDBTest::FlushCompactDuringBackupCheckpoint:Before");
       FillDB(db_.get(), keys_iteration, 2 * keys_iteration);
       ASSERT_OK(db_->Flush(FlushOptions()));
       DBImpl* dbi = static_cast<DBImpl*>(db_.get());
@@ -1837,11 +1838,28 @@ TEST_F(BackupableDBTest, FlushCompactDuringBackupCheckpoint) {
       ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
       dbi->TEST_WaitForCompact();
       TEST_SYNC_POINT(
-          "BackupableDBTest::FlushDuringBackupCheckpoint:AfterFlush");
+          "BackupableDBTest::FlushCompactDuringBackupCheckpoint:After");
     }};
     ASSERT_OK(backup_engine_->CreateNewBackup(db_.get()));
     flush_thread.join();
     CloseDBAndBackupEngine();
+    SyncPoint::GetInstance()->DisableProcessing();
+    SyncPoint::GetInstance()->ClearAllCallBacks();
+    if (sopt == kShareWithChecksum) {
+      // Ensure we actually got DB manifest checksums by inspecting
+      // shared_checksum file names for hex checksum component
+      std::regex expected("[^_]+_[0-9A-F]{8}_[^_]+.sst");
+      std::vector<FileAttributes> children;
+      const std::string dir = backupdir_ + "/shared_checksum";
+      ASSERT_OK(file_manager_->GetChildrenFileAttributes(dir, &children));
+      for (const auto& child : children) {
+        if (child.name == "." || child.name == ".." || child.size_bytes == 0) {
+          continue;
+        }
+        const std::string match("match");
+        EXPECT_EQ(match, std::regex_replace(child.name, expected, match));
+      }
+    }
     AssertBackupConsistency(0, 0, keys_iteration);
   }
 }

utilities/checkpoint/checkpoint_impl.cc

@@ -27,6 +27,7 @@
 #include "rocksdb/utilities/checkpoint.h"
 #include "test_util/sync_point.h"
 #include "util/cast_util.h"
+#include "util/file_checksum_helper.h"
 
 namespace ROCKSDB_NAMESPACE {
 
@@ -180,7 +181,7 @@ Status CheckpointImpl::CreateCustomCheckpoint(
                          FileType type)>
         create_file_cb,
     uint64_t* sequence_number, uint64_t log_size_for_flush,
-    const bool& get_live_table_checksum) {
+    bool get_live_table_checksum) {
   Status s;
   std::vector<std::string> live_files;
   uint64_t manifest_file_size = 0;
@@ -262,19 +263,17 @@ Status CheckpointImpl::CreateCustomCheckpoint(
 
   size_t wal_size = live_wal_files.size();
 
-  // get table file checksums if get_live_table_checksum is true
-  std::unique_ptr<FileChecksumList> checksum_list(NewFileChecksumList());
-  Status checksum_status;
-  if (get_live_table_checksum) {
-    checksum_status = db_->GetLiveFilesChecksumInfo(checksum_list.get());
-  }
-
-  // copy/hard link live_files
+  // process live files, non-table files first
   std::string manifest_fname, current_fname;
-  for (size_t i = 0; s.ok() && i < live_files.size(); ++i) {
+  // record table files for processing next
+  std::vector<std::pair<std::string, uint64_t>> live_table_files;
+  for (auto& live_file : live_files) {
+    if (!s.ok()) {
+      break;
+    }
     uint64_t number;
     FileType type;
-    bool ok = ParseFileName(live_files[i], &number, &type);
+    bool ok = ParseFileName(live_file, &number, &type);
     if (!ok) {
       s = Status::Corruption("Can't parse file name. This is very bad");
       break;
@@ -282,54 +281,80 @@ Status CheckpointImpl::CreateCustomCheckpoint(
     // we should only get sst, options, manifest and current files here
     assert(type == kTableFile || type == kDescriptorFile ||
            type == kCurrentFile || type == kOptionsFile);
-    assert(live_files[i].size() > 0 && live_files[i][0] == '/');
+    assert(live_file.size() > 0 && live_file[0] == '/');
     if (type == kCurrentFile) {
       // We will craft the current file manually to ensure it's consistent with
       // the manifest number. This is necessary because current's file contents
       // can change during checkpoint creation.
-      current_fname = live_files[i];
+      current_fname = live_file;
       continue;
     } else if (type == kDescriptorFile) {
-      manifest_fname = live_files[i];
+      manifest_fname = live_file;
+    }
+    if (type != kTableFile) {
+      // copy non-table files here
+      // * if it's kDescriptorFile, limit the size to manifest_file_size
+      s = copy_file_cb(db_->GetName(), live_file,
+                       (type == kDescriptorFile) ? manifest_file_size : 0, type,
+                       kUnknownFileChecksumFuncName, kUnknownFileChecksum);
+    } else {
+      // process table files below
+      live_table_files.push_back(make_pair(live_file, number));
+    }
+  }
+
+  // get checksum info for table files
+  // get table file checksums if get_live_table_checksum is true
+  std::unique_ptr<FileChecksumList> checksum_list;
+
+  if (s.ok() && get_live_table_checksum) {
+    checksum_list.reset(NewFileChecksumList());
+    // should succeed even without checksum info present, else manifest
+    // is corrupt
+    s = GetFileChecksumsFromManifest(db_->GetEnv(),
+                                     db_->GetName() + manifest_fname,
+                                     manifest_file_size, checksum_list.get());
+  }
+
+  // copy/hard link live table files
+  for (auto& ltf : live_table_files) {
+    if (!s.ok()) {
+      break;
     }
-    std::string src_fname = live_files[i];
+    std::string& src_fname = ltf.first;
+    uint64_t number = ltf.second;
 
