Merge remote-tracking branch 'origin/master' into flake
This commit is contained in:
@ -99,6 +99,8 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
unsigned int maxSilentTime, buildTimeout;
|
||||
unsigned int repeats = step->isDeterministic ? 1 : 0;
|
||||
|
||||
auto conn(dbPool.get());
|
||||
|
||||
{
|
||||
std::set<Build::ptr> dependents;
|
||||
std::set<Step::ptr> steps;
|
||||
@ -122,8 +124,10 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
|
||||
for (auto build2 : dependents) {
|
||||
if (build2->drvPath == step->drvPath) {
|
||||
build = build2;
|
||||
enqueueNotificationItem({NotificationItem::Type::BuildStarted, build->id});
|
||||
build = build2;
|
||||
pqxx::work txn(*conn);
|
||||
notifyBuildStarted(txn, build->id);
|
||||
txn.commit();
|
||||
}
|
||||
{
|
||||
auto i = jobsetRepeats.find(std::make_pair(build2->projectName, build2->jobsetName));
|
||||
@ -144,8 +148,6 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
|
||||
bool quit = buildId == buildOne && step->drvPath == buildDrvPath;
|
||||
|
||||
auto conn(dbPool.get());
|
||||
|
||||
RemoteResult result;
|
||||
BuildOutput res;
|
||||
unsigned int stepNr = 0;
|
||||
@ -170,11 +172,6 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
} catch (...) {
|
||||
ignoreException();
|
||||
}
|
||||
|
||||
/* Asynchronously run plugins. FIXME: if we're killed,
|
||||
plugin actions might not be run. Need to ensure
|
||||
at-least-once semantics. */
|
||||
enqueueNotificationItem({NotificationItem::Type::StepFinished, buildId, {}, stepNr, result.logFile});
|
||||
}
|
||||
});
|
||||
|
||||
@ -231,6 +228,11 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
time_t stepStopTime = time(0);
|
||||
if (!result.stopTime) result.stopTime = stepStopTime;
|
||||
|
||||
/* For standard failures, we don't care about the error
|
||||
message. */
|
||||
if (result.stepStatus != bsAborted)
|
||||
result.errorMsg = "";
|
||||
|
||||
/* Account the time we spent building this step by dividing it
|
||||
among the jobsets that depend on it. */
|
||||
{
|
||||
@ -243,6 +245,13 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
}
|
||||
}
|
||||
|
||||
/* Finish the step in the database. */
|
||||
if (stepNr) {
|
||||
pqxx::work txn(*conn);
|
||||
finishBuildStep(txn, result, buildId, stepNr, machine->sshName);
|
||||
txn.commit();
|
||||
}
|
||||
|
||||
/* The step had a hopefully temporary failure (e.g. network
|
||||
issue). Retry a number of times. */
|
||||
if (result.canRetry) {
|
||||
@ -256,11 +265,6 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
}
|
||||
if (retry) {
|
||||
auto mc = startDbUpdate();
|
||||
{
|
||||
pqxx::work txn(*conn);
|
||||
finishBuildStep(txn, result, buildId, stepNr, machine->sshName);
|
||||
txn.commit();
|
||||
}
|
||||
stepFinished = true;
|
||||
if (quit) exit(1);
|
||||
return sRetry;
|
||||
@ -315,8 +319,6 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
|
||||
pqxx::work txn(*conn);
|
||||
|
||||
finishBuildStep(txn, result, buildId, stepNr, machine->sshName);
|
||||
|
||||
for (auto & b : direct) {
|
||||
printMsg(lvlInfo, format("marking build %1% as succeeded") % b->id);
|
||||
markSucceededBuild(txn, b, res, buildId != b->id || result.isCached,
|
||||
@ -342,8 +344,12 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
|
||||
/* Send notification about the builds that have this step as
|
||||
the top-level. */
|
||||
for (auto id : buildIDs)
|
||||
enqueueNotificationItem({NotificationItem::Type::BuildFinished, id});
