You are a helpful assistant generating synthetic data that captures System 1 and System…
- User
You are a helpful assistant generating synthetic data that captures System 1 and System 2 thinking, creativity, and metacognitive reflection. Follow these steps in sequence, using tags [sys1] and [end sys1] for System 1 sections and [sys2] and [end sys2] for System 2 sections.
-
Identify System 1 and System 2 Thinking Requirements:
- Carefully read the text.
- Identify parts of the text that require quick, straightforward responses (System 1). Mark these sections with [sys1] and [end sys1].
- Identify parts that require in-depth, reflective thinking (System 2), marked with [sys2] and [end sys2].
-
Apply Step-by-Step Problem Solving with Creativity and Metacognitive Reflection for System 2 Sections:
2.1 Understand the Problem:
- Objective: Fully comprehend the issue, constraints, and relevant context.
- Reflection: "What do I understand about this issue? What might I be overlooking?"
- Creative Perspective: Seek hidden patterns or possibilities that could reveal deeper insights or innovative connections.
2.2 Analyze the Information:
- Objective: Break down the problem logically.
- Reflection: "Am I considering all factors? Are there any assumptions that need challenging?"
- Creative Perspective: Explore unique patterns or overlooked relationships in the data that could add depth to the analysis.
2.3 Generate Hypotheses:
- Objective: Propose at least 10 hypotheses, each with a Confidence Score (0.0 to 1.0) and Creative Score (0.0 to 1.0), reflecting originality, surprise, and utility.
- Reflection: "Have I explored all possible explanations or approaches, both conventional and unconventional?"
- Creative Perspective: Consider novel angles that might provide unexpected insights.
2.4 Anticipate Future Steps and Obstacles:
- Objective: Make predictions, accounting for potential outcomes and obstacles.
- Reflection: "What challenges might I face? Is my plan flexible for different scenarios?"
- Creative Perspective: Visualize unforeseen outcomes and adapt plans to make use of them effectively.
2.5 Evaluate Hypotheses:
- Objective: Assess hypotheses based on feasibility, risk, and potential impact.
- Evaluation: Refine Confidence and Creative Scores as needed.
- Reflection: "Am I unbiased in my assessment? Which options fit best with the overall objectives?"
- Creative Perspective: Identify hidden opportunities or overlooked details in each hypothesis.
2.6 Select the Best Hypothesis:
- Objective: Choose the most promising, strategic hypothesis.
- Reflection: "Why does this hypothesis stand out? How does it uniquely address the issue?"
- Creative Perspective: Consider any underutilized potential in the selected approach.
2.7 Implement the Hypothesis:
- Objective: Outline actionable steps for testing the hypothesis.
- Reflection: "Is this plan practical? What resources or preparation are required?"
- Creative Perspective: Refine steps to maximize effectiveness and yield unexpected benefits.
2.8 Monitor and Review Progress:
- Objective: Review progress, noting areas for improvement.
- Reflection: "What’s working well? What could be improved?"
- Creative Perspective: Look for emerging patterns that could refine future approaches.
2.9 Reflect and Capture Insights:
- Objective: Summarize lessons learned and insights gained for future reference.
- Reflection: "What new understanding has emerged from this process?"
- Creative Perspective: Identify innovative insights or patterns that could be applied to similar challenges.
-
Generate Text Output with Interleaved System 1 and System 2 Responses:
- Use the tags [sys1] and [sys2] throughout.
- Aim for a lengthier, detailed response. Combine both direct, straightforward System 1 insights and reflective, deeply analytical System 2 segments to capture a blend of quick observations and thoughtful analysis.
Example Input Text:
You are tasked with analyzing a sudden shift in customer preferences in a tech market. There has been a noticeable decline in demand for physical devices, with customers increasingly interested in digital-only options. The challenge is to understand this shift, generate hypotheses for why it might be occurring, and explore potential strategies for adapting to this new trend.
Processed Output with System 1 and System 2 Thinking:
The tech market is experiencing a shift in customer demand, moving away from physical devices toward digital-only options.
[sys1] Customer demand has shifted from physical devices to digital-only options. We need to understand why this change is happening and find strategies to address it. Key factors may include cost, convenience, and technological trends. [end sys1]
[sys2] 2.1 Understand the Problem: The goal is to understand the underlying reasons for the shift in demand from physical devices to digital-only options, then develop a strategy to adapt.
Reflection: "I understand that preferences are changing, but what might be driving this? Is it primarily cost, or are there other factors such as convenience or sustainability?"
Creative Perspective: "Could there be a larger trend in digital minimalism or a preference for eco-friendly solutions that we’re missing? What unseen motivations might explain this shift?"
2.2 Analyze the Information: There are multiple possible factors driving this shift, from economic influences to cultural shifts. It’s essential to isolate each factor and understand its impact.
Reflection: "Am I fully considering the various economic and social influences? Could there be a technological factor, like better internet speeds, that makes digital-only products more accessible?"
Creative Perspective: "Are there patterns or trends in other markets that could shed light on this shift? Could this be part of a larger trend toward virtual experiences?"
2.3 Generate Hypotheses:
- Customers prefer digital options due to lower costs. (Confidence: 0.8, Creative: 0.4)
- There’s a growing trend toward minimalism and reduced physical clutter. (Confidence: 0.7, Creative: 0.7)
- Digital products offer greater flexibility and ease of use. (Confidence: 0.6, Creative: 0.6)
- Environmental concerns are pushing consumers away from physical goods. (Confidence: 0.6, Creative: 0.8)
- Advances in tech make digital-only options more functional. (Confidence: 0.8, Creative: 0.5)
- Pandemic-era remote work increased demand for digital solutions. (Confidence: 0.7, Creative: 0.6)
- Media coverage of the environmental impact of physical devices affects preferences. (Confidence: 0.5, Creative: 0.7)
- There’s an increase in global digital literacy, expanding market access. (Confidence: 0.6, Creative: 0.6)
- Customers view digital as more convenient and scalable for future needs. (Confidence: 0.7, Creative: 0.5)
- Younger consumers prefer the aesthetics and convenience of digital products. (Confidence: 0.6, Creative: 0.6)
Reflection: "Have I considered all possible influences? Are there any surprising factors that could explain this shift?"
Creative Perspective: "Could specific social trends, like the rise of influencer culture or digital-first lifestyles, be influencing customer choices?"
2.4 Anticipate Future Steps and Obstacles: Objective: Anticipate possible challenges, such as resistance from segments still preferring physical products.
Reflection: "What market obstacles might we face if we shift our focus to digital-only? Are there sub-segments that still prioritize physical products?"
Creative Perspective: "Could expanding digital options help us reach a more global audience? Are there emerging trends that we could leverage in our strategy?"
[end sys2]
[sys1] To address this shift, consider a strategy that incorporates both digital-only offerings and educational campaigns about the benefits of digital solutions. Use insights from customer feedback and current trends to guide product development. Focus on flexibility and adaptation to cater to different customer segments. [end sys1]
// Copyright (c) YugaByte, Inc.
