mongodb/mongo
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

SERVER-114992 Enable more tests and support WCOS metrics in UWE (#45617)

Browse Source 
GitOrigin-RevId: ad14164d4e7cc495d3b5b67b680289fd996e3eda
This commit is contained in:
Mihai Andrei 2025-12-24 11:54:07 -05:00 committed by MongoDB Bot
parent 4295703db1
commit 836d6223d4
14 changed files with 111 additions and 66 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
jstests/libs/bulk_write_utils.js
View File

@ -1,3 +1,5 @@
import {isUweEnabled} from "jstests/libs/query/uwe_utils.js";
/**
* Helper function to check a BulkWrite cursorEntry.
*/
@ -387,7 +389,9 @@ export class BulkWriteMetricChecker {
targetedInsert *= retryCount;
}
if (this.timeseries && !this.bulkWrite && updateShardField === "manyShards") {
// Note that in the case of the UWE, we do not double count retryable timeseries updates.
const uweEnabled = isUweEnabled(this.testDB);
if (this.timeseries && !this.bulkWrite && !uweEnabled && updateShardField === "manyShards") {
targetedUpdate = targetedUpdate * 2;
}

6
jstests/noPassthrough/crud/bulk_write_metrics.js
View File

@ -6,7 +6,6 @@
import {getTimeseriesCollForDDLOps} from "jstests/core/timeseries/libs/viewless_timeseries_util.js";
import {BulkWriteMetricChecker} from "jstests/libs/bulk_write_utils.js";
import {isUweEnabled} from "jstests/libs/query/uwe_utils.js";
import {ReplSetTest} from "jstests/libs/replsettest.js";
import {ShardingTest} from "jstests/libs/shardingtest.js";
@ -225,10 +224,7 @@ function runTest(isMongos, cluster, bulkWrite, retryCount, timeseries) {
},
);
// TODO SERVER-114992: Not running in master because of tag but passes with flag off and fails
// with flag on.
const uweEnabled = isUweEnabled(testDB);
if (isMongos && !uweEnabled) {
if (isMongos) {
// Update modifying owning shard requires a transaction or retryable write, we do not want
// actual retries here.
metricChecker.retryCount = 1;

11
jstests/sharding/analyze_shard_key/sample_write_queries_sharded.js
View File

@ -6,7 +6,6 @@
import {withTxnAndAutoRetryOnMongos} from "jstests/libs/auto_retry_transaction_in_sharding.js";
import {ShardingTest} from "jstests/libs/shardingtest.js";
import {QuerySamplingUtil} from "jstests/sharding/analyze_shard_key/libs/query_sampling_util.js";
import {isUweEnabled} from "jstests/libs/query/uwe_utils.js";
// Make the periodic jobs for refreshing sample rates and writing sampled queries and diffs have a
// period of 1 second to speed up the test.
@ -25,8 +24,6 @@ const st = new ShardingTest({
mongosOptions: {setParameter: {queryAnalysisSamplerConfigurationRefreshSecs}},
});
const uweEnabled = isUweEnabled(st.s);
const dbName = "testDb";
const collName = "testColl";
const ns = dbName + "." + collName;
@ -154,10 +151,6 @@ const expectedSampledQueryDocs = [];
}),
);
// TODO SERVER-114992: failed to find sampled query.
if (uweEnabled) {
return;
}
// This is a WouldChangeOwningShard update. It causes the document to move from shard0 to
// shard1.
const singleUpdateOp2 = {
@ -297,10 +290,6 @@ const expectedSampledQueryDocs = [];
mongosDB.runCommand({explain: {findAndModify: collName, query: {x: 501}, update: {$set: {y: 501}}}}),
);
// TODO SERVER-114992: failed to find sampled query.
if (uweEnabled) {
return;
}
// This is a WouldChangeOwningShard update. It causes the document to move from shard1 to
// shard2.
const originalCmdObj1 = {