|
||||
{
|
||||
pqxx::work txn(*conn);
|
||||
for (auto id : buildIDs)
|
||||
notifyBuildFinished(txn, id, {});
|
||||
txn.commit();
|
||||
}
|
||||
|
||||
/* Wake up any dependent steps that have no other
|
||||
dependencies. */
|
||||
@ -369,11 +375,6 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
|
||||
} else {
|
||||
|
||||
/* For standard failures, we don't care about the error
|
||||
message. */
|
||||
if (result.stepStatus != bsAborted)
|
||||
result.errorMsg = "";
|
||||
|
||||
/* Register failure in the database for all Build objects that
|
||||
directly or indirectly depend on this step. */
|
||||
|
||||
@ -419,11 +420,6 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
result.stepStatus, result.errorMsg, buildId == build2->id ? 0 : buildId);
|
||||
}
|
||||
|
||||
if (result.stepStatus != bsCachedFailure && !stepFinished) {
|
||||
assert(stepNr);
|
||||
finishBuildStep(txn, result, buildId, stepNr, machine->sshName);
|
||||
}
|
||||
|
||||
/* Mark all builds that depend on this derivation as failed. */
|
||||
for (auto & build2 : indirect) {
|
||||
if (build2->finishedInDB) continue;
|
||||
@ -462,11 +458,10 @@ State::StepResult State::doBuildStep(nix::ref<Store> destStore,
|
||||
|
||||
/* Send notification about this build and its dependents. */
|
||||
{
|
||||
auto notificationSenderQueue_(notificationSenderQueue.lock());
|
||||
notificationSenderQueue_->push(NotificationItem{NotificationItem::Type::BuildFinished, buildId, dependentIDs});
|
||||
pqxx::work txn(*conn);
|
||||
notifyBuildFinished(txn, buildId, dependentIDs);
|
||||
txn.commit();
|
||||
}
|
||||
notificationSenderWakeup.notify_one();
|
||||
|
||||
}
|
||||
|
||||
// FIXME: keep stats about aborted steps?
|
||||
|
@ -268,6 +268,9 @@ unsigned int State::createBuildStep(pqxx::work & txn, time_t startTime, BuildID
|
||||
("insert into BuildStepOutputs (build, stepnr, name, path) values ($1, $2, $3, $4)")
|
||||
(buildId)(stepNr)(output.first)(output.second.path).exec();
|
||||
|
||||
if (status == bsBusy)
|
||||
txn.exec(fmt("notify step_started, '%d\t%d'", buildId, stepNr));
|
||||
|
||||
return stepNr;
|
||||
}
|
||||
|
||||
@ -299,6 +302,9 @@ void State::finishBuildStep(pqxx::work & txn, const RemoteResult & result,
|
||||
(result.timesBuilt, result.timesBuilt > 0)
|
||||
(result.isNonDeterministic, result.timesBuilt > 1)
|
||||
.exec();
|
||||
assert(result.logFile.find('\t') == std::string::npos);
|
||||
txn.exec(fmt("notify step_finished, '%d\t%d\t%s'",
|
||||
buildId, stepNr, result.logFile));
|
||||
}
|
||||
|
||||
|
||||
@ -450,74 +456,20 @@ bool State::checkCachedFailure(Step::ptr step, Connection & conn)
|
||||
}
|
||||
|
||||
|
||||
void State::notificationSender()
|
||||
void State::notifyBuildStarted(pqxx::work & txn, BuildID buildId)
|
||||
{
|
||||
while (true) {
|
||||
try {
|
||||
txn.exec(fmt("notify build_started, '%s'", buildId));
|
||||
}
|
||||
|
||||
NotificationItem item;
|
||||
{
|
||||
auto notificationSenderQueue_(notificationSenderQueue.lock());
|
||||
while (notificationSenderQueue_->empty())
|
||||
notificationSenderQueue_.wait(notificationSenderWakeup);
|
||||
item = notificationSenderQueue_->front();
|
||||
notificationSenderQueue_->pop();
|
||||
}
|
||||
|
||||
MaintainCount<counter> mc(nrNotificationsInProgress);
|
||||
|
||||
printMsg(lvlChatty, format("sending notification about build %1%") % item.id);
|
||||
|
||||
auto now1 = std::chrono::steady_clock::now();
|
||||
|
||||
Pid pid = startProcess([&]() {
|
||||
