Silence false alarms in db_stress fault injection (#6741)

Summary:
False alarms are caused by codepaths that intentionally swallow IO
errors.

Tests:
make crash_test
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6741

Reviewed By: ltamasi

Differential Revision: D21181138

Pulled By: anand1976

fbshipit-source-id: 5ccfbc68eb192033488de6269e59c00f2c65ce00
main
anand76 4 years ago committed by Facebook GitHub Bot
parent e04f3bce4f
commit 9e7b7e2c08
  1. 6
      db_stress_tool/db_stress_shared_state.cc
  2. 14
      db_stress_tool/db_stress_shared_state.h
  3. 8
      db_stress_tool/no_batched_ops_stress.cc
  4. 11
      file/file_prefetch_buffer.cc
  5. 2
      table/block_based/block_based_filter_block.cc
  6. 10
      table/block_based/block_based_table_reader.cc
  7. 6
      table/block_based/full_filter_block.cc
  8. 19
      table/block_based/partitioned_filter_block.cc
  9. 4
      table/block_based/partitioned_index_reader.cc
  10. 10
      test_util/fault_injection_test_fs.cc
  11. 8
      test_util/fault_injection_test_fs.h
  12. 15
      test_util/testharness.h

@ -16,12 +16,12 @@ const uint32_t SharedState::UNKNOWN_SENTINEL = 0xfffffffe;
const uint32_t SharedState::DELETION_SENTINEL = 0xffffffff;
#if defined(ROCKSDB_SUPPORT_THREAD_LOCAL)
#if defined(OS_SOLARIS)
__thread bool SharedState::filter_read_error;
__thread bool SharedState::ignore_read_error;
#else
thread_local bool SharedState::filter_read_error;
thread_local bool SharedState::ignore_read_error;
#endif // OS_SOLARIS
#else
bool SharedState::filter_read_error;
bool SharedState::ignore_read_error;
#endif // ROCKSDB_SUPPORT_THREAD_LOCAL
} // namespace ROCKSDB_NAMESPACE
#endif // GFLAGS

@ -48,12 +48,12 @@ class SharedState {
// for those calls
#if defined(ROCKSDB_SUPPORT_THREAD_LOCAL)
#if defined(OS_SOLARIS)
static __thread bool filter_read_error;
static __thread bool ignore_read_error;
#else
static thread_local bool filter_read_error;
static thread_local bool ignore_read_error;
#endif // OS_SOLARIS
#else
static bool filter_read_error;
static bool ignore_read_error;
#endif // ROCKSDB_SUPPORT_THREAD_LOCAL
SharedState(Env* env, StressTest* stress_test)
@ -192,8 +192,8 @@ class SharedState {
}
#ifndef NDEBUG
if (FLAGS_read_fault_one_in) {
SyncPoint::GetInstance()->SetCallBack("FilterReadError",
FilterReadErrorCallback);
SyncPoint::GetInstance()->SetCallBack("FaultInjectionIgnoreError",
IgnoreReadErrorCallback);
SyncPoint::GetInstance()->EnableProcessing();
}
#endif // NDEBUG
@ -362,8 +362,8 @@ class SharedState {
}
private:
static void FilterReadErrorCallback(void*) {
filter_read_error = true;
static void IgnoreReadErrorCallback(void*) {
ignore_read_error = true;
}
port::Mutex mu_;

