Skip to content

Recovery trigger observability #12490

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 2 commits into
base: main
Choose a base branch
from
Draft

Recovery trigger observability #12490

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d34923f
Recovery trigger observability
spraza Oct 21, 2025
88d9323
using standalone stringref
spraza Oct 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 11 additions & 9 deletions fdbserver/ClusterController.actor.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -297,15 +297,17 @@ ACTOR Future<Void> clusterWatchDatabase(ClusterControllerData* cluster,
// really don't want to have to start over
loop choose {
when(wait(recoveryCore)) {}
when(wait(waitFailureClient(
iMaster.waitFailure,
db->masterRegistrationCount
? SERVER_KNOBS->MASTER_FAILURE_REACTION_TIME
: (now() - recoveryStart) * SERVER_KNOBS->MASTER_FAILURE_SLOPE_DURING_RECOVERY,
db->masterRegistrationCount ? -SERVER_KNOBS->MASTER_FAILURE_REACTION_TIME /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY
: SERVER_KNOBS->MASTER_FAILURE_SLOPE_DURING_RECOVERY) ||
db->forceMasterFailure.onTrigger())) {
when(wait(
traceAfter(waitFailureClient(
iMaster.waitFailure,
db->masterRegistrationCount
? SERVER_KNOBS->MASTER_FAILURE_REACTION_TIME
: (now() - recoveryStart) * SERVER_KNOBS->MASTER_FAILURE_SLOPE_DURING_RECOVERY,
db->masterRegistrationCount ? -SERVER_KNOBS->MASTER_FAILURE_REACTION_TIME /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY
: SERVER_KNOBS->MASTER_FAILURE_SLOPE_DURING_RECOVERY),
"CCWDBMasterFailed") ||
traceAfter(db->forceMasterFailure.onTrigger(), "CCWDBForceMasterFailureTriggered"))) {
break;
}
when(wait(db->serverInfo->onChange())) {}
Expand Down
55 changes: 31 additions & 24 deletions fdbserver/TagPartitionedLogSystem.actor.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -475,33 +475,36 @@ ACTOR Future<Void> TagPartitionedLogSystem::onError_internal(TagPartitionedLogSy
for (auto& it : self->tLogs) {
for (auto& t : it->logServers) {
if (t->get().present()) {
failed.push_back(
waitFailureClient(t->get().interf().waitFailure,
SERVER_KNOBS->TLOG_TIMEOUT,
-SERVER_KNOBS->TLOG_TIMEOUT / SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/*trace=*/true));
failed.push_back(waitFailureClient(t->get().interf().waitFailure,
/* failureReactionTime */ SERVER_KNOBS->TLOG_TIMEOUT,
/* failureReactionSlope */ -SERVER_KNOBS->TLOG_TIMEOUT /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/* trace */ true,
/* traceMsg */ "TLogFailed"_sr));
} else {
changes.push_back(t->onChange());
}
}
for (auto& t : it->logRouters) {
if (t->get().present()) {
failed.push_back(
waitFailureClient(t->get().interf().waitFailure,
SERVER_KNOBS->TLOG_TIMEOUT,
-SERVER_KNOBS->TLOG_TIMEOUT / SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/*trace=*/true));
failed.push_back(waitFailureClient(t->get().interf().waitFailure,
/* failureReactionTime */ SERVER_KNOBS->TLOG_TIMEOUT,
/* failureReactionSlope */ -SERVER_KNOBS->TLOG_TIMEOUT /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/* trace */ true,
/* traceMsg */ "LogRouterFailed"_sr));
} else {
changes.push_back(t->onChange());
}
}
for (const auto& worker : it->backupWorkers) {
if (worker->get().present()) {
backupFailed.push_back(
waitFailureClient(worker->get().interf().waitFailure,
SERVER_KNOBS->BACKUP_TIMEOUT,
-SERVER_KNOBS->BACKUP_TIMEOUT / SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/*trace=*/true));
backupFailed.push_back(waitFailureClient(worker->get().interf().waitFailure,
/* failureReactionTime */ SERVER_KNOBS->BACKUP_TIMEOUT,
/* failureReactionSlope */ -SERVER_KNOBS->BACKUP_TIMEOUT /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/* trace */ true,
/* traceMsg */ "BackupWorkerFailed"_sr));
} else {
changes.push_back(worker->onChange());
}
Expand All @@ -514,10 +517,11 @@ ACTOR Future<Void> TagPartitionedLogSystem::onError_internal(TagPartitionedLogSy
for (auto& t : it->logRouters) {
if (t->get().present()) {
failed.push_back(waitFailureClient(t->get().interf().waitFailure,
SERVER_KNOBS->TLOG_TIMEOUT,
-SERVER_KNOBS->TLOG_TIMEOUT /
/* failureReactionTime */ SERVER_KNOBS->TLOG_TIMEOUT,
/* failureReactionSlope */ -SERVER_KNOBS->TLOG_TIMEOUT /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/*trace=*/true));
/* trace */ true,
/* traceMsg */ "OldLogRouterFailed"_sr));
} else {
changes.push_back(t->onChange());
}
Expand All @@ -526,11 +530,13 @@ ACTOR Future<Void> TagPartitionedLogSystem::onError_internal(TagPartitionedLogSy
// Monitor changes of backup workers for old epochs.
for (const auto& worker : old.tLogs[0]->backupWorkers) {
if (worker->get().present()) {
backupFailed.push_back(waitFailureClient(worker->get().interf().waitFailure,
SERVER_KNOBS->BACKUP_TIMEOUT,
-SERVER_KNOBS->BACKUP_TIMEOUT /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/*trace=*/true));
backupFailed.push_back(
waitFailureClient(worker->get().interf().waitFailure,
/* failureReactionTime */ SERVER_KNOBS->BACKUP_TIMEOUT,
/* failureReactionSlope */ -SERVER_KNOBS->BACKUP_TIMEOUT /
SERVER_KNOBS->SECONDS_BEFORE_NO_FAILURE_DELAY,
/* trace */ true,
/* traceMsg */ "OldBackupWorkerFailed"_sr));
} else {
changes.push_back(worker->onChange());
}
Expand All @@ -546,7 +552,8 @@ ACTOR Future<Void> TagPartitionedLogSystem::onError_internal(TagPartitionedLogSy
changes.push_back(self->backupWorkerChanged.onTrigger());