11
jstests/sharding/analyze_shard_key/shard_key_updates.js
View File

@ -9,7 +9,6 @@ import {ShardingTest} from "jstests/libs/shardingtest.js";
import {extractUUIDFromObject} from "jstests/libs/uuid_util.js";
import {AnalyzeShardKeyUtil} from "jstests/sharding/analyze_shard_key/libs/analyze_shard_key_util.js";
import {QuerySamplingUtil} from "jstests/sharding/analyze_shard_key/libs/query_sampling_util.js";
import {isUweEnabled} from "jstests/libs/query/uwe_utils.js";
// This command involves running commands outside a session.
TestData.disableImplicitSessions = true;
@ -40,16 +39,6 @@ const st = new ShardingTest({
mongosOptions: {setParameter: {queryAnalysisSamplerConfigurationRefreshSecs}},
});
// TODO SERVER-114992: failed to find sampled query.
let uweEnabled = false;
st.forEachConnection((conn) => {
uweEnabled = uweEnabled || isUweEnabled(conn);
});
if (uweEnabled) {
st.stop();
quit();
}
const execCtxTypes = {
kNoClientSession: 1,
kClientSessionNotRetryableWrite: 2,

9
jstests/sharding/query/compound_hashed_shard_key_targeting.js
View File

@ -16,7 +16,6 @@ import {
import {assertStagesForExplainOfCommand} from "jstests/libs/query/analyze_plan.js";
import {ShardingTest} from "jstests/libs/shardingtest.js";
import {findChunksUtil} from "jstests/sharding/libs/find_chunks_util.js";
import {isUweEnabled} from "jstests/libs/query/uwe_utils.js";
const st = new ShardingTest({shards: 2});
const kDbName = jsTestName();
@ -29,7 +28,6 @@ const sessionDB = session.getDatabase(kDbName);
const coll = sessionDB["coll"];
const shard0DB = st.shard0.getDB(kDbName);
const shard1DB = st.shard1.getDB(kDbName);
const uweEnabled = isUweEnabled(st.s);
/**
* Enables profiling on both shards so that we can verify the targeting behaviour.
@ -174,11 +172,8 @@ res = assert.commandWorked(coll.update({a: 22, b: {subObj: "str_0"}, c: "update"
assert.eq(res.nModified, 1, res);
// Verify that the 'update' command gets targeted to 'shard1DB'.
// TODO SERVER-114992: couldn't find matching bulkWrite.
if (!uweEnabled) {
profilerHasAtLeastOneMatchingEntryOrThrow({profileDB: shard1DB, filter: {ns: ns, "op": shardCmdName}});
profilerHasZeroMatchingEntriesOrThrow({profileDB: shard0DB, filter: {ns: ns, "op": shardCmdName}});
}
profilerHasAtLeastOneMatchingEntryOrThrow({profileDB: shard1DB, filter: {ns: ns, "op": shardCmdName}});
profilerHasZeroMatchingEntriesOrThrow({profileDB: shard0DB, filter: {ns: ns, "op": shardCmdName}});
// Verify that the 'count' command gets targeted to 'shard0DB' after the update.
assert.eq(coll.count(updateObj["$set"]), 1);