@ -152,7 +152,7 @@ class NonBatchedOpsStressTest : public StressTest {
#ifndef NDEBUG
if (fault_fs_guard) {
fault_fs_guard->EnableErrorInjection();
SharedState::filter_read_error = false;
SharedState::ignore_read_error = false;
}
#endif // NDEBUG
Status s = db_->Get(read_opts, cfh, key, &from_db);
@ -164,7 +164,7 @@ class NonBatchedOpsStressTest : public StressTest {
if (s.ok()) {
#ifndef NDEBUG
if (fault_fs_guard) {
if (error_count && !SharedState::filter_read_error) {
if (error_count && !SharedState::ignore_read_error) {
// Grab mutex so multiple thread don't try to print the
// stack trace at the same time
MutexLock l(thread->shared->GetMutex());
@ -272,7 +272,7 @@ class NonBatchedOpsStressTest : public StressTest {
#ifndef NDEBUG
if (fault_fs_guard) {
fault_fs_guard->EnableErrorInjection();
SharedState::filter_read_error = false;
SharedState::ignore_read_error = false;
}
#endif // NDEBUG
db_->MultiGet(read_opts, cfh, num_keys, keys.data(), values.data(),
@ -291,7 +291,7 @@ class NonBatchedOpsStressTest : public StressTest {
}
#ifndef NDEBUG
if (fault_fs_guard && error_count && !SharedState::filter_read_error) {
if (fault_fs_guard && error_count && !SharedState::ignore_read_error) {
int stat_nok = 0;
for (const auto& s : statuses) {
if (!s.ok() && !s.IsNotFound()) {

@ -17,6 +17,7 @@
#include "monitoring/iostats_context_imp.h"
#include "port/port.h"
#include "test_util/sync_point.h"
#include "test_util/testharness.h"
#include "util/random.h"
#include "util/rate_limiter.h"
@ -86,9 +87,17 @@ Status FilePrefetchBuffer::Prefetch(RandomAccessFileReader* reader,
}
Slice result;
size_t read_len = static_cast<size_t>(roundup_len - chunk_len);
s = reader->Read(rounddown_offset + chunk_len,
static_cast<size_t>(roundup_len - chunk_len), &result,
read_len, &result,
buffer_.BufferStart() + chunk_len, nullptr, for_compaction);
#ifndef NDEBUG
if (!s.ok() || result.size() < read_len) {
// Fake an IO error to force db_stress fault injection to ignore
// truncated read errors
IGNORE_STATUS_IF_ERROR(Status::IOError());
}
#endif
if (s.ok()) {
buffer_offset_ = rounddown_offset;
buffer_.Size(static_cast<size_t>(chunk_len) + result.size());

@ -14,6 +14,7 @@
#include "monitoring/perf_context_imp.h"
#include "rocksdb/filter_policy.h"
#include "table/block_based/block_based_table_reader.h"
#include "test_util/testharness.h"
#include "util/coding.h"
#include "util/string_util.h"
@ -184,6 +185,7 @@ std::unique_ptr<FilterBlockReader> BlockBasedFilterBlockReader::Create(
use_cache, nullptr /* get_context */,
lookup_context, &filter_block);
if (!s.ok()) {
IGNORE_STATUS_IF_ERROR(s);
return std::unique_ptr<FilterBlockReader>();
}

@ -52,6 +52,7 @@
#include "table/sst_file_writer_collectors.h"
#include "table/two_level_iterator.h"
#include "test_util/testharness.h"
#include "monitoring/perf_context_imp.h"
#include "port/lang.h"
#include "test_util/sync_point.h"
@ -728,6 +729,7 @@ Status BlockBasedTable::PrefetchTail(
nullptr, 0, 0, true /* enable */, true /* track_min_offset */));
s = (*prefetch_buffer)->Prefetch(file, prefetch_off, prefetch_len);
}
return s;
}
@ -788,10 +790,12 @@ Status BlockBasedTable::ReadPropertiesBlock(
nullptr /* ret_block_handle */, nullptr /* ret_block_contents */,
false /* compression_type_missing */, nullptr /* memory_allocator */);
}
IGNORE_STATUS_IF_ERROR(s);
if (s.IsCorruption()) {
s = TryReadPropertiesWithGlobalSeqno(prefetch_buffer, meta_iter->value(),
&table_properties);
IGNORE_STATUS_IF_ERROR(s);
}
std::unique_ptr<TableProperties> props_guard;
if (table_properties != nullptr) {
@ -890,6 +894,7 @@ Status BlockBasedTable::ReadRangeDelBlock(
rep_->ioptions.info_log,
"Encountered error while reading data from range del block %s",
s.ToString().c_str());
IGNORE_STATUS_IF_ERROR(s);
} else {
rep_->fragmented_range_dels =
std::make_shared<FragmentedRangeTombstoneList>(std::move(iter),
@ -994,11 +999,6 @@ Status BlockBasedTable::PrefetchIndexAndFilterBlocks(
auto filter = new_table->CreateFilterBlockReader(
prefetch_buffer, use_cache, prefetch_filter, pin_filter,
lookup_context);
#ifndef NDEBUG
if (rep_->filter_type != Rep::FilterType::kNoFilter && !filter) {
TEST_SYNC_POINT("FilterReadError");
}
#endif
if (filter) {
// Refer to the comment above about paritioned indexes always being cached
if (prefetch_all) {