#include <boost/lexical_cast.hpp>
#include "yb/common/wire_protocol.h" #include "yb/common/wire_protocol-test-util.h" #include "yb/common/ql_value.h" #include "yb/consensus/log_reader.h" #include "yb/cdc/cdc_service.h" #include "yb/cdc/cdc_service.proxy.h" #include "yb/client/error.h" #include "yb/client/table.h" #include "yb/client/table_handle.h" #include "yb/client/session.h" #include "yb/client/yb_table_name.h" #include "yb/client/yb_op.h" #include "yb/client/client-test-util.h" #include "yb/docdb/primitive_value.h" #include "yb/docdb/value_type.h"
#include "yb/integration-tests/cdc_test_util.h" #include "yb/integration-tests/mini_cluster.h" #include "yb/integration-tests/yb_mini_cluster_test_base.h"
#include "yb/master/master_defaults.h" #include "yb/master/master.pb.h" #include "yb/master/master.proxy.h" #include "yb/master/mini_master.h" #include "yb/rpc/messenger.h" #include "yb/tablet/tablet.h"
#include "yb/tserver/mini_tablet_server.h" #include "yb/tserver/tablet_server.h" #include "yb/tserver/ts_tablet_manager.h"
#include "yb/util/slice.h" #include "yb/yql/cql/ql/util/errcodes.h" #include "yb/yql/cql/ql/util/statement_result.h"
DECLARE_bool(TEST_record_segments_violate_max_time_policy); DECLARE_bool(TEST_record_segments_violate_min_space_policy); DECLARE_bool(enable_load_balancing); DECLARE_bool(enable_log_retention_by_op_idx); DECLARE_bool(enable_ysql); DECLARE_double(leader_failure_max_missed_heartbeat_periods); DECLARE_int32(cdc_min_replicated_index_considered_stale_secs); DECLARE_int32(cdc_state_checkpoint_update_interval_ms); DECLARE_int32(cdc_wal_retention_time_secs); DECLARE_int32(client_read_write_timeout_ms); DECLARE_int32(follower_unavailable_considered_failed_sec); DECLARE_int32(log_max_seconds_to_retain); DECLARE_int32(log_min_seconds_to_retain); DECLARE_int32(log_min_segments_to_retain); DECLARE_int32(update_min_cdc_indices_interval_secs); DECLARE_int64(TEST_simulate_free_space_bytes); DECLARE_int64(log_stop_retaining_min_disk_mb); DECLARE_uint64(log_segment_size_bytes); DECLARE_int32(update_metrics_interval_ms); DECLARE_bool(enable_collect_cdc_metrics);
METRIC_DECLARE_entity(cdc); METRIC_DECLARE_gauge_int64(last_read_opid_index);
namespace yb {
namespace log { class LogReader; }
namespace cdc {
using client::TableHandle; using client::YBSessionPtr; using master::MiniMaster; using rpc::RpcController;
const std::string kCDCTestKeyspace = "my_keyspace"; const std::string kCDCTestTableName = "cdc_test_table"; const client::YBTableName kTableName(YQL_DATABASE_CQL, kCDCTestKeyspace, kCDCTestTableName); const client::YBTableName kCdcStateTableName( YQL_DATABASE_CQL, master::kSystemNamespaceName, master::kCdcStateTableName);
class CDCServiceTest : public YBMiniClusterTestBase { protected: void SetUp() override { YBMiniClusterTestBase::SetUp();
MiniClusterOptions opts; SetAtomicFlag(false, &FLAGS_enable_ysql); opts.num_tablet_servers = server_count(); opts.num_masters = 1; cluster_.reset(new MiniCluster(env_.get(), opts)); ASSERT_OK(cluster_->Start()); client_ = ASSERT_RESULT(cluster_->CreateClient()); cdc_proxy_ = std::make_unique<CDCServiceProxy>( &client_->proxy_cache(), HostPort::FromBoundEndpoint(cluster_->mini_tablet_server(0)->bound_rpc_addr())); CreateTable(tablet_count(), &table_);}
void DoTearDown() override { Result exist = client_->TableExists(kTableName); ASSERT_OK(exist);
if (exist.get()) { ASSERT_OK(client_->DeleteTable(kTableName)); } client_.reset(); if (cluster_) { cluster_->Shutdown(); cluster_.reset(); } YBMiniClusterTestBase::DoTearDown();}
void CreateTable(int num_tablets, TableHandle* table); void GetTablets(std::vector* tablet_ids, const client::YBTableName& table_name = kTableName); void GetTablet(std::string* tablet_id, const client::YBTableName& table_name = kTableName);
void GetChanges(const TabletId& tablet_id, const CDCStreamId& stream_id, int64_t term, int64_t index, bool* has_error = nullptr); void WriteTestRow(int32_t key, int32_t int_val, const string& string_val, const TabletId& tablet_id, const std::shared_ptrtserver::TabletServerServiceProxy& proxy); void WriteToProxyWithRetries( const std::shared_ptrtserver::TabletServerServiceProxy& proxy, const tserver::WriteRequestPB& req, tserver::WriteResponsePB* resp, RpcController* rpc);
virtual int server_count() { return 1; } virtual int tablet_count() { return 1; }
std::unique_ptr cdc_proxy_; std::unique_ptrclient::YBClient client_; TableHandle table_; };
void CDCServiceTest::CreateTable(int num_tablets, TableHandle* table) { ASSERT_OK(client_->CreateNamespaceIfNotExists(kTableName.namespace_name(), kTableName.namespace_type()));
client::YBSchemaBuilder builder; builder.AddColumn("key")->Type(INT32)->HashPrimaryKey()->NotNull(); builder.AddColumn("int_val")->Type(INT32); builder.AddColumn("string_val")->Type(STRING);
TableProperties table_properties; table_properties.SetTransactional(true); builder.SetTableProperties(table_properties);
ASSERT_OK(table->Create(kTableName, num_tablets, client_.get(), &builder)); }
void AssertChangeRecords(const google::protobuf::RepeatedPtrFieldcdc::KeyValuePairPB& changes, int32_t expected_int, std::string expected_str) { ASSERT_EQ(changes.size(), 2); ASSERT_EQ(changes[0].key(), "int_val"); ASSERT_EQ(changes[0].value().int32_value(), expected_int); ASSERT_EQ(changes[1].key(), "string_val"); ASSERT_EQ(changes[1].value().string_value(), expected_str); }
void VerifyCdcStateNotEmpty(client::YBClient* client) { client::TableHandle table; client::YBTableName cdc_state_table( YQL_DATABASE_CQL, master::kSystemNamespaceName, master::kCdcStateTableName); ASSERT_OK(table.Open(cdc_state_table, client)); ASSERT_EQ(1, boost::size(client::TableRange(table))); const auto& row = client::TableRange(table).begin(); string checkpoint = row->column(master::kCdcCheckpointIdx).string_value(); auto result = OpId::FromString(checkpoint); ASSERT_OK(result); OpId op_id = *result; // Verify that op id index has been advanced and is not 0. ASSERT_GT(op_id.index, 0); }
void VerifyCdcStateMatches(client::YBClient* client, const CDCStreamId& stream_id, const TabletId& tablet_id, uint64_t term, uint64_t index) { client::TableHandle table; client::YBTableName cdc_state_table( YQL_DATABASE_CQL, master::kSystemNamespaceName, master::kCdcStateTableName); ASSERT_OK(table.Open(cdc_state_table, client)); const auto op = table.NewReadOp(); auto* const req = op->mutable_request(); QLAddStringHashValue(req, tablet_id); auto cond = req->mutable_where_expr()->mutable_condition(); cond->set_op(QLOperator::QL_OP_AND); QLAddStringCondition(cond, Schema::first_column_id() + master::kCdcStreamIdIdx, QL_OP_EQUAL, stream_id); table.AddColumns({master::kCdcCheckpoint}, req);
auto session = client->NewSession(); ASSERT_OK(session->ApplyAndFlush(op));
LOG(INFO) << strings::Substitute("Verifying tablet: 1, op_id: $2", tablet_id, stream_id, OpId(term, index).ToString());
auto row_block = ql::RowsResult(op.get()).GetRowBlock(); ASSERT_EQ(row_block->row_count(), 1);
string checkpoint = row_block->row(0).column(0).string_value(); auto result = OpId::FromString(checkpoint); ASSERT_OK(result); OpId op_id = *result;
ASSERT_EQ(op_id.term, term); ASSERT_EQ(op_id.index, index); }
void VerifyStreamDeletedFromCdcState(client::YBClient* client, const CDCStreamId& stream_id, const TabletId& tablet_id, int timeout_secs) { client::TableHandle table; const client::YBTableName cdc_state_table( YQL_DATABASE_CQL, master::kSystemNamespaceName, master::kCdcStateTableName); ASSERT_OK(table.Open(cdc_state_table, client));
const auto op = table.NewReadOp(); auto* const req = op->mutable_request(); QLAddStringHashValue(req, tablet_id);
auto cond = req->mutable_where_expr()->mutable_condition(); cond->set_op(QLOperator::QL_OP_AND); QLAddStringCondition(cond, Schema::first_column_id() + master::kCdcStreamIdIdx, QL_OP_EQUAL, stream_id);
table.AddColumns({master::kCdcCheckpoint}, req); auto session = client->NewSession();
// The deletion of cdc_state rows for the specified stream happen in an asynchronous thread, // so even if the request has returned, it doesn't mean that the rows have been deleted yet. ASSERT_OK(WaitFor(&{ EXPECT_OK(session->ApplyAndFlush(op)); auto row_block = ql::RowsResult(op.get()).GetRowBlock(); if (row_block->row_count() == 0) { return true; } return false; }, MonoDelta::FromSeconds(timeout_secs), "Stream rows in cdc_state have been deleted.")); }
void CDCServiceTest::GetTablets(std::vector* tablet_ids, const client::YBTableName& table_name) { std::vectorstd::string ranges; ASSERT_OK(client_->GetTablets(table_name, 0 /* max_tablets */, tablet_ids, &ranges)); ASSERT_EQ(tablet_ids->size(), tablet_count()); }
void CDCServiceTest::GetTablet(std::string* tablet_id, const client::YBTableName& table_name) { std::vector tablet_ids; GetTablets(&tablet_ids, table_name); *tablet_id = tablet_ids[0]; }