     // rules:
-    // * if it's kTableFile, then it's shared
-    // * if it's kDescriptorFile, limit the size to manifest_file_size
-    // * always copy if cross-device link
-    if ((type == kTableFile) && same_fs) {
-      s = link_file_cb(db_->GetName(), src_fname, type);
+    // * for kTableFile, attempt hard link instead of copy.
+    // * but can't hard link across filesystems.
+    if (same_fs) {
+      s = link_file_cb(db_->GetName(), src_fname, kTableFile);
       if (s.IsNotSupported()) {
         same_fs = false;
         s = Status::OK();
       }
     }
-    if ((type != kTableFile) || (!same_fs)) {
+    if (!same_fs) {
       std::string checksum_name = kUnknownFileChecksumFuncName;
       std::string checksum_value = kUnknownFileChecksum;
+
       // we ignore the checksums either they are not required or we failed to
       // obtain the checksum lsit for old table files that have no file
       // checksums
-      if (type == kTableFile && get_live_table_checksum &&
-          checksum_status.ok()) {
+      if (get_live_table_checksum) {
         // find checksum info for table files
-        s = checksum_list->SearchOneFileChecksum(number, &checksum_value,
-                                                 &checksum_name);
+        Status search = checksum_list->SearchOneFileChecksum(
+            number, &checksum_value, &checksum_name);
 
-        // XXX/FIXME(peterd): There's currently a race between GetLiveFiles
-        // and GetLiveFilesChecksumInfo that could lead to not finding
-        // checksum info on a file that has it. For now, we can accept
-        // that and treat it like a legacy file lacking checksum info.
-        if (!s.ok()) {
+        // could be a legacy file lacking checksum info. overall OK if
+        // not found
+        if (!search.ok()) {
           assert(checksum_name == kUnknownFileChecksumFuncName);
           assert(checksum_value == kUnknownFileChecksum);
-          s = Status::OK();
         }
       }
-      s = copy_file_cb(db_->GetName(), src_fname,
-                       (type == kDescriptorFile) ? manifest_file_size : 0, type,
-                       checksum_name, checksum_value);
+      s = copy_file_cb(db_->GetName(), src_fname, 0, kTableFile, checksum_name,
+                       checksum_value);
     }
   }
   if (s.ok() && !current_fname.empty() && !manifest_fname.empty()) {

utilities/checkpoint/checkpoint_impl.h

@@ -58,7 +58,7 @@ class CheckpointImpl : public Checkpoint {
                            const std::string& contents, FileType type)>
           create_file_cb,
       uint64_t* sequence_number, uint64_t log_size_for_flush,
-      const bool& get_live_table_checksum = false);
+      bool get_live_table_checksum = false);
 
  private:
   void CleanStagingDirectory(const std::string& path, Logger* info_log);