rocksdb/db_stress_tool/expected_state.h

288 lines
9.9 KiB
C
Raw Permalink Normal View History

Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
// Copyright (c) 2021-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
#ifdef GFLAGS
#pragma once
#include <stdint.h>
#include <atomic>
#include <memory>
#include "db/dbformat.h"
#include "file/file_util.h"
#include "rocksdb/db.h"
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
#include "rocksdb/env.h"
#include "rocksdb/file_system.h"
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
#include "rocksdb/rocksdb_namespace.h"
#include "rocksdb/types.h"
#include "util/string_util.h"
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
namespace ROCKSDB_NAMESPACE {
// An `ExpectedState` provides read/write access to expected values for every
// key.
class ExpectedState {
public:
explicit ExpectedState(size_t max_key, size_t num_column_families);
virtual ~ExpectedState() {}
// Requires external locking preventing concurrent execution with any other
// member function.
virtual Status Open(bool create) = 0;
// Requires external locking covering all keys in `cf`.
void ClearColumnFamily(int cf);
// @param pending True if the update may have started but is not yet
// guaranteed finished. This is useful for crash-recovery testing when the
// process may crash before updating the expected values array.
//
// Requires external locking covering `key` in `cf`.
void Put(int cf, int64_t key, uint32_t value_base, bool pending);
// Requires external locking covering `key` in `cf`.
uint32_t Get(int cf, int64_t key) const;
// @param pending See comment above Put()
// Returns true if the key was not yet deleted.
//
// Requires external locking covering `key` in `cf`.
bool Delete(int cf, int64_t key, bool pending);
// @param pending See comment above Put()
// Returns true if the key was not yet deleted.
//
// Requires external locking covering `key` in `cf`.
bool SingleDelete(int cf, int64_t key, bool pending);
// @param pending See comment above Put()
// Returns number of keys deleted by the call.
//
// Requires external locking covering keys in `[begin_key, end_key)` in `cf`.
int DeleteRange(int cf, int64_t begin_key, int64_t end_key, bool pending);
// Requires external locking covering `key` in `cf`.
bool Exists(int cf, int64_t key);
private:
// Requires external locking covering `key` in `cf`.
std::atomic<uint32_t>& Value(int cf, int64_t key) const {
return values_[cf * max_key_ + key];
}
const size_t max_key_;
const size_t num_column_families_;
protected:
size_t GetValuesLen() const {
return sizeof(std::atomic<uint32_t>) * num_column_families_ * max_key_;
}
// Requires external locking preventing concurrent execution with any other
// member function.
void Reset();
std::atomic<uint32_t>* values_;
};
// A `FileExpectedState` implements `ExpectedState` backed by a file.
class FileExpectedState : public ExpectedState {
public:
explicit FileExpectedState(std::string expected_state_file_path,
size_t max_key, size_t num_column_families);
// Requires external locking preventing concurrent execution with any other
// member function.
Status Open(bool create) override;
private:
const std::string expected_state_file_path_;
std::unique_ptr<MemoryMappedFileBuffer> expected_state_mmap_buffer_;
};
// An `AnonExpectedState` implements `ExpectedState` backed by a memory
// allocation.
class AnonExpectedState : public ExpectedState {
public:
explicit AnonExpectedState(size_t max_key, size_t num_column_families);
// Requires external locking preventing concurrent execution with any other
// member function.
Status Open(bool create) override;
private:
std::unique_ptr<std::atomic<uint32_t>[]> values_allocation_;
};
// An `ExpectedStateManager` manages data about the expected state of the
// database. It exposes operations for reading and modifying the latest
// expected state.
class ExpectedStateManager {
public:
explicit ExpectedStateManager(size_t max_key, size_t num_column_families);
virtual ~ExpectedStateManager();
// Requires external locking preventing concurrent execution with any other
// member function.
virtual Status Open() = 0;
// Saves expected values for the current state of `db` and begins tracking
// changes. Following a successful `SaveAtAndAfter()`, `Restore()` can be
// called on the same DB, as long as its state does not roll back to before
// its current state.
//
// Requires external locking preventing concurrent execution with any other
// member function. Furthermore, `db` must not be mutated while this function
// is executing.
virtual Status SaveAtAndAfter(DB* db) = 0;
// Returns true if at least one state of historical expected values can be
// restored.
//
// Requires external locking preventing concurrent execution with any other
// member function.
virtual bool HasHistory() = 0;
// Restores expected values according to the current state of `db`. See
// `SaveAtAndAfter()` for conditions where this can be called.
//
// Requires external locking preventing concurrent execution with any other
// member function. Furthermore, `db` must not be mutated while this function
// is executing.
virtual Status Restore(DB* db) = 0;
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
// Requires external locking covering all keys in `cf`.
void ClearColumnFamily(int cf) { return latest_->ClearColumnFamily(cf); }
// @param pending True if the update may have started but is not yet
// guaranteed finished. This is useful for crash-recovery testing when the
// process may crash before updating the expected values array.
//
// Requires external locking covering `key` in `cf`.
void Put(int cf, int64_t key, uint32_t value_base, bool pending) {
return latest_->Put(cf, key, value_base, pending);
}
// Requires external locking covering `key` in `cf`.
uint32_t Get(int cf, int64_t key) const { return latest_->Get(cf, key); }
// @param pending See comment above Put()
// Returns true if the key was not yet deleted.
//
// Requires external locking covering `key` in `cf`.
bool Delete(int cf, int64_t key, bool pending) {
return latest_->Delete(cf, key, pending);
}
// @param pending See comment above Put()
// Returns true if the key was not yet deleted.