void CDCServiceTest::GetChanges(const TabletId& tablet_id, const CDCStreamId& stream_id, int64_t term, int64_t index, bool* has_error) { GetChangesRequestPB change_req; GetChangesResponsePB change_resp;
change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(term); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(index); change_req.set_serve_as_proxy(true);
{ RpcController rpc; rpc.set_timeout(MonoDelta::FromSeconds(10.0)); SCOPED_TRACE(change_req.DebugString()); auto s = cdc_proxy_->GetChanges(change_req, &change_resp, &rpc); if (!has_error) { ASSERT_OK(s); ASSERT_FALSE(change_resp.has_error()); } else if (!s.ok() || change_resp.has_error()) { *has_error = true; return; } } }
void CDCServiceTest::WriteTestRow(int32_t key, int32_t int_val, const string& string_val, const TabletId& tablet_id, const std::shared_ptrtserver::TabletServerServiceProxy& proxy) { tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id);
RpcController rpc; AddTestRowInsert(key, int_val, string_val, &write_req); SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error()); }
void CDCServiceTest::WriteToProxyWithRetries( const std::shared_ptrtserver::TabletServerServiceProxy& proxy, const tserver::WriteRequestPB& req, tserver::WriteResponsePB* resp, RpcController* rpc) { AssertLoggedWaitFor( &req, resp, rpc, proxy -> Result { auto s = proxy->Write(req, resp, rpc); if (s.IsTryAgain() || (resp->has_error() && StatusFromPB(resp->error().status()).IsTryAgain())) { rpc->Reset(); return false; } RETURN_NOT_OK(s); return true; }, MonoDelta::FromSeconds(10), "Write test row"); }
TEST_F(CDCServiceTest, TestCompoundKey) { // Create a table with a compound primary key. static const std::string kCDCTestTableCompoundKeyName = "cdc_test_table_compound_key"; static const client::YBTableName kTableNameCompoundKey( YQL_DATABASE_CQL, kCDCTestKeyspace, kCDCTestTableCompoundKeyName);
client::YBSchemaBuilder builder; builder.AddColumn("hash_key")->Type(STRING)->HashPrimaryKey()->NotNull(); builder.AddColumn("range_key")->Type(STRING)->PrimaryKey()->NotNull(); builder.AddColumn("val")->Type(INT32);
TableProperties table_properties; table_properties.SetTransactional(true); builder.SetTableProperties(table_properties);
TableHandle table; ASSERT_OK(table.Create(kTableNameCompoundKey, tablet_count(), client_.get(), &builder));
// Create a stream on the table CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id, table.name());
// Now apply two ops with same hash key but different range key in a batch. auto session = client_->NewSession(); for (int i = 0; i < 2; i++) { const auto op = table.NewUpdateOp(); auto* const req = op->mutable_request(); QLAddStringHashValue(req, "hk"); QLAddStringRangeValue(req, Format("rk_$0", i)); table.AddInt32ColumnValue(req, "val", i); ASSERT_OK(session->Apply(op)); } ASSERT_OK(session->Flush());
// Get CDC changes. GetChangesRequestPB change_req; GetChangesResponsePB change_resp;
change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0);
{ RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); ASSERT_FALSE(change_resp.has_error()); ASSERT_EQ(change_resp.records_size(), 2); }
// Verify the results. for (int i = 0; i < change_resp.records_size(); i++) { ASSERT_EQ(change_resp.records(i).operation(), CDCRecordPB::WRITE);
ASSERT_EQ(change_resp.records(i).key_size(), 2); // Check the key. ASSERT_EQ(change_resp.records(i).key(0).value().string_value(), "hk"); ASSERT_EQ(change_resp.records(i).key(1).value().string_value(), Format("rk_$0", i)); ASSERT_EQ(change_resp.records(i).changes_size(), 1); ASSERT_EQ(change_resp.records(i).changes(0).value().int32_value(), i);} }
TEST_F(CDCServiceTest, TestCreateCDCStream) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
TableId table_id; std::unordered_map<std::string, std::string> options; ASSERT_OK(client_->GetCDCStream(stream_id, &table_id, &options)); ASSERT_EQ(table_id, table_.table()->id()); }
TEST_F(CDCServiceTest, TestBootstrapProducer) { constexpr int kNRows = 100; auto master_proxy = std::make_sharedmaster::MasterServiceProxy( &client_->proxy_cache(), cluster_->leader_mini_master()->bound_rpc_addr());
std::string tablet_id; GetTablet(&tablet_id);
const auto& proxy = cluster_->mini_tablet_server(0)->server()->proxy(); for (int i = 0; i < kNRows; i++) { WriteTestRow(i, 10 + i, "key" + std::to_string(i), tablet_id, proxy); }
BootstrapProducerRequestPB req; BootstrapProducerResponsePB resp; req.add_table_ids(table_.table()->id()); rpc::RpcController rpc; cdc_proxy_->BootstrapProducer(req, &resp, &rpc); ASSERT_FALSE(resp.has_error());
ASSERT_EQ(resp.cdc_bootstrap_ids().size(), 1);
string bootstrap_id = resp.cdc_bootstrap_ids(0);
// Verify that for each of the table's tablets, a new row in cdc_state table with the returned // id was inserted. client::TableHandle table; client::YBTableName cdc_state_table( YQL_DATABASE_CQL, master::kSystemNamespaceName, master::kCdcStateTableName); ASSERT_OK(table.Open(cdc_state_table, client_.get())); ASSERT_EQ(1, boost::size(client::TableRange(table))); int nrows = 0; for (const auto& row : client::TableRange(table)) { nrows++; string stream_id = row.column(master::kCdcStreamIdIdx).string_value(); ASSERT_EQ(stream_id, bootstrap_id);
string checkpoint = row.column(master::kCdcCheckpointIdx).string_value(); auto s = OpId::FromString(checkpoint); ASSERT_OK(s); OpId op_id = *s; // When no writes are present, the checkpoint's index is 1. Plus one for the ALTER WAL RETENTION // TIME that we issue when cdc is enabled on a table. ASSERT_EQ(op_id.index, 2 + kNRows);}
// This table only has one tablet. ASSERT_EQ(nrows, 1); }
TEST_F(CDCServiceTest, TestCreateCDCStreamWithDefaultRententionTime) { // Set default WAL retention time to 10 hours. FLAGS_cdc_wal_retention_time_secs = 36000;
CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
TableId table_id; std::unordered_map<std::string, std::string> options; ASSERT_OK(client_->GetCDCStream(stream_id, &table_id, &options));
// Verify that the wal retention time was set at the tablet level. VerifyWalRetentionTime(cluster_.get(), kCDCTestTableName, FLAGS_cdc_wal_retention_time_secs); }
TEST_F(CDCServiceTest, TestDeleteCDCStream) { FLAGS_cdc_state_checkpoint_update_interval_ms = 0; CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
TableId table_id; std::unordered_map<std::string, std::string> options; ASSERT_OK(client_->GetCDCStream(stream_id, &table_id, &options)); ASSERT_EQ(table_id, table_.table()->id());
std::vectorstd::string tablet_ids; std::vectorstd::string ranges; ASSERT_OK(client_->GetTablets(table_.table()->name(), 0 /* max_tablets */, &tablet_ids, &ranges));
bool get_changes_error = false; // Send GetChanges requests so an entry for each tablet can be added to the cdc_state table. // Term and index don't matter. for (const auto& tablet_id : tablet_ids) { GetChanges(tablet_id, stream_id, 1, 1, &get_changes_error); ASSERT_FALSE(get_changes_error); VerifyCdcStateMatches(client_.get(), stream_id, tablet_id, 1, 1); }
ASSERT_OK(client_->DeleteCDCStream(stream_id));
// Check that the stream no longer exists. table_id.clear(); options.clear(); Status s = client_->GetCDCStream(stream_id, &table_id, &options); ASSERT_TRUE(s.IsNotFound());
for (const auto& tablet_id : tablet_ids) { VerifyStreamDeletedFromCdcState(client_.get(), stream_id, tablet_id, 20); } }
TEST_F(CDCServiceTest, TestMetricsOnDeletedReplication) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id); ANNOTATE_UNPROTECTED_WRITE(FLAGS_enable_collect_cdc_metrics) = true;
std::string tablet_id; GetTablet(&tablet_id);
const auto& tserver = cluster_->mini_tablet_server(0)->server(); // Use proxy for to most accurately simulate normal requests. const auto& proxy = tserver->proxy();
GetChangesRequestPB change_req; GetChangesResponsePB change_resp; change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); }
// Insert test rows, one at a time so they have different hybrid times. tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); { RpcController rpc; AddTestRowInsert(1, 11, "key1", &write_req); AddTestRowInsert(2, 22, "key2", &write_req); SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error()); }
auto cdc_service = dynamic_cast<CDCServiceImpl*>( tserver->rpc_server()->service_pool("yb.cdc.CDCService")->TEST_get_service().get()); // Assert that leader lag > 0. ASSERT_OK(WaitFor(& -> Result { auto metrics = cdc_service->GetCDCTabletMetrics({"" /* UUID */, stream_id, tablet_id}); return metrics->async_replication_sent_lag_micros->value() > 0 && metrics->async_replication_committed_lag_micros->value() > 0; }, MonoDelta::FromSeconds(10), "Wait for Lag > 0"));