33
src/mongo/s/cluster_write.cpp
View File

@ -38,6 +38,7 @@
#include "mongo/logv2/log.h"
#include "mongo/s/write_ops/bulk_write_exec.h"
#include "mongo/s/write_ops/fle.h"
#include "mongo/s/write_ops/unified_write_executor/stats.h"
#include "mongo/s/write_ops/unified_write_executor/unified_write_executor.h"
#include "mongo/util/decorable.h"
@ -48,6 +49,28 @@
#define MONGO_LOGV2_DEFAULT_COMPONENT ::mongo::logv2::LogComponent::kSharding
namespace mongo {
namespace {
// Utility to port 'uweStats' to 'batchStats'. This is done to enable re-using the existing logic
// for cluster write statistics.
// TODO SERVER-109104: This can be removed once we delete BatchWriteExec.
void fillOutBatchWriteStats(unified_write_executor::Stats& uweStats,
BatchWriteExecStats& batchStats) {
const auto& map = uweStats.getTargetingStatsMap();
if (map.empty()) {
return;
}
tassert(11499200,
"Expected only a single entry in targeting stats map for a batch write",
map.size() == 1 && map.count(0) != 0);
const auto& targetingStats = map.at(0);
batchStats.noteNumShardsOwningChunks(targetingStats.numShardsOwningChunks);
batchStats.noteTargetedCollectionIsSharded(targetingStats.isSharded);
for (const auto& shard : uweStats.getTargetedShards()) {
batchStats.noteTargetedShard(shard);
}
}
} // namespace
namespace cluster {
void write(OperationContext* opCtx,
@ -59,6 +82,7 @@ void write(OperationContext* opCtx,
NotPrimaryErrorTracker::Disabled scopeDisabledTracker(
&NotPrimaryErrorTracker::get(opCtx->getClient()));
tassert(11499201, "Should have initialized 'stats' object", stats);
CollectionRoutingInfoTargeter targeter(opCtx, request.getNS(), targetEpoch);
if (nss) {
*nss = targeter.getNS();
@ -68,9 +92,10 @@ void write(OperationContext* opCtx,
4817400, 2, {logv2::LogComponent::kShardMigrationPerf}, "Starting batch write");
if (unified_write_executor::isEnabled(opCtx)) {
*response = unified_write_executor::write(opCtx, request, targetEpoch);
unified_write_executor::Stats uweStats;
*response = unified_write_executor::write(opCtx, request, uweStats, targetEpoch);
// SERVER-109104 This can be removed once we delete BatchWriteExec.
stats->markIgnore();
fillOutBatchWriteStats(uweStats, *stats);
} else {
// Create an RAII object that prints the collection's shard key in the case of a tassert
// or crash.
@ -102,8 +127,8 @@ bulk_write_exec::BulkWriteReplyInfo bulkWrite(
const std::vector<std::unique_ptr<NSTargeter>>& targeters,
bulk_write_exec::BulkWriteExecStats& execStats) {
if (unified_write_executor::isEnabled(opCtx)) {
execStats.markIgnore();
return unified_write_executor::bulkWrite(opCtx, request);
unified_write_executor::Stats uweStats;
return unified_write_executor::bulkWrite(opCtx, request, uweStats);
} else {
if (request.getNsInfo()[0].getEncryptionInformation().has_value()) {
auto [result, replies] = attemptExecuteFLE(opCtx, request);

11
src/mongo/s/commands/query_cmd/cluster_bulk_write_cmd.h
View File

@ -74,6 +74,7 @@
#include "mongo/s/query/exec/router_stage_queued_data.h"
#include "mongo/s/would_change_owning_shard_exception.h"
#include "mongo/s/write_ops/batched_command_request.h"
#include "mongo/s/write_ops/unified_write_executor/stats.h"
#include "mongo/s/write_ops/unified_write_executor/unified_write_executor.h"
#include "mongo/util/assert_util.h"
#include "mongo/util/decorable.h"
@ -254,8 +255,10 @@ public:
bulk_write_exec::BulkWriteExecStats execStats;
bulk_write_exec::BulkWriteReplyInfo replyInfo;
unified_write_executor::Stats uweStats;
if (unified_write_executor::isEnabled(opCtx)) {
replyInfo = unified_write_executor::bulkWrite(opCtx, bulkRequest);
replyInfo = unified_write_executor::bulkWrite(opCtx, bulkRequest, uweStats);
execStats.markIgnore();
} else {
// This is used only for the ScopedDebugInfo construction below.
stdx::unordered_map<NamespaceString, boost::optional<BSONObj>>
@ -306,7 +309,11 @@ public:
handleWouldChangeOwningShardError(opCtx, bulkRequest, replyInfo, targeters);
// TODO SERVER-83869 handle BulkWriteExecStats for batches of size > 1 containing
// updates that modify a document’s owning shard.
execStats.updateMetrics(opCtx, targeters, updatedShardKey);
if (!execStats.getIgnore()) {
execStats.updateMetrics(opCtx, targeters, updatedShardKey);
} else {
uweStats.updateMetrics(opCtx, updatedShardKey);
}
response = populateCursorReply(opCtx, bulkRequest, request.body, std::move(replyInfo));