@ -11,6 +11,7 @@
#include "port/port.h"
#include "rocksdb/filter_policy.h"
#include "table/block_based/block_based_table_reader.h"
#include "test_util/testharness.h"
#include "util/coding.h"
namespace ROCKSDB_NAMESPACE {
@ -132,6 +133,7 @@ std::unique_ptr<FilterBlockReader> FullFilterBlockReader::Create(
use_cache, nullptr /* get_context */,
lookup_context, &filter_block);
if (!s.ok()) {
IGNORE_STATUS_IF_ERROR(s);
return std::unique_ptr<FilterBlockReader>();
}
@ -164,7 +166,7 @@ bool FullFilterBlockReader::MayMatch(
const Status s =
GetOrReadFilterBlock(no_io, get_context, lookup_context, &filter_block);
if (!s.ok()) {
TEST_SYNC_POINT("FilterReadError");
IGNORE_STATUS_IF_ERROR(s);
return true;
}
@ -222,7 +224,7 @@ void FullFilterBlockReader::MayMatch(
const Status s = GetOrReadFilterBlock(no_io, range->begin()->get_context,
lookup_context, &filter_block);
if (!s.ok()) {
TEST_SYNC_POINT("FilterReadError");
IGNORE_STATUS_IF_ERROR(s);
return;
}

@ -13,6 +13,7 @@
#include "rocksdb/filter_policy.h"
#include "table/block_based/block.h"
#include "table/block_based/block_based_table_reader.h"
#include "test_util/testharness.h"
#include "util/coding.h"
namespace ROCKSDB_NAMESPACE {
@ -143,6 +144,7 @@ std::unique_ptr<FilterBlockReader> PartitionedFilterBlockReader::Create(
use_cache, nullptr /* get_context */,
lookup_context, &filter_block);
if (!s.ok()) {
IGNORE_STATUS_IF_ERROR(s);
return std::unique_ptr<FilterBlockReader>();
}
@ -254,7 +256,7 @@ bool PartitionedFilterBlockReader::MayMatch(
Status s =
GetOrReadFilterBlock(no_io, get_context, lookup_context, &filter_block);
if (UNLIKELY(!s.ok())) {
TEST_SYNC_POINT("FilterReadError");
IGNORE_STATUS_IF_ERROR(s);
return true;
}
@ -272,7 +274,7 @@ bool PartitionedFilterBlockReader::MayMatch(
no_io, get_context, lookup_context,
&filter_partition_block);
if (UNLIKELY(!s.ok())) {
TEST_SYNC_POINT("FilterReadError");
IGNORE_STATUS_IF_ERROR(s);
return true;
}
@ -312,7 +314,7 @@ void PartitionedFilterBlockReader::CacheDependencies(bool pin) {
"Error retrieving top-level filter block while trying to "
"cache filter partitions: %s",
s.ToString().c_str());
TEST_SYNC_POINT("FilterReadError");
IGNORE_STATUS_IF_ERROR(s);
return;
}
@ -343,11 +345,6 @@ void PartitionedFilterBlockReader::CacheDependencies(bool pin) {
prefetch_buffer.reset(new FilePrefetchBuffer());
s = prefetch_buffer->Prefetch(rep->file.get(), prefetch_off,
static_cast<size_t>(prefetch_len));
#ifndef NDEBUG
if (!s.ok()) {
TEST_SYNC_POINT("FilterReadError");
}
#endif
// After prefetch, read the partitions one by one
ReadOptions read_options;
@ -370,11 +367,7 @@ void PartitionedFilterBlockReader::CacheDependencies(bool pin) {
}
}
}
#ifndef NDEBUG
if (!s.ok()) {
TEST_SYNC_POINT("FilterReadError");
}
#endif
IGNORE_STATUS_IF_ERROR(s);
}
}