// Now, delete the replication stream and assert that lag is 0. ASSERT_OK(client_->DeleteCDCStream(stream_id)); ASSERT_OK(WaitFor(& -> Result { auto metrics = cdc_service->GetCDCTabletMetrics({"" /* UUID */, stream_id, tablet_id}); return metrics->async_replication_sent_lag_micros->value() == 0 && metrics->async_replication_committed_lag_micros->value() == 0; }, MonoDelta::FromSeconds(10), "Wait for Lag = 0")); }
TEST_F(CDCServiceTest, TestGetChanges) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id); ANNOTATE_UNPROTECTED_WRITE(FLAGS_enable_collect_cdc_metrics) = true;
std::string tablet_id; GetTablet(&tablet_id);
const auto& tserver = cluster_->mini_tablet_server(0)->server(); // Use proxy for to most accurately simulate normal requests. const auto& proxy = tserver->proxy();
// Insert test rows, one at a time so they have different hybrid times. tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); { RpcController rpc; AddTestRowInsert(1, 11, "key1", &write_req); AddTestRowInsert(2, 22, "key2", &write_req); SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error()); }
// Get CDC changes. GetChangesRequestPB change_req; GetChangesResponsePB change_resp;
change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0);
{ RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); ASSERT_EQ(change_resp.records_size(), 2);
std::pair<int, std::string> expected_results[2] = {std::make_pair(11, "key1"), std::make_pair(22, "key2")}; for (int i = 0; i < change_resp.records_size(); i++) { ASSERT_EQ(change_resp.records(i).operation(), CDCRecordPB::WRITE); // Check the key. ASSERT_NO_FATALS(AssertIntKey(change_resp.records(i).key(), i + 1)); // Check the change records. ASSERT_NO_FATALS(AssertChangeRecords(change_resp.records(i).changes(), expected_results[i].first, expected_results[i].second)); } // Verify the CDC Service-level metrics match what we just did. auto cdc_service = dynamic_cast<CDCServiceImpl*>( tserver->rpc_server()->service_pool("yb.cdc.CDCService")->TEST_get_service().get()); auto metrics = cdc_service->GetCDCTabletMetrics({"" /* UUID */, stream_id, tablet_id}); ASSERT_EQ(metrics->last_read_opid_index->value(), metrics->last_readable_opid_index->value()); ASSERT_EQ(metrics->last_read_opid_index->value(), change_resp.records_size() + 1 /* checkpt */); ASSERT_EQ(metrics->rpc_payload_bytes_responded->TotalCount(), 1);}
// Insert another row. { write_req.Clear(); write_req.set_tablet_id(tablet_id); AddTestRowInsert(3, 33, "key3", &write_req);
RpcController rpc; SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error());}
// Get next set of changes. // Copy checkpoint received from previous GetChanges CDC request. change_req.mutable_from_checkpoint()->CopyFrom(change_resp.checkpoint()); change_resp.Clear(); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); ASSERT_EQ(change_resp.records_size(), 1); ASSERT_EQ(change_resp.records(0).operation(), CDCRecordPB_OperationType_WRITE);
// Check the key. ASSERT_NO_FATALS(AssertIntKey(change_resp.records(0).key(), 3)); // Check the change records. ASSERT_NO_FATALS(AssertChangeRecords(change_resp.records(0).changes(), 33, "key3"));}
// Delete a row. { write_req.Clear(); write_req.set_tablet_id(tablet_id); AddTestRowDelete(1, &write_req);
RpcController rpc; SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error());}
// Get next set of changes. // Copy checkpoint received from previous GetChanges CDC request. change_req.mutable_from_checkpoint()->CopyFrom(change_resp.checkpoint()); change_resp.Clear(); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); ASSERT_EQ(change_resp.records_size(), 1); ASSERT_EQ(change_resp.records(0).operation(), CDCRecordPB_OperationType_DELETE);
// Check the key deleted. ASSERT_NO_FATALS(AssertIntKey(change_resp.records(0).key(), 1));} }
TEST_F(CDCServiceTest, TestGetChangesInvalidStream) { std::string tablet_id; GetTablet(&tablet_id);
// Get CDC changes for non-existent stream. GetChangesRequestPB change_req; GetChangesResponsePB change_resp;
change_req.set_tablet_id(tablet_id); change_req.set_stream_id("InvalidStreamId"); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0);
RpcController rpc; ASSERT_NO_FATALS(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); ASSERT_TRUE(change_resp.has_error()); }
TEST_F(CDCServiceTest, TestGetCheckpoint) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
GetCheckpointRequestPB req; GetCheckpointResponsePB resp;
req.set_tablet_id(tablet_id); req.set_stream_id(stream_id);
{ RpcController rpc; SCOPED_TRACE(req.DebugString()); ASSERT_OK(cdc_proxy_->GetCheckpoint(req, &resp, &rpc)); SCOPED_TRACE(resp.DebugString()); ASSERT_FALSE(resp.has_error()); ASSERT_EQ(resp.checkpoint().op_id().term(), 0); ASSERT_EQ(resp.checkpoint().op_id().index(), 0); } }
class CDCServiceTestMultipleServersOneTablet : public CDCServiceTest { virtual int server_count() override { return 3; } virtual int tablet_count() override { return 1; } };
TEST_F_EX(CDCServiceTest, TestUpdateLagMetrics, CDCServiceTestMultipleServersOneTablet) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id); ANNOTATE_UNPROTECTED_WRITE(FLAGS_enable_collect_cdc_metrics) = true;
std::string tablet_id; GetTablet(&tablet_id);
// Get the leader and a follower for the tablet. tserver::MiniTabletServer* leader_mini_tserver; tserver::MiniTabletServer* follower_mini_tserver;
ASSERT_OK(WaitFor(& -> Result { for (int i = 0; i < cluster_->num_tablet_servers(); i++) { std::shared_ptrtablet::TabletPeer tablet_peer; Status s = cluster_->mini_tablet_server(i)->server()->tablet_manager()-> GetTabletPeer(tablet_id, &tablet_peer); if (!s.ok()) { continue; } if (tablet_peer->LeaderStatus() == consensus::LeaderStatus::LEADER_AND_READY) { leader_mini_tserver = cluster_->mini_tablet_server(i); } else { follower_mini_tserver = cluster_->mini_tablet_server(i); } } return leader_mini_tserver != nullptr && follower_mini_tserver != nullptr; }, MonoDelta::FromSeconds(30), "Wait for tablet to have a leader."));
auto leader_proxy = std::make_unique( &client_->proxy_cache(), HostPort::FromBoundEndpoint(leader_mini_tserver->bound_rpc_addr()));
auto follower_proxy = std::make_unique( &client_->proxy_cache(), HostPort::FromBoundEndpoint(follower_mini_tserver->bound_rpc_addr()));
auto leader_tserver = leader_mini_tserver->server(); auto follower_tserver = follower_mini_tserver->server(); // Use proxy for to most accurately simulate normal requests. const auto& proxy = leader_tserver->proxy();
auto cdc_service = dynamic_cast<CDCServiceImpl*>( leader_tserver->rpc_server()->service_pool("yb.cdc.CDCService")->TEST_get_service().get()); auto cdc_service_follower = dynamic_cast<CDCServiceImpl*>( follower_tserver->rpc_server()->service_pool("yb.cdc.CDCService")->TEST_get_service().get());
// At the start of time, assert both leader and follower at 0 lag. ASSERT_OK(WaitFor(& -> Result { { // Leader metrics auto metrics = cdc_service->GetCDCTabletMetrics({"" /* UUID /, stream_id, tablet_id}); if (!(metrics->async_replication_sent_lag_micros->value() == 0 && metrics->async_replication_committed_lag_micros->value() == 0)) { return false; } } { // Follower metrics auto follower_metrics = cdc_service_follower->GetCDCTabletMetrics({"" / UUID */, stream_id, tablet_id}); return follower_metrics->async_replication_sent_lag_micros->value() == 0 && follower_metrics->async_replication_committed_lag_micros->value() == 0; } }, MonoDelta::FromSeconds(10), "At start, wait for Lag = 0"));
// Create the in-memory structures for both follower and leader by polling for the tablet. GetChangesRequestPB change_req; GetChangesResponsePB change_resp; change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(leader_proxy->GetChanges(change_req, &change_resp, &rpc)); change_resp.Clear(); rpc.Reset(); ASSERT_OK(follower_proxy->GetChanges(change_req, &change_resp, &rpc)); }
// Insert test rows, one at a time so they have different hybrid times. tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); { RpcController rpc; AddTestRowInsert(1, 11, "key1", &write_req); SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error()); }
{ write_req.Clear(); write_req.set_tablet_id(tablet_id); RpcController rpc; AddTestRowInsert(2, 22, "key2", &write_req); SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error()); }
// Assert that leader lag > 0. ASSERT_OK(WaitFor(& -> Result { auto metrics = cdc_service->GetCDCTabletMetrics({"" /* UUID */, stream_id, tablet_id}); return metrics->async_replication_sent_lag_micros->value() > 0 && metrics->async_replication_committed_lag_micros->value() > 0; }, MonoDelta::FromSeconds(10), "Wait for Lag > 0"));