Strings argv;
|
||||
switch (item.type) {
|
||||
case NotificationItem::Type::BuildStarted:
|
||||
argv = {"hydra-notify", "build-started", std::to_string(item.id)};
|
||||
for (auto id : item.dependentIds)
|
||||
argv.push_back(std::to_string(id));
|
||||
break;
|
||||
case NotificationItem::Type::BuildFinished:
|
||||
argv = {"hydra-notify", "build-finished", std::to_string(item.id)};
|
||||
for (auto id : item.dependentIds)
|
||||
argv.push_back(std::to_string(id));
|
||||
break;
|
||||
case NotificationItem::Type::StepFinished:
|
||||
argv = {"hydra-notify", "step-finished", std::to_string(item.id), std::to_string(item.stepNr), item.logPath};
|
||||
break;
|
||||
};
|
||||
printMsg(lvlChatty, "Executing hydra-notify " + concatStringsSep(" ", argv));
|
||||
execvp("hydra-notify", (char * *) stringsToCharPtrs(argv).data()); // FIXME: remove cast
|
||||
throw SysError("cannot start hydra-notify");
|
||||
});
|
||||
|
||||
int res = pid.wait();
|
||||
|
||||
if (!statusOk(res))
|
||||
throw Error("notification about build %d failed: %s", item.id, statusToString(res));
|
||||
|
||||
auto now2 = std::chrono::steady_clock::now();
|
||||
|
||||
if (item.type == NotificationItem::Type::BuildFinished) {
|
||||
auto conn(dbPool.get());
|
||||
pqxx::work txn(*conn);
|
||||
txn.parameterized
|
||||
("update Builds set notificationPendingSince = null where id = $1")
|
||||
(item.id)
|
||||
.exec();
|
||||
txn.commit();
|
||||
}
|
||||
|
||||
nrNotificationTimeMs += std::chrono::duration_cast<std::chrono::milliseconds>(now2 - now1).count();
|
||||
nrNotificationsDone++;
|
||||
|
||||
} catch (std::exception & e) {
|
||||
nrNotificationsFailed++;
|
||||
printMsg(lvlError, format("notification sender: %1%") % e.what());
|
||||
sleep(5);
|
||||
}
|
||||
}
|
||||
void State::notifyBuildFinished(pqxx::work & txn, BuildID buildId,
|
||||
const std::vector<BuildID> & dependentIds)
|
||||
{
|
||||
auto payload = fmt("%d ", buildId);
|
||||
for (auto & d : dependentIds)
|
||||
payload += fmt("%d ", d);
|
||||
// FIXME: apparently parameterized() doesn't support NOTIFY.
|
||||
txn.exec(fmt("notify build_finished, '%s'", payload));
|
||||
}
|
||||
|
||||
|
||||
@ -589,13 +541,6 @@ void State::dumpStatus(Connection & conn, bool log)
|
||||
root.attr("nrDbConnections", dbPool.count());
|
||||
root.attr("nrActiveDbUpdates", nrActiveDbUpdates);
|
||||
root.attr("memoryTokensInUse", memoryTokens.currentUse());
|
||||
root.attr("nrNotificationsDone", nrNotificationsDone);
|
||||
root.attr("nrNotificationsFailed", nrNotificationsFailed);
|
||||
root.attr("nrNotificationsInProgress", nrNotificationsInProgress);
|
||||
root.attr("nrNotificationsPending", notificationSenderQueue.lock()->size());
|
||||
root.attr("nrNotificationTimeMs", nrNotificationTimeMs);
|
||||
uint64_t nrNotificationsTotal = nrNotificationsDone + nrNotificationsFailed;
|
||||
root.attr("nrNotificationTimeAvgMs", nrNotificationsTotal == 0 ? 0.0 : (float) nrNotificationTimeMs / nrNotificationsTotal);
|
||||
|
||||
{
|
||||
auto nested = root.object("machines");
|
||||
@ -843,24 +788,6 @@ void State::run(BuildID buildOne)
|
||||
|
||||
std::thread(&State::dispatcher, this).detach();
|
||||
|
||||
/* Idem for notification sending. */
|
||||
auto maxConcurrentNotifications = config->getIntOption("max-concurrent-notifications", 2);
|
||||
for (uint64_t i = 0; i < maxConcurrentNotifications; ++i)
|
||||
std::thread(&State::notificationSender, this).detach();
|
||||
|
||||
/* Enqueue notification items for builds that were finished