28
src/mongo/s/write_ops/unified_write_executor/stats.cpp
View File

@ -59,21 +59,27 @@ void Stats::recordTargetingStats(const std::vector<ShardEndpoint>& targetedShard
}
}
void Stats::updateMetrics(OperationContext* opCtx) {
void Stats::updateMetrics(OperationContext* opCtx, bool updatedShardKey) {
// Record the number of shards targeted by this write.
// TODO SERVER-114992 increment 'nShards' by 1 if we've targeted shards and updated the shard
// key.
CurOp::get(opCtx)->debug().nShards = _targetedShards.size();
for (const auto& [nsIdx, targetingStats] : _targetingStatsMap) {
const bool isSharded = targetingStats.isSharded;
const int nShardsOwningChunks = targetingStats.numShardsOwningChunks;
for (const auto& [writeType, shards] : targetingStats.targetedShardsByWriteType) {
const int perWriteNShards = shards.size();
if (nShardsOwningChunks == 0) {
continue;
}
for (const auto& [writeType, shards] : targetingStats.targetedShardsByWriteType) {
int perWriteNShards = shards.size();
// If we have no information on the shards targeted, ignore updatedShardKey,
// updateHostsTargetedMetrics will report this as TargetType::kManyShards.
if (perWriteNShards != 0 && updatedShardKey) {
perWriteNShards += 1;
}
// TODO SERVER-114992: add one to 'nShards' if updated shard key. This information is
// returned from WCOS handling, see batch_write_exec.cpp
NumHostsTargetedMetrics::QueryType metricsWriteType;
switch (writeType) {
case WriteType::kInsert:
@ -95,12 +101,16 @@ void Stats::updateMetrics(OperationContext* opCtx) {
}
}
void Stats::incrementOpCounters(OperationContext* opCtx, WriteCommandRef::OpRef op) {
void Stats::incrementOpCounters(OperationContext* opCtx,
WriteCommandRef::OpRef op,
bool statusOkOrNotWCOS) {
if (op.isInsertOp()) {
serviceOpCounters(ClusterRole::RouterServer).gotInsert();
} else if (op.isUpdateOp()) {
serviceOpCounters(ClusterRole::RouterServer).gotUpdate();
if (statusOkOrNotWCOS) {
serviceOpCounters(ClusterRole::RouterServer).gotUpdate();
}
auto updateRef = op.getUpdateOp();
// 'isMulti' is set to false as the metrics for multi updates were registered

14
src/mongo/s/write_ops/unified_write_executor/stats.h
View File

@ -67,12 +67,22 @@ public:
/**
* Method to update CurOp and NumHostsTargetedMetrics.
*/
void updateMetrics(OperationContext* opCtx);
void updateMetrics(OperationContext* opCtx, bool updatedShardKey);
/**
* Helper to increment query counters for 'op'.
*/
void incrementOpCounters(OperationContext* opCtx, WriteCommandRef::OpRef op);
void incrementOpCounters(OperationContext* opCtx,
WriteCommandRef::OpRef op,
bool statusOkOrNotWCOS);
const stdx::unordered_map<size_t, TargetingStats>& getTargetingStatsMap() {
return _targetingStatsMap;
}
const stdx::unordered_set<ShardId>& getTargetedShards() {
return _targetedShards;
}
private:
stdx::unordered_map<size_t, TargetingStats> _targetingStatsMap;

13
src/mongo/s/write_ops/unified_write_executor/unified_write_executor.cpp
View File

@ -69,6 +69,7 @@ bool isNonVerboseWriteCommand(OperationContext* opCtx, WriteCommandRef cmdRef) {
WriteCommandResponse executeWriteCommand(OperationContext* opCtx,
WriteCommandRef cmdRef,
unified_write_executor::Stats& stats,
BSONObj originalCommand,
boost::optional<OID> targetEpoch) {
@ -76,7 +77,6 @@ WriteCommandResponse executeWriteCommand(OperationContext* opCtx,
const bool isNonVerbose = isNonVerboseWriteCommand(opCtx, cmdRef);
Stats stats;
WriteOpProducer producer(cmdRef);
WriteOpAnalyzerImpl analyzer = WriteOpAnalyzerImpl(stats);
@ -94,12 +94,12 @@ WriteCommandResponse executeWriteCommand(OperationContext* opCtx,
WriteBatchScheduler scheduler(cmdRef, *batcher, executor, processor, targetEpoch);
scheduler.run(opCtx);
stats.updateMetrics(opCtx);
return processor.generateClientResponse(opCtx);
}
BatchedCommandResponse write(OperationContext* opCtx,
const BatchedCommandRequest& request,
unified_write_executor::Stats& stats,
boost::optional<OID> targetEpoch) {
if (request.hasEncryptionInformation()) {
BatchedCommandResponse response;
@ -112,11 +112,12 @@ BatchedCommandResponse write(OperationContext* opCtx,
}
return std::get<BatchedCommandResponse>(
executeWriteCommand(opCtx, WriteCommandRef{request}, BSONObj(), targetEpoch));
executeWriteCommand(opCtx, WriteCommandRef{request}, stats, BSONObj(), targetEpoch));
}
bulk_write_exec::BulkWriteReplyInfo bulkWrite(OperationContext* opCtx,
const BulkWriteCommandRequest& request,
unified_write_executor::Stats& stats,
BSONObj originalCommand) {
if (request.getNsInfo()[0].getEncryptionInformation().has_value()) {
auto [result, replyInfo] = attemptExecuteFLE(opCtx, request);
@ -128,7 +129,7 @@ bulk_write_exec::BulkWriteReplyInfo bulkWrite(OperationContext* opCtx,
}
return std::get<bulk_write_exec::BulkWriteReplyInfo>(
executeWriteCommand(opCtx, WriteCommandRef{request}, originalCommand));
executeWriteCommand(opCtx, WriteCommandRef{request}, stats, originalCommand));
}
// TODO(SERVER-115515) Clean up FAM code in UWE (here and across files in
@ -157,8 +158,10 @@ FindAndModifyCommandResponse findAndModify(
"Cannot specify runtime constants option to a mongos",
request.getLegacyRuntimeConstants() == boost::none);
request.setLegacyRuntimeConstants(Variables::generateRuntimeConstants(opCtx));
unified_write_executor::Stats stats;
return std::get<FindAndModifyCommandResponse>(
executeWriteCommand(opCtx, WriteCommandRef{request}, originalCommand));
executeWriteCommand(opCtx, WriteCommandRef{request}, stats, originalCommand));
}
bool isEnabled(OperationContext* opCtx) {