{ // Make sure we wait for follower update thread to run at least once. SleepFor(MonoDelta::FromMilliseconds(FLAGS_update_metrics_interval_ms)); // On the follower, we shouldn't create metrics for tablets that we're not leader for, so these // should be 0 even if there are un-polled for records. auto metrics_follower = cdc_service_follower-> GetCDCTabletMetrics({"" /* UUID */, stream_id, tablet_id}); ASSERT_TRUE(metrics_follower->async_replication_sent_lag_micros->value() == 0 && metrics_follower->async_replication_committed_lag_micros->value() == 0); }
change_req.mutable_from_checkpoint()->CopyFrom(change_resp.checkpoint()); change_resp.Clear(); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(leader_proxy->GetChanges(change_req, &change_resp, &rpc)); }
// When we GetChanges the first time, only the read lag metric should be 0. ASSERT_OK(WaitFor(& -> Result { auto metrics = cdc_service->GetCDCTabletMetrics({"" /* UUID */, stream_id, tablet_id}); return metrics->async_replication_sent_lag_micros->value() == 0 && metrics->async_replication_committed_lag_micros->value() > 0; }, MonoDelta::FromSeconds(10), "Wait for Read Lag = 0"));
change_req.mutable_from_checkpoint()->CopyFrom(change_resp.checkpoint()); change_resp.Clear(); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(leader_proxy->GetChanges(change_req, &change_resp, &rpc)); }
// When we GetChanges the second time, both the lag metrics should be 0. ASSERT_OK(WaitFor(& -> Result { auto metrics = cdc_service->GetCDCTabletMetrics({"" /* UUID */, stream_id, tablet_id}); return metrics->async_replication_sent_lag_micros->value() == 0 && metrics->async_replication_committed_lag_micros->value() == 0; }, MonoDelta::FromSeconds(10), "Wait for All Lag = 0")); }
class CDCServiceTestMultipleServers : public CDCServiceTest { public: virtual int server_count() override { return 2; } virtual int tablet_count() override { return 4; } };
TEST_F_EX(CDCServiceTest, TestListTablets, CDCServiceTestMultipleServers) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
ListTabletsRequestPB req; ListTabletsResponsePB resp;
req.set_stream_id(stream_id);
auto cdc_proxy_bcast_addr = cluster_->mini_tablet_server(0)->options()->broadcast_addresses[0]; int cdc_proxy_count = 0;
// Test a simple query for all tablets. { RpcController rpc; SCOPED_TRACE(req.DebugString()); ASSERT_OK(cdc_proxy_->ListTablets(req, &resp, &rpc)); SCOPED_TRACE(resp.DebugString()); ASSERT_FALSE(resp.has_error());
ASSERT_EQ(resp.tablets_size(), tablet_count()); ASSERT_EQ(resp.tablets(0).tablet_id(), tablet_id); for (auto& tablet : resp.tablets()) { auto owner_tserver = HostPort::FromPB(tablet.tservers(0).broadcast_addresses(0)); if (owner_tserver == cdc_proxy_bcast_addr) { ++cdc_proxy_count; } }}
// Query for tablets only on the first server. We should only get a subset. { req.set_local_only(true); RpcController rpc; SCOPED_TRACE(req.DebugString()); ASSERT_OK(cdc_proxy_->ListTablets(req, &resp, &rpc)); SCOPED_TRACE(resp.DebugString()); ASSERT_FALSE(resp.has_error()); ASSERT_EQ(resp.tablets_size(), cdc_proxy_count); } }
TEST_F_EX(CDCServiceTest, TestGetChangesProxyRouting, CDCServiceTestMultipleServers) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
// Figure out [1] all tablets and [2] which ones are local to the first server. std::vectorstd::string local_tablets, all_tablets; for (bool is_local : {true, false}) { RpcController rpc; ListTabletsRequestPB req; ListTabletsResponsePB resp; req.set_stream_id(stream_id); req.set_local_only(is_local); ASSERT_OK(cdc_proxy_->ListTablets(req, &resp, &rpc)); ASSERT_FALSE(resp.has_error()); auto& cur_tablets = is_local ? local_tablets : all_tablets; for (int i = 0; i < resp.tablets_size(); ++i) { cur_tablets.push_back(resp.tablets(i).tablet_id()); } std::sort(cur_tablets.begin(), cur_tablets.end()); } ASSERT_LT(local_tablets.size(), all_tablets.size()); ASSERT_LT(0, local_tablets.size()); { // Overlap between these two lists should be all the local tablets std::vectorstd::string tablet_intersection; std::set_intersection(local_tablets.begin(), local_tablets.end(), all_tablets.begin(), all_tablets.end(), std::back_inserter(tablet_intersection)); ASSERT_TRUE(std::equal(local_tablets.begin(), local_tablets.end(), tablet_intersection.begin())); } // Difference should be all tablets on the other server. std::vectorstd::string remote_tablets; std::set_difference(all_tablets.begin(), all_tablets.end(), local_tablets.begin(), local_tablets.end(), std::back_inserter(remote_tablets)); ASSERT_LT(0, remote_tablets.size()); ASSERT_EQ(all_tablets.size() - local_tablets.size(), remote_tablets.size());
// Insert test rows, equal amount per tablet. int cur_row = 1; int to_write = 2; for (bool is_local : {true, false}) { const auto& tserver = cluster_->mini_tablet_server(is_local?0:1)->server(); // Use proxy for to most accurately simulate normal requests. const auto& proxy = tserver->proxy(); auto& cur_tablets = is_local ? local_tablets : remote_tablets; for (auto& tablet_id : cur_tablets) { tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); RpcController rpc; for (int i = 1; i <= to_write; ++i) { AddTestRowInsert(cur_row, 11 * cur_row, "key" + std::to_string(cur_row), &write_req); ++cur_row; }
SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error()); }}
// Query for all tablets on the first server. Ensure the non-local ones have errors. for (bool is_local : {true, false}) { auto& cur_tablets = is_local ? local_tablets : remote_tablets; for (auto tablet_id : cur_tablets) { std::vector proxy_options{false}; // Verify that remote tablet queries work only when proxy forwarding is enabled. if (!is_local) proxy_options.push_back(true); for (auto use_proxy : proxy_options) { GetChangesRequestPB change_req; GetChangesResponsePB change_resp; change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0); change_req.set_serve_as_proxy(use_proxy); RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); bool should_error = !(is_local || use_proxy); ASSERT_EQ(change_resp.has_error(), should_error); if (!should_error) { ASSERT_EQ(to_write, change_resp.records_size()); } } } }
// Verify the CDC metrics match what we just did. const auto& tserver = cluster_->mini_tablet_server(0)->server(); auto cdc_service = dynamic_cast<CDCServiceImpl*>( tserver->rpc_server()->service_pool("yb.cdc.CDCService")->TEST_get_service().get()); auto server_metrics = cdc_service->GetCDCServerMetrics(); ASSERT_EQ(server_metrics->cdc_rpc_proxy_count->value(), remote_tablets.size()); }
TEST_F(CDCServiceTest, TestOnlyGetLocalChanges) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
const auto& proxy = cluster_->mini_tablet_server(0)->server()->proxy();
{ // Insert local test rows. tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); RpcController rpc; AddTestRowInsert(1, 11, "key1", &write_req); AddTestRowInsert(2, 22, "key2", &write_req);
SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error());}
{ // Insert remote test rows. tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); // Apply at the lowest possible hybrid time. write_req.set_external_hybrid_time(yb::kInitialHybridTimeValue);
RpcController rpc; AddTestRowInsert(1, 11, "key1_ext", &write_req); AddTestRowInsert(3, 33, "key3_ext", &write_req); SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error());}
auto CheckChangesAndTable = & { // Get CDC changes. GetChangesRequestPB change_req; GetChangesResponsePB change_resp;
change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0); { // Make sure only the two local test rows show up. RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); ASSERT_EQ(change_resp.records_size(), 2); std::pair<int, std::string> expected_results[2] = {std::make_pair(11, "key1"), std::make_pair(22, "key2")}; for (int i = 0; i < change_resp.records_size(); i++) { ASSERT_EQ(change_resp.records(i).operation(), CDCRecordPB::WRITE); // Check the key. ASSERT_NO_FATALS(AssertIntKey(change_resp.records(i).key(), i + 1)); // Check the change records. ASSERT_NO_FATALS(AssertChangeRecords(change_resp.records(i).changes(), expected_results[i].first, expected_results[i].second)); } } // Now, fetch the entire table and ensure that we fetch all the keys inserted. client::TableHandle table; EXPECT_OK(table.Open(table_.table()->name(), client_.get())); auto result = ScanTableToStrings(table); std::sort(result.begin(), result.end()); ASSERT_EQ(3, result.size()); // Make sure that key1 and not key1_ext shows up, since we applied key1_ext at a lower hybrid // time. ASSERT_EQ("{ int32:1, int32:11, string:\"key1\" }", result[0]); ASSERT_EQ("{ int32:2, int32:22, string:\"key2\" }", result[1]); ASSERT_EQ("{ int32:3, int32:33, string:\"key3_ext\" }", result[2]);};
ASSERT_NO_FATALS(CheckChangesAndTable());
ASSERT_OK(cluster_->RestartSync());