//
// Requires external locking covering `key` in `cf`.
bool SingleDelete(int cf, int64_t key, bool pending) {
return latest_->SingleDelete(cf, key, pending);
}
// @param pending See comment above Put()
// Returns number of keys deleted by the call.
//
// Requires external locking covering keys in `[begin_key, end_key)` in `cf`.
int DeleteRange(int cf, int64_t begin_key, int64_t end_key, bool pending) {
return latest_->DeleteRange(cf, begin_key, end_key, pending);
}
// Requires external locking covering `key` in `cf`.
bool Exists(int cf, int64_t key) { return latest_->Exists(cf, key); }
protected:
const size_t max_key_;
const size_t num_column_families_;
std::unique_ptr<ExpectedState> latest_;
};
// A `FileExpectedStateManager` implements an `ExpectedStateManager` backed by
// a directory of files containing data about the expected state of the
// database.
class FileExpectedStateManager : public ExpectedStateManager {
public:
explicit FileExpectedStateManager(size_t max_key, size_t num_column_families,
std::string expected_state_dir_path);
// Requires external locking preventing concurrent execution with any other
// member function.
Status Open() override;
// See `ExpectedStateManager::SaveAtAndAfter()` API doc.
//
// This implementation makes a copy of "LATEST.state" into
// "<current seqno>.state", and starts a trace in "<current seqno>.trace".
// Due to using external files, a following `Restore()` can happen even
// from a different process.
Status SaveAtAndAfter(DB* db) override;
// See `ExpectedStateManager::HasHistory()` API doc.
bool HasHistory() override;
// See `ExpectedStateManager::Restore()` API doc.
//
// Say `db->GetLatestSequenceNumber()` was `a` last time `SaveAtAndAfter()`
// was called and now it is `b`. Then this function replays `b - a` write
// operations from "`a`.trace" onto "`a`.state", and then copies the resulting
// file into "LATEST.state".
Status Restore(DB* db) override;
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
private:
// Requires external locking preventing concurrent execution with any other
// member function.
Status Clean();
std::string GetTempPathForFilename(const std::string& filename);
std::string GetPathForFilename(const std::string& filename);
static const std::string kLatestBasename;
static const std::string kStateFilenameSuffix;
static const std::string kTraceFilenameSuffix;
static const std::string kTempFilenamePrefix;
static const std::string kTempFilenameSuffix;
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
const std::string expected_state_dir_path_;
SequenceNumber saved_seqno_ = kMaxSequenceNumber;
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
};
// An `AnonExpectedStateManager` implements an `ExpectedStateManager` backed by
// a memory allocation containing data about the expected state of the database.
class AnonExpectedStateManager : public ExpectedStateManager {
public:
explicit AnonExpectedStateManager(size_t max_key, size_t num_column_families);
// See `ExpectedStateManager::SaveAtAndAfter()` API doc.
//
// This implementation returns `Status::NotSupported` since we do not
// currently have a need to keep history of expected state within a process.
Status SaveAtAndAfter(DB* /* db */) override {
return Status::NotSupported();
}
// See `ExpectedStateManager::HasHistory()` API doc.
bool HasHistory() override { return false; }
// See `ExpectedStateManager::Restore()` API doc.
//
// This implementation returns `Status::NotSupported` since we do not
// currently have a need to keep history of expected state within a process.
Status Restore(DB* /* db */) override { return Status::NotSupported(); }
Refactor expected state in stress/crash test (#8913) Summary: This is a precursor refactoring to enable an upcoming feature: persistence failure correctness testing. - Changed `--expected_values_path` to `--expected_values_dir` and migrated "db_crashtest.py" to use the new flag. For persistence failure correctness testing there are multiple possible correct states since unsynced data is allowed to be dropped. Making it possible to restore all these possible correct states will eventually involve files containing snapshots of expected values and DB trace files. - The expected values directory is managed by an `ExpectedStateManager` instance. Managing expected state files is separated out of `SharedState` to prevent `SharedState` from becoming too complex when the new files and features (snapshotting, tracing, and restoring) are introduced. - Migrated expected values file access/management out of `SharedState` into a separate class called `ExpectedState`. This is not exposed directly to the test but rather the `ExpectedState` for the latest values file is accessed via a pass-through API on `ExpectedStateManager`. This forces the test to always access the single latest `ExpectedState`. - Changed the initialization of the latest expected values file to use a tempfile followed by rename, and also add cleanup logic for possible stranded tempfiles. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8913 Test Plan: run in several ways; try to make sure it's not obviously broken. - crashtest blackbox without TEST_TMPDIR ``` $ python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest blackbox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none ``` - crashtest whitebox with TEST_TMPDIR ``` $ TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py whitebox --simple --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --duration=120 --interval=10 --compression_type=none --blob_compression_type=none --random_kill_odd=88887 ``` - db_stress without expected_values_dir ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true ``` - db_stress with expected_values_dir and manual corruption ``` $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=true --expected_values_dir=./ // modify one byte in "./LATEST.state" $ ./db_stress --write_buffer_size=1048576 --target_file_size_base=1048576 --max_bytes_for_level_base=4194304 --max_key=100000 --value_size_mult=33 --compression_type=none --ops_per_thread=10000 --clear_column_family_one_in=0 --destroy_db_initially=false --expected_values_dir=./ ... Verification failed for column family 0 key 0000000000000000 (0): Value not found: NotFound: ... ``` Reviewed By: riversand963 Differential Revision: D30921951 Pulled By: ajkr fbshipit-source-id: babfe218062e55d018c9b046536c0289fb78f41c
2021-09-28 23:12:23 +02:00
// Requires external locking preventing concurrent execution with any other
// member function.
Status Open() override;
};
} // namespace ROCKSDB_NAMESPACE
#endif // GFLAGS