4
src/mongo/s/write_ops/unified_write_executor/unified_write_executor.h
View File

@ -34,6 +34,7 @@
#include "mongo/s/write_ops/batched_command_request.h"
#include "mongo/s/write_ops/batched_command_response.h"
#include "mongo/s/write_ops/bulk_write_reply_info.h"
#include "mongo/s/write_ops/unified_write_executor/stats.h"
#include "mongo/s/write_ops/write_command_ref.h"
#include "mongo/util/modules.h"
@ -53,6 +54,7 @@ using WriteCommandResponse = std::variant<BatchedCommandResponse,
*/
WriteCommandResponse executeWriteCommand(OperationContext* opCtx,
WriteCommandRef cmdRef,
unified_write_executor::Stats& stats,
BSONObj originalCommand = BSONObj(),
boost::optional<OID> targetEpoch = boost::none);
@ -61,6 +63,7 @@ WriteCommandResponse executeWriteCommand(OperationContext* opCtx,
*/
BatchedCommandResponse write(OperationContext* opCtx,
const BatchedCommandRequest& request,
unified_write_executor::Stats& stats,
boost::optional<OID> targetEpoch = boost::none);
/**
@ -68,6 +71,7 @@ BatchedCommandResponse write(OperationContext* opCtx,
*/
bulk_write_exec::BulkWriteReplyInfo bulkWrite(OperationContext* opCtx,
const BulkWriteCommandRequest& request,
unified_write_executor::Stats& stats,
BSONObj originalCommand = BSONObj());
/**

21
src/mongo/s/write_ops/unified_write_executor/unified_write_executor_test.cpp
View File