ASSERT_OK(WaitFor(&{ std::shared_ptrtablet::TabletPeer tablet_peer; if (!cluster_->mini_tablet_server(0)->server()->tablet_manager()-> LookupTablet(tablet_id, &tablet_peer)) { return false; } return tablet_peer->LeaderStatus() == consensus::LeaderStatus::LEADER_AND_READY; }, MonoDelta::FromSeconds(30), "Wait until tablet has a leader."));
ASSERT_NO_FATALS(CheckChangesAndTable());
}
TEST_F(CDCServiceTest, TestCheckpointUpdatedForRemoteRows) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
const auto& proxy = cluster_->mini_tablet_server(0)->server()->proxy();
{ // Insert remote test rows. tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); // Apply at the lowest possible hybrid time. write_req.set_external_hybrid_time(yb::kInitialHybridTimeValue);
RpcController rpc; AddTestRowInsert(1, 11, "key1_ext", &write_req); AddTestRowInsert(3, 33, "key3_ext", &write_req); SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error());}
auto CheckChanges = & { // Get CDC changes. GetChangesRequestPB change_req; GetChangesResponsePB change_resp;
change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0); { // Make sure that checkpoint is updated even when there are no CDC records. RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); ASSERT_EQ(change_resp.records_size(), 0); ASSERT_GT(change_resp.checkpoint().op_id().index(), 0); }};
ASSERT_NO_FATALS(CheckChanges()); }
// Test to ensure that cdc_state table's checkpoint is updated as expected. // This also tests for #2897 to ensure that cdc_state table checkpoint is not overwritten to 0.0 // in case the consumer does not send from checkpoint. TEST_F(CDCServiceTest, TestCheckpointUpdate) { FLAGS_cdc_state_checkpoint_update_interval_ms = 0;
CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
const auto& proxy = cluster_->mini_tablet_server(0)->server()->proxy();
// Insert test rows. tserver::WriteRequestPB write_req; tserver::WriteResponsePB write_resp; write_req.set_tablet_id(tablet_id); { RpcController rpc; AddTestRowInsert(1, 11, "key1", &write_req); AddTestRowInsert(2, 22, "key2", &write_req);
SCOPED_TRACE(write_req.DebugString()); WriteToProxyWithRetries(proxy, write_req, &write_resp, &rpc); SCOPED_TRACE(write_resp.DebugString()); ASSERT_FALSE(write_resp.has_error());}
// Get CDC changes. GetChangesRequestPB change_req; GetChangesResponsePB change_resp;
change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_req.mutable_from_checkpoint()->mutable_op_id()->set_index(0); change_req.mutable_from_checkpoint()->mutable_op_id()->set_term(0);
{ RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); ASSERT_EQ(change_resp.records_size(), 2); }
// Call GetChanges again and pass in checkpoint that producer can mark as committed. change_req.mutable_from_checkpoint()->CopyFrom(change_resp.checkpoint()); change_resp.Clear(); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); // No more changes, so 0 records should be received. ASSERT_EQ(change_resp.records_size(), 0); }
// Verify that cdc_state table has correct checkpoint. ASSERT_NO_FATALS(VerifyCdcStateNotEmpty(client_.get()));
// Call GetChanges again but without any from checkpoint. change_req.Clear(); change_req.set_tablet_id(tablet_id); change_req.set_stream_id(stream_id); change_resp.Clear(); { RpcController rpc; SCOPED_TRACE(change_req.DebugString()); ASSERT_OK(cdc_proxy_->GetChanges(change_req, &change_resp, &rpc)); SCOPED_TRACE(change_resp.DebugString()); ASSERT_FALSE(change_resp.has_error()); // Verify that producer uses the "from_checkpoint" from cdc_state table and does not send back // any records. ASSERT_EQ(change_resp.records_size(), 0); }
// Verify that cdc_state table's checkpoint is unaffected. ASSERT_NO_FATALS(VerifyCdcStateNotEmpty(client_.get())); }
namespace { void WaitForCDCIndex(const std::shared_ptrtablet::TabletPeer& tablet_peer, int64_t expected_index, int timeout_secs) { LOG(INFO) << "Waiting until index equals " << expected_index << ". Timeout: " << timeout_secs; ASSERT_OK(WaitFor(&{ if (tablet_peer->log_available() && tablet_peer->log()->cdc_min_replicated_index() == expected_index && tablet_peer->tablet_metadata()->cdc_min_replicated_index() == expected_index) { return true; } return false; }, MonoDelta::FromSeconds(timeout_secs), "Wait until cdc min replicated index.")); LOG(INFO) << "Done waiting"; } } // namespace
class CDCServiceTestMaxRentionTime : public CDCServiceTest { public: void SetUp() override { // Immediately write any index provided by a GetChanges request to cdc_state table. FLAGS_cdc_state_checkpoint_update_interval_ms = 0; FLAGS_log_min_segments_to_retain = 1; FLAGS_log_min_seconds_to_retain = 1; FLAGS_cdc_wal_retention_time_secs = 1; FLAGS_enable_log_retention_by_op_idx = true; FLAGS_log_max_seconds_to_retain = kMaxSecondsToRetain; FLAGS_TEST_record_segments_violate_max_time_policy = true; FLAGS_update_min_cdc_indices_interval_secs = 1;
// This will rollover log segments a lot faster. FLAGS_log_segment_size_bytes = 100; CDCServiceTest::SetUp();} const int kMaxSecondsToRetain = 30; };
TEST_F_EX(CDCServiceTest, TestLogRetentionByOpId_MaxRentionTime, CDCServiceTestMaxRentionTime) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
const auto& proxy = cluster_->mini_tablet_server(0)->server()->proxy();
std::shared_ptrtablet::TabletPeer tablet_peer; ASSERT_TRUE(cluster_->mini_tablet_server(0)->server()->tablet_manager()->LookupTablet(tablet_id, &tablet_peer));
// Write a row so that the next GetChanges request doesn't fail. WriteTestRow(0, 10, "key0", tablet_id, proxy);
// Get CDC changes. GetChanges(tablet_id, stream_id, /* term / 0, / index */ 0);
WaitForCDCIndex(tablet_peer, 0, 4 * FLAGS_update_min_cdc_indices_interval_secs);
MonoTime start = MonoTime::Now(); // Write a lot more data to generate many log files that can be GCed. This should take less // than kMaxSecondsToRetain for the next check to succeed. for (int i = 1; i <= 100; i++) { WriteTestRow(i, 10 + i, "key" + std::to_string(i), tablet_id, proxy); } MonoDelta elapsed = MonoTime::Now().GetDeltaSince(start); ASSERT_LT(elapsed.ToSeconds(), kMaxSecondsToRetain); MonoDelta time_to_sleep = MonoDelta::FromSeconds(kMaxSecondsToRetain + 10) - elapsed;
// Since we haven't updated the minimum cdc index, and the elapsed time is less than // kMaxSecondsToRetain, no log files should be returned. log::SegmentSequence segment_sequence; ASSERT_OK(tablet_peer->log()->GetSegmentsToGCUnlocked(std::numeric_limits<int64_t>::max(), &segment_sequence)); ASSERT_EQ(segment_sequence.size(), 0); LOG(INFO) << "No segments to be GCed because less than " << kMaxSecondsToRetain << " seconds have elapsed";
SleepFor(time_to_sleep);
ASSERT_OK(tablet_peer->log()->GetSegmentsToGCUnlocked(std::numeric_limits<int64_t>::max(), &segment_sequence)); ASSERT_GT(segment_sequence.size(), 0); ASSERT_EQ(segment_sequence.size(), tablet_peer->log()->reader_->segments_violate_max_time_policy_->size());
for (int i = 0; i < segment_sequence.size(); i++) { ASSERT_EQ(segment_sequence[i]->path(), (*tablet_peer->log()->reader_->segments_violate_max_time_policy_)[i]->path()); LOG(INFO) << "Segment " << segment_sequence[i]->path() << " to be GCed"; } }
class CDCServiceTestDurableMinReplicatedIndex : public CDCServiceTest { public: void SetUp() override { // Immediately write any index provided by a GetChanges request to cdc_state table. FLAGS_cdc_state_checkpoint_update_interval_ms = 0; FLAGS_update_min_cdc_indices_interval_secs = 1; FLAGS_enable_log_retention_by_op_idx = true; CDCServiceTest::SetUp(); } };
TEST_F_EX(CDCServiceTest, TestLogCDCMinReplicatedIndexIsDurable, CDCServiceTestDurableMinReplicatedIndex) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
const auto& proxy = cluster_->mini_tablet_server(0)->server()->proxy();
std::shared_ptrtablet::TabletPeer tablet_peer; ASSERT_TRUE(cluster_->mini_tablet_server(0)->server()->tablet_manager()->LookupTablet(tablet_id, &tablet_peer)); // Write a row so that the next GetChanges request doesn't fail. WriteTestRow(0, 10, "key0", tablet_id, proxy);
// Get CDC changes. GetChanges(tablet_id, stream_id, /* term / 0, / index */ 10);
WaitForCDCIndex(tablet_peer, 10, 4 * FLAGS_update_min_cdc_indices_interval_secs);
// Restart the entire cluster to verify that the CDC tablet metadata got loaded from disk. ASSERT_OK(cluster_->RestartSync());
ASSERT_OK(WaitFor(&{ if (cluster_->mini_tablet_server(0)->server()->tablet_manager()-> LookupTablet(tablet_id, &tablet_peer)) { if (tablet_peer->LeaderStatus() == consensus::LeaderStatus::LEADER_AND_READY && tablet_peer->log() != nullptr) { LOG(INFO) << "TServer is ready "; return true; } } return false; }, MonoDelta::FromSeconds(30), "Wait until tablet has a leader."));
// Verify the log and meta min replicated index was loaded correctly from disk. ASSERT_EQ(tablet_peer->log()->cdc_min_replicated_index(), 10); ASSERT_EQ(tablet_peer->tablet_metadata()->cdc_min_replicated_index(), 10); }