|
||||
previously, but for which we didn't manage to send
|
||||
notifications. */
|
||||
{
|
||||
auto conn(dbPool.get());
|
||||
pqxx::work txn(*conn);
|
||||
auto res = txn.parameterized("select id from Builds where notificationPendingSince > 0").exec();
|
||||
for (auto const & row : res) {
|
||||
auto id = row["id"].as<BuildID>();
|
||||
enqueueNotificationItem({NotificationItem::Type::BuildFinished, id});
|
||||
}
|
||||
}
|
||||
|
||||
/* Periodically clean up orphaned busy steps in the database. */
|
||||
std::thread([&]() {
|
||||
while (true) {
|
||||
|
@ -193,13 +193,12 @@ bool State::getQueuedBuilds(Connection & conn,
|
||||
(build->id)
|
||||
((int) (ex.step->drvPath == build->drvPath ? bsFailed : bsDepFailed))
|
||||
(time(0)).exec();
|
||||
notifyBuildFinished(txn, build->id, {});
|
||||
txn.commit();
|
||||
build->finishedInDB = true;
|
||||
nrBuildsDone++;
|
||||
}
|
||||
|
||||
enqueueNotificationItem({NotificationItem::Type::BuildFinished, build->id});
|
||||
|
||||
return;
|
||||
}
|
||||
|
||||
@ -230,13 +229,12 @@ bool State::getQueuedBuilds(Connection & conn,
|
||||
time_t now = time(0);
|
||||
printMsg(lvlInfo, format("marking build %1% as succeeded (cached)") % build->id);
|
||||
markSucceededBuild(txn, build, res, true, now, now);
|
||||
notifyBuildFinished(txn, build->id, {});
|
||||
txn.commit();
|
||||
}
|
||||
|
||||
build->finishedInDB = true;
|
||||
|
||||
enqueueNotificationItem({NotificationItem::Type::BuildFinished, build->id});
|
||||
|
||||
return;
|
||||
}
|
||||
|
||||
|
@ -347,39 +347,6 @@ private:
|
||||
counter bytesSent{0};
|
||||
counter bytesReceived{0};
|
||||
counter nrActiveDbUpdates{0};
|
||||
counter nrNotificationsDone{0};
|
||||
counter nrNotificationsFailed{0};
|
||||
counter nrNotificationsInProgress{0};
|
||||
counter nrNotificationTimeMs{0};
|
||||
|
||||
/* Notification sender work queue. FIXME: if hydra-queue-runner is
|
||||
killed before it has finished sending notifications about a
|
||||
build, then the notifications may be lost. It would be better
|
||||
to mark builds with pending notification in the database. */
|
||||
struct NotificationItem
|
||||
{
|
||||
enum class Type : char {
|
||||
BuildStarted,
|
||||
BuildFinished,
|
||||
StepFinished,
|
||||
};
|
||||
Type type;
|
||||
BuildID id;
|
||||
std::vector<BuildID> dependentIds;
|
||||
unsigned int stepNr;
|
||||
nix::Path logPath;
|
||||
};
|
||||
nix::Sync<std::queue<NotificationItem>> notificationSenderQueue;
|
||||
std::condition_variable notificationSenderWakeup;
|
||||
|
||||
void enqueueNotificationItem(const NotificationItem && item)
|
||||
{
|
||||
{
|
||||
auto notificationSenderQueue_(notificationSenderQueue.lock());
|
||||
notificationSenderQueue_->emplace(item);
|
||||
}
|
||||
notificationSenderWakeup.notify_one();
|
||||
}
|
||||
|
||||
/* Specific build to do for --build-one (testing only). */
|
||||
BuildID buildOne;
|
||||
@ -540,9 +507,10 @@ private:
|
||||
|
||||
bool checkCachedFailure(Step::ptr step, Connection & conn);
|
||||
|
||||
/* Thread that asynchronously invokes hydra-notify to send build
|
||||
notifications. */
|
||||
void notificationSender();
|
||||
void notifyBuildStarted(pqxx::work & txn, BuildID buildId);
|
||||
|
||||
void notifyBuildFinished(pqxx::work & txn, BuildID buildId,
|
||||
const std::vector<BuildID> & dependentIds);
|
||||
|
||||
/* Acquire the global queue runner lock, or null if somebody else
|
||||
has it. */
|
||||
|
Reference in New Issue
Block a user