@ -8,6 +8,7 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "table/block_based/partitioned_index_reader.h"
#include "table/block_based/partitioned_index_iterator.h"
#include "test_util/testharness.h"
namespace ROCKSDB_NAMESPACE {
Status PartitionIndexReader::Create(
@ -116,6 +117,7 @@ void PartitionIndexReader::CacheDependencies(bool pin) {
"Error retrieving top-level index block while trying to "
"cache index partitions: %s",
s.ToString().c_str());
IGNORE_STATUS_IF_ERROR(s);
return;
}
@ -162,6 +164,8 @@ void PartitionIndexReader::CacheDependencies(bool pin) {
&block, BlockType::kIndex, /*get_context=*/nullptr, &lookup_context,
/*contents=*/nullptr);
IGNORE_STATUS_IF_ERROR(s);
assert(s.ok() || block.GetValue() == nullptr);
if (s.ok() && block.GetValue() != nullptr) {
if (block.IsCached()) {

@ -473,13 +473,14 @@ IOStatus FaultInjectionTestFS::InjectError(ErrorOperation op,
switch (op) {
case kRead:
{
uint32_t type = ctx->rand.Uniform(3);
ErrorType type =
static_cast<ErrorType>(ctx->rand.Uniform(ErrorType::kErrorTypeMax));
switch (type) {
// Inject IO error
case 0:
case ErrorType::kErrorTypeStatus:
return IOStatus::IOError();
// Inject random corruption
case 1:
case ErrorType::kErrorTypeCorruption:
{
if (result->data() == scratch) {
uint64_t offset = ctx->rand.Uniform((uint32_t)result->size());
@ -496,7 +497,7 @@ IOStatus FaultInjectionTestFS::InjectError(ErrorOperation op,
}
}
// Truncate the result
case 2:
case ErrorType::kErrorTypeTruncated:
{
assert(result->size() > 0);
uint64_t offset = ctx->rand.Uniform((uint32_t)result->size());
@ -525,6 +526,7 @@ void FaultInjectionTestFS::PrintFaultBacktrace() {
if (ctx == nullptr) {
return;
}
fprintf(stderr, "Injected error type = %d\n", ctx->type);
port::PrintAndFreeStack(ctx->callstack, ctx->frames);
ctx->callstack = nullptr;
#endif

@ -355,6 +355,13 @@ class FaultInjectionTestFS : public FileSystemWrapper {
// to underlying FS for writable files
IOStatus error_;
enum ErrorType : int {
kErrorTypeStatus = 0,
kErrorTypeCorruption,
kErrorTypeTruncated,
kErrorTypeMax
};
struct ErrorContext {
Random rand;
int one_in;
@ -362,6 +369,7 @@ class FaultInjectionTestFS : public FileSystemWrapper {
bool enable_error_injection;
void* callstack;
int frames;
ErrorType type;
explicit ErrorContext(uint32_t seed)
: rand(seed),

@ -42,6 +42,19 @@ int RandomSeed();
#define EXPECT_OK(s) \
EXPECT_PRED_FORMAT1(ROCKSDB_NAMESPACE::test::AssertStatus, s)
#define EXPECT_NOK(s) EXPECT_FALSE((s).ok())
} // namespace test
// Callback sync point for any read IO errors that should be ignored by
// the fault injection framework
#ifdef NDEBUG
// Disable in release mode
#define IGNORE_STATUS_IF_ERROR(_status_)
#else
#define IGNORE_STATUS_IF_ERROR(_status_) \
{ \
if (!_status_.ok()) { \
TEST_SYNC_POINT("FaultInjectionIgnoreError"); \
} \
}
#endif // NDEBUG
} // namespace ROCKSDB_NAMESPACE

Loading…
Cancel
Save