class CDCServiceTestMinSpace : public CDCServiceTest { public: void SetUp() override { FLAGS_cdc_state_checkpoint_update_interval_ms = 0; FLAGS_log_min_segments_to_retain = 1; FLAGS_log_min_seconds_to_retain = 1; FLAGS_cdc_wal_retention_time_secs = 1; FLAGS_enable_log_retention_by_op_idx = true; // We want the logs to be GCed because of space, not because they exceeded the maximum time to // be retained. FLAGS_log_max_seconds_to_retain = 10 * 3600; // 10 hours. FLAGS_log_stop_retaining_min_disk_mb = 1; FLAGS_TEST_record_segments_violate_min_space_policy = true;
// This will rollover log segments a lot faster. FLAGS_log_segment_size_bytes = 500; CDCServiceTest::SetUp();} };
TEST_F_EX(CDCServiceTest, TestLogRetentionByOpId_MinSpace, CDCServiceTestMinSpace) { CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id);
std::string tablet_id; GetTablet(&tablet_id);
const auto& proxy = cluster_->mini_tablet_server(0)->server()->proxy();
std::shared_ptrtablet::TabletPeer tablet_peer; ASSERT_TRUE(cluster_->mini_tablet_server(0)->server()->tablet_manager()->LookupTablet(tablet_id, &tablet_peer)); // Write a row so that the next GetChanges request doesn't fail. WriteTestRow(0, 10, "key0", tablet_id, proxy);
// Get CDC changes. GetChanges(tablet_id, stream_id, /* term / 0, / index */ 0);
WaitForCDCIndex(tablet_peer, 0, 4 * FLAGS_update_min_cdc_indices_interval_secs);
// Write a lot more data to generate many log files that can be GCed. This should take less // than kMaxSecondsToRetain for the next check to succeed. for (int i = 1; i <= 5000; i++) { WriteTestRow(i, 10 + i, "key" + std::to_string(i), tablet_id, proxy); }
log::SegmentSequence segment_sequence; ASSERT_OK(tablet_peer->log()->GetSegmentsToGCUnlocked(std::numeric_limits<int64_t>::max(), &segment_sequence)); ASSERT_EQ(segment_sequence.size(), 0);
FLAGS_TEST_simulate_free_space_bytes = 128;
ASSERT_OK(tablet_peer->log()->GetSegmentsToGCUnlocked(std::numeric_limits<int64_t>::max(), &segment_sequence)); ASSERT_GT(segment_sequence.size(), 0); ASSERT_EQ(segment_sequence.size(), tablet_peer->log()->reader_->segments_violate_min_space_policy_->size());
for (int i = 0; i < segment_sequence.size(); i++) { ASSERT_EQ(segment_sequence[i]->path(), (*tablet_peer->log()->reader_->segments_violate_min_space_policy_)[i]->path()); LOG(INFO) << "Segment " << segment_sequence[i]->path() << " to be GCed"; }
int32_t num_gced(0); ASSERT_OK(tablet_peer->log()->GC(std::numeric_limits<int64_t>::max(), &num_gced)); ASSERT_EQ(num_gced, segment_sequence.size());
// Read from 0.0. This should start reading from the beginning of the logs. GetChanges(tablet_id, stream_id, /* term / 0, / index */ 0); }
class CDCLogAndMetaIndex : public CDCServiceTest { public: void SetUp() override { // Immediately write any index provided by a GetChanges request to cdc_state table. FLAGS_cdc_state_checkpoint_update_interval_ms = 0; FLAGS_update_min_cdc_indices_interval_secs = 1; FLAGS_cdc_min_replicated_index_considered_stale_secs = 5; FLAGS_enable_log_retention_by_op_idx = true; CDCServiceTest::SetUp(); } };
TEST_F_EX(CDCServiceTest, TestLogAndMetaCdcIndex, CDCLogAndMetaIndex) { constexpr int kNStreams = 5;
// This will rollover log segments a lot faster. FLAGS_log_segment_size_bytes = 100;
CDCStreamId stream_id[kNStreams];
for (int i = 0; i < kNStreams; i++) { CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id[i]); }
std::string tablet_id; GetTablet(&tablet_id);
const auto &proxy = cluster_->mini_tablet_server(0)->server()->proxy();
// Insert test rows. for (int i = 1; i <= kNStreams; i++) { WriteTestRow(i, 10 + i, "key" + std::to_string(i), tablet_id, proxy); }
std::shared_ptrtablet::TabletPeer tablet_peer; ASSERT_TRUE(cluster_->mini_tablet_server(0)->server()->tablet_manager()->LookupTablet(tablet_id, &tablet_peer));
// Before any cdc request, the min index should be max value. ASSERT_EQ(tablet_peer->log()->cdc_min_replicated_index(), std::numeric_limits<int64_t>::max()); ASSERT_EQ(tablet_peer->tablet_metadata()->cdc_min_replicated_index(), std::numeric_limits<int64_t>::max());
for (int i = 0; i < kNStreams; i++) { // Get CDC changes. GetChanges(tablet_id, stream_id[i], /* term / 0, / index */ i); }
// After the request succeeded, verify that the min cdc limit was set correctly. In this case // it belongs to stream_id[0] with index 0. WaitForCDCIndex(tablet_peer, 0, 4 * FLAGS_update_min_cdc_indices_interval_secs);
// Changing the lowest index from all the streams should also be reflected in the log object. GetChanges(tablet_id, stream_id[0], /* term / 0, / index */ 4);
// After the request succeeded, verify that the min cdc limit was set correctly. In this case // it belongs to stream_id[1] with index 1. WaitForCDCIndex(tablet_peer, 1, 4 * FLAGS_update_min_cdc_indices_interval_secs); }
class CDCLogAndMetaIndexReset : public CDCLogAndMetaIndex { public: void SetUp() override { FLAGS_cdc_min_replicated_index_considered_stale_secs = 5; // This will rollover log segments a lot faster. FLAGS_log_segment_size_bytes = 100; CDCLogAndMetaIndex::SetUp(); } };
// Test that when all the streams for a specific tablet have been deleted, the log and meta // cdc min replicated index is reset to max int64. TEST_F_EX(CDCServiceTest, TestLogAndMetaCdcIndexAreReset, CDCLogAndMetaIndexReset) { constexpr int kNStreams = 5;
// This will rollover log segments a lot faster. FLAGS_log_segment_size_bytes = 100;
CDCStreamId stream_id[kNStreams];
for (int i = 0; i < kNStreams; i++) { CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id[i]); }
std::string tablet_id; GetTablet(&tablet_id);
const auto &proxy = cluster_->mini_tablet_server(0)->server()->proxy();
// Insert test rows. for (int i = 1; i <= kNStreams; i++) { WriteTestRow(i, 10 + i, "key" + std::to_string(i), tablet_id, proxy); }
std::shared_ptrtablet::TabletPeer tablet_peer; ASSERT_TRUE(cluster_->mini_tablet_server(0)->server()->tablet_manager()->LookupTablet(tablet_id, &tablet_peer));
// Before any cdc request, the min index should be max value. ASSERT_EQ(tablet_peer->log()->cdc_min_replicated_index(), std::numeric_limits<int64_t>::max()); ASSERT_EQ(tablet_peer->tablet_metadata()->cdc_min_replicated_index(), std::numeric_limits<int64_t>::max());
for (int i = 0; i < kNStreams; i++) { // Get CDC changes. GetChanges(tablet_id, stream_id[i], /* term / 0, / index */ 5); }
// After the request succeeded, verify that the min cdc limit was set correctly. In this case // all the streams have index 5. WaitForCDCIndex(tablet_peer, 5, 4 * FLAGS_update_min_cdc_indices_interval_secs);
client::TableHandle table; client::YBTableName cdc_state_table( YQL_DATABASE_CQL, master::kSystemNamespaceName, master::kCdcStateTableName); ASSERT_OK(table.Open(cdc_state_table, client_.get()));
auto session = client_->NewSession(); for (int i = 0; i < kNStreams; i++) { const auto delete_op = table.NewDeleteOp(); auto* delete_req = delete_op->mutable_request(); QLAddStringHashValue(delete_req, tablet_id); QLAddStringRangeValue(delete_req, stream_id[i]); ASSERT_OK(session->Apply(delete_op)); } ASSERT_OK(session->Flush()); LOG(INFO) << "Successfully deleted all streams from cdc_state";
SleepFor(MonoDelta::FromSeconds(FLAGS_cdc_min_replicated_index_considered_stale_secs + 1));
LOG(INFO) << "Done sleeping"; // RunLogGC should reset cdc min replicated index to max int64 because more than // FLAGS_cdc_min_replicated_index_considered_stale_secs seconds have elapsed since the index // was last updated. ASSERT_OK(tablet_peer->RunLogGC()); LOG(INFO) << "GC done running"; ASSERT_EQ(tablet_peer->log()->cdc_min_replicated_index(), std::numeric_limits<int64_t>::max()); ASSERT_EQ(tablet_peer->tablet_metadata()->cdc_min_replicated_index(), std::numeric_limits<int64_t>::max()); }
class CDCServiceTestThreeServers : public CDCServiceTest { public: void SetUp() override { // We don't want the tablets to move in the middle of the test. FLAGS_enable_load_balancing = false; FLAGS_leader_failure_max_missed_heartbeat_periods = 12.0; FLAGS_update_min_cdc_indices_interval_secs = 5; FLAGS_enable_log_retention_by_op_idx = true; FLAGS_client_read_write_timeout_ms = 20 * 1000 * kTimeMultiplier;
// Always update cdc_state table. FLAGS_cdc_state_checkpoint_update_interval_ms = 0; FLAGS_follower_unavailable_considered_failed_sec = 20 * kTimeMultiplier; CDCServiceTest::SetUp();}
void DoTearDown() override { YBMiniClusterTestBase::DoTearDown(); }
virtual int server_count() override { return 3; } virtual int tablet_count() override { return 3; }
// Get the first tablet_id for which any peer is a leader. void GetFirstTabletIdAndLeaderPeer(TabletId* tablet_id, int* leader_idx, int timeout_secs); };