@ -206,7 +206,8 @@ TEST_F(UnifiedWriteExecutorTest, BulkWriteBasic) {
{NamespaceInfoEntry(nss1), NamespaceInfoEntry(nss2)});
auto future = launchAsync([&]() {
auto replyInfo = bulkWrite(operationContext(), request);
Stats uweStats;
auto replyInfo = bulkWrite(operationContext(), request, uweStats);
auto reply = populateCursorReply(operationContext(), request, request.toBSON(), replyInfo);
auto replyItems = reply.getCursor().getFirstBatch();
ASSERT_EQ(replyItems.size(), 2);
@ -272,7 +273,8 @@ TEST_F(UnifiedWriteExecutorTest, BatchWriteBasic) {
auto future = launchAsync([&]() {
auto resp = write(operationContext(), insertRequest);
Stats uweStats;
auto resp = write(operationContext(), insertRequest, uweStats);
ASSERT(resp.getOk());
ASSERT_FALSE(resp.isErrDetailsSet());
ASSERT_EQ(resp.getN(), 2);
@ -308,7 +310,8 @@ TEST_F(UnifiedWriteExecutorTest, BatchWriteBasic) {
}());
auto updateFuture = launchAsync([&]() {
auto resp = write(operationContext(), updateRequest);
Stats uweStats;
auto resp = write(operationContext(), updateRequest, uweStats);
ASSERT(resp.getOk());
ASSERT_FALSE(resp.isErrDetailsSet());
ASSERT_EQ(resp.getN(), 1);
@ -339,7 +342,8 @@ TEST_F(UnifiedWriteExecutorTest, BatchWriteBasic) {
}());
auto deleteFuture = launchAsync([&]() {
auto resp = write(operationContext(), deleteRequest);
Stats uweStats;
auto resp = write(operationContext(), deleteRequest, uweStats);
ASSERT(resp.getOk());
ASSERT_FALSE(resp.isErrDetailsSet());
ASSERT_EQ(resp.getN(), 1);
@ -361,7 +365,8 @@ TEST_F(UnifiedWriteExecutorTest, BulkWriteImplicitCollectionCreation) {
{NamespaceInfoEntry(nss1)});
auto future = launchAsync([&]() {
auto replyInfo = bulkWrite(operationContext(), request);
Stats uweStats;
auto replyInfo = bulkWrite(operationContext(), request, uweStats);
auto reply = populateCursorReply(operationContext(), request, request.toBSON(), replyInfo);
auto replyItems = reply.getCursor().getFirstBatch();
ASSERT_EQ(replyItems.size(), 1);
@ -432,7 +437,8 @@ TEST_F(UnifiedWriteExecutorTest, OrderedBulkWriteErrorsAndStops) {
request.setOrdered(true);
auto future = launchAsync([&]() {
auto replyInfo = bulkWrite(operationContext(), request);
Stats uweStats;
auto replyInfo = bulkWrite(operationContext(), request, uweStats);
auto reply = populateCursorReply(operationContext(), request, request.toBSON(), replyInfo);
auto replyItems = reply.getCursor().getFirstBatch();
ASSERT_EQ(replyItems.size(), 1);
@ -482,7 +488,8 @@ TEST_F(UnifiedWriteExecutorTest, UnorderedBulkWriteErrorsAndStops) {
auto future = launchAsync([&]() {
auto replyInfo = bulkWrite(operationContext(), request);
Stats uweStats;
auto replyInfo = bulkWrite(operationContext(), request, uweStats);
auto reply = populateCursorReply(operationContext(), request, request.toBSON(), replyInfo);
auto replyItems = reply.getCursor().getFirstBatch();
ASSERT_EQ(replyItems.size(), 2);

8
src/mongo/s/write_ops/unified_write_executor/write_batch_response_processor.cpp
View File

@ -1228,6 +1228,11 @@ WriteBatchResponseProcessor::generateClientResponseForBulkWriteCommand(Operation
auto finalResults = finalizeRepliesForOps(opCtx);
for (auto& [id, item] : finalResults) {
// Check to see if we've recieved a WCOS error. If we have, then we should skip incrementing
// the relevant op counters as this will doulbe count our update. Instead, we will increment
// this as part of the WCOS handling in the command layer.
const bool isNotAWCOSError = !item || item->getStatus().isOK() ||
item->getStatus() != ErrorCodes::WouldChangeOwningShard || _cmdRef.getNumOps() > 1;
if (!_isNonVerbose && (!errorsOnly || (item && !item->getStatus().isOK()))) {
tassert(11182221, "Expected a BulkWriteReplyItem", item.has_value());
@ -1238,10 +1243,9 @@ WriteBatchResponseProcessor::generateClientResponseForBulkWriteCommand(Operation
item->getIdx(),
id),
static_cast<WriteOpId>(item->getIdx()) == id);
results.emplace_back(std::move(*item));
}
_stats.incrementOpCounters(opCtx, _cmdRef.getOp(id));
_stats.incrementOpCounters(opCtx, _cmdRef.getOp(id), isNotAWCOSError);
}
bulk_write_exec::SummaryFields fields(

2
src/mongo/s/write_ops/unified_write_executor/write_op_analyzer.cpp
View File

@ -139,6 +139,8 @@ StatusWith<Analysis> WriteOpAnalyzerImpl::analyze(OperationContext* opCtx,
!(tr.useTwoPhaseWriteProtocol || tr.isNonTargetedRetryableWriteWithId));
// Note we do not translate viewful timeseries collection namespace here, it will be
// translated within the transaction when we analyze the request again.
// Note also that we do not record any targeting stats here as we will do so when we analyze
// the request a second time.
return Analysis{AnalysisType::kInternalTransaction,
std::move(tr.endpoints),
false /* isTimeseries */,