ASSERT(failed.size() >= 1);
wait(quorum(changes, 1) || tagError<Void>(quorum(failed, 1), tlog_failed()) ||
wait(quorum(changes, 1) ||
tagError<Void>(traceAfter(quorum(failed, 1), "TPLSOnErrorLogSystemFailed"), tlog_failed()) ||
tagError<Void>(quorum(backupFailed, 1), backup_worker_failed()));
}
}
Expand Down
15 changes: 12 additions & 3 deletions fdbserver/WaitFailure.actor.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,6 +42,7 @@ ACTOR Future<Void> waitFailureClient(RequestStream<ReplyPromise<Void>> waitFailu
double reactionTime,
double reactionSlope,
bool trace,
Optional<Standalone<StringRef>> traceMsg,
TaskPriority taskID) {
loop {
try {
Expand All @@ -50,9 +51,12 @@ ACTOR Future<Void> waitFailureClient(RequestStream<ReplyPromise<Void>> waitFailu
wait(waitFailure.getReplyUnlessFailedFor(ReplyPromise<Void>(), reactionTime, reactionSlope, taskID));
if (!x.present()) {
if (trace) {
TraceEvent("WaitFailureClient")
.detail("FailedEndpoint", waitFailure.getEndpoint().getPrimaryAddress().toString())
TraceEvent te("WaitFailureClient");
te.detail("FailedEndpoint", waitFailure.getEndpoint().getPrimaryAddress().toString())
.detail("Token", waitFailure.getEndpoint().token);
if (traceMsg.present()) {
te.detail("Context", traceMsg.get());
}
}
return Void();
}
Expand All @@ -72,7 +76,12 @@ ACTOR Future<Void> waitFailureClientStrict(RequestStream<ReplyPromise<Void>> wai
double failureReactionTime,
TaskPriority taskID) {
loop {
wait(waitFailureClient(waitFailure, 0, 0, false, taskID));
wait(waitFailureClient(waitFailure,
/* reactionTime */ 0,
/* reactionSlope */ 0,
/* trace */ false,
/* traceMsg */ Optional<Standalone<StringRef>>(),
taskID));
wait(delay(failureReactionTime, taskID) ||
IFailureMonitor::failureMonitor().onStateEqual(waitFailure.getEndpoint(), FailureStatus(false)));
if (IFailureMonitor::failureMonitor().getState(waitFailure.getEndpoint()).isFailed()) {
Expand Down
1 change: 1 addition & 0 deletions fdbserver/include/fdbserver/WaitFailure.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,7 @@ Future<Void> waitFailureClient(const RequestStream<ReplyPromise<Void>>& waitFail
double const& failureReactionTime = 0,
double const& failureReactionSlope = 0,
bool const& trace = false,
Optional<Standalone<StringRef>> const& traceMsg = Optional<Standalone<StringRef>>(),
TaskPriority const& taskID = TaskPriority::DefaultEndpoint);

// talks to a wait failure server, returns Void on failure, reaction time is always waited
Expand Down