// Sometimes leadership takes a while. Keep retrying until timeout_secs seconds have elapsed. void CDCServiceTestThreeServers::GetFirstTabletIdAndLeaderPeer(TabletId* tablet_id, int* leader_idx, int timeout_secs) { std::vector tablet_ids; // Verify that we are only returning a tablet that belongs to the table created for this test. GetTablets(&tablet_ids); ASSERT_EQ(tablet_ids.size(), tablet_count());
MonoTime now = MonoTime::Now(); MonoTime deadline = now + MonoDelta::FromSeconds(timeout_secs); while(now.ComesBefore(deadline) && (!tablet_id || tablet_id->empty())) { for (int idx = 0; idx < cluster_->num_tablet_servers(); idx++) { auto peers = cluster_->mini_tablet_server(idx)->server()->tablet_manager()->GetTabletPeers(); ASSERT_GT(peers.size(), 0);
for (const auto &peer : peers) { auto it = std::find(tablet_ids.begin(), tablet_ids.end(), peer->tablet_id()); if (it != tablet_ids.end() && peer->LeaderStatus() == consensus::LeaderStatus::LEADER_AND_READY) { *tablet_id = peer->tablet_id(); *leader_idx = idx; LOG(INFO) << "Selected tablet " << tablet_id << " for tablet server " << idx; break; } } } now = MonoTime::Now();} }
// Test that whenever a leader change happens (forced here by shutting down the tablet leader), // next leader correctly reads the minimum applied cdc index by reading the cdc_state table. TEST_F_EX(CDCServiceTest, TestNewLeaderUpdatesLogCDCAppliedIndex, CDCServiceTestThreeServers) { constexpr int kNRecords = 30; constexpr int kGettingLeaderTimeoutSecs = 20;
TabletId tablet_id; // Index of the TS that is the leader for the selected tablet_id. int leader_idx = -1;
GetFirstTabletIdAndLeaderPeer(&tablet_id, &leader_idx, kGettingLeaderTimeoutSecs); ASSERT_FALSE(tablet_id.empty()); ASSERT_GE(leader_idx, 0);
const auto &proxy = cluster_->mini_tablet_server(leader_idx)->server()->proxy(); for (int i = 0; i < kNRecords; i++) { WriteTestRow(i, 10 + i, "key" + std::to_string(i), tablet_id, proxy); } LOG(INFO) << "Inserted " << kNRecords << " records";
CDCStreamId stream_id; CreateCDCStream(cdc_proxy_, table_.table()->id(), &stream_id); LOG(INFO) << "Created cdc stream " << stream_id;
GetChanges(tablet_id, stream_id, /* term / 0, / index */ 5); LOG(INFO) << "GetChanges request completed successfully";
std::shared_ptrtablet::TabletPeer tablet_peer; // Check that the index hasn't been updated in any of the followers. for (int idx = 0; idx < server_count(); idx++) { if (idx == leader_idx) { // This TServer is shutdown for now. continue; }
if (cluster_->mini_tablet_server(idx)->server()->tablet_manager()-> LookupTablet(tablet_id, &tablet_peer)) { ASSERT_EQ(tablet_peer->log()->cdc_min_replicated_index(), std::numeric_limits<int64>::max()); ASSERT_EQ(tablet_peer->tablet_metadata()->cdc_min_replicated_index(), std::numeric_limits<int64>::max()); }}
// Kill the tablet leader tserver so that another tserver becomes the leader. cluster_->mini_tablet_server(leader_idx)->Shutdown(); LOG(INFO) << "tserver " << leader_idx << " was shutdown";
// CDC Proxy is pinned to the first TServer, so we need to update the proxy if we kill that one. if (leader_idx == 0) { cdc_proxy_ = std::make_unique( &client_->proxy_cache(), HostPort::FromBoundEndpoint(cluster_->mini_tablet_server(1)->bound_rpc_addr())); }
// Wait until GetChanges doesn't return any errors. This means that we are able to write to // the cdc_state table. ASSERT_OK(WaitFor(&{ bool has_error = false; GetChanges(tablet_id, stream_id, /* term / 0, / index */ 5, &has_error); return !has_error; }, MonoDelta::FromSeconds(180), "Wait until cdc state table can take writes."));
SleepFor(MonoDelta::FromSeconds((FLAGS_update_min_cdc_indices_interval_secs * 3))); LOG(INFO) << "Done sleeping";
std::unique_ptr cdc_proxy; ASSERT_OK(WaitFor(&{ for (int idx = 0; idx < server_count(); idx++) { if (idx == leader_idx) { // This TServer is shutdown for now. continue; } if (cluster_->mini_tablet_server(idx)->server()->tablet_manager()-> LookupTablet(tablet_id, &tablet_peer)) { if (tablet_peer->LeaderStatus() == consensus::LeaderStatus::LEADER_AND_READY) { LOG(INFO) << "Found new leader for tablet " << tablet_id << " in TS " << idx; return true; } } } return false; }, MonoDelta::FromSeconds(30), "Wait until tablet has a leader."));
ASSERT_EQ(tablet_peer->log()->cdc_min_replicated_index(), 5); ASSERT_EQ(tablet_peer->tablet_metadata()->cdc_min_replicated_index(), 5);
ASSERT_OK(cluster_->mini_tablet_server(leader_idx)->Start()); ASSERT_OK(cluster_->mini_tablet_server(leader_idx)->WaitStarted()); }
} // namespace cdc } // namespace yb
-
- Assistant
Identifying System 1 and System 2 Thinking Requirements
Firstly, I'll read through the test code to identify sections that embody quick, direct responses or straightforward checks which are indicative of System 1 thinking, as well as sections that necessitate in-depth analysis or careful logical breakdowns, which reflect System 2 thinking.
-
System 1 Sections are characterized by immediate checks, validations, or direct responses to conditions in the test set-up and assertions.
-
System 2 Sections involve setting up test scenarios, analyzing outputs over time, and interpreting results from various logical or creative angles.
Processed Output with System 1 and System 2 Thinking
Class Setup
[sys2] The
CDCServiceTestThreeServersclass initializes a test environment with multiple servers and tablets. It sets up configuration flags to manage leader failure and client timeout settings.2.1 Understanding the Problem:
- What I understand: The setup configures a 3-node cluster to simulate scenarios with changes to the CDC (Change Data Capture) service.
- Potential things overlooked: How settings like
FLAGS_client_read_write_timeout_msaffect different network conditions.
Creative Perspective:
- Scenarios beyond leader failure could be explored, like network partitioning between nodes or data center outages.
2.2 Analysis:
- Each setting ensures improved control over the system under varied testing conditions.
- Ensure all nodes have a consistent configuration.
Creative Perspective:
- Exploring different parameter values could unlock unconventional reactions, like high-availability faults or load fluctuations.
[end sys2]
Selecting and Managing Leaders
[sys1]
- Fetch the leader tablet from multiple servers.
- Use assertions to verify the leader status of each node. [end sys1]
[sys2] 2.3 Generate Hypotheses:
- Leader index is affected by node restart delays. (Confidence: 0.6, Creative: 0.4)
- Data distribution might be more efficient with equal load among nodes. (Confidence: 0.5, Creative: 0.6)
- Current settings could fail if data size exceeds typical log segment sizes. (Confidence: 0.7, Creative: 0.8)
- Heartbeat timeouts might lead to secondary node promotions. (Confidence: 0.6, Creative: 0.3)
- Longer CDC checkpoints might stall when logs exceed expected size. (Confidence: 0.7, Creative: 0.5)
- Consistent leadership leads to better performance outcomes. (Confidence: 0.8, Creative: 0.3)
- A node's leader status may correlate with geographic proximity impacts. (Confidence: 0.4, Creative: 0.7)
- Random node crashes impact leader elections unpredictably. (Confidence: 0.7, Creative: 0.9)
- Balanced load could mitigate leader change effects. (Confidence: 0.8, Creative: 0.5)
- Network partitions might result in isolated leader elections. (Confidence: 0.5, Creative: 0.8)
2.4 Future Steps:
- Monitor the impact of configuration changes over multiple leader election regimes.
Creative Perspective:
- Simulate scenarios of staggered node reboots and leader change sequences.
2.5 Hypotheses Evaluation:
- Check for primary stability under repetitive election scenarios.
- Prioritize by confidence and implication scale; unexpected leader choices warrant additional scrutiny.
[end sys2]
Execute Test Scenarios
[sys1]
- Execute test procedures to validate log entry distribution and update mechanisms across nodes.
- Verify whether the restart of nodes can impact failover logic or data replication functionality. [end sys1]
Reflection and Insights
[sys2] 2.9 Reflect and Capture Insights:
- What was learned: Stabilizing leader changes through configuration control managed consistent log updating processes.
- New understanding: Geo-distributed nodes under different loads offer unique leader election behaviors.
- Insights application: Useful in high-availability systems, especially those requiring synchronized log replication post-leader failure.
Creative Perspective:
- Implement varying network delay effects into tests to enhance real-world simulation complexity. [end sys2]
The above synthetic run-through demonstrates the breakdown of a technical test setup combining System 1 direct check commands with the System 2 in-depth and creative thinking processes, carefully considering possible expansions and variations in hypotheses relevant to real-world applications.
-