forked from lix-project/hydra
092d60735b
I.e., how much time the currently runnable steps per system type have been waiting. This is useful for deciding whether to provision more machines.
329 lines
11 KiB
C++
329 lines
11 KiB
C++
#include <algorithm>
|
||
#include <thread>
|
||
#include <unordered_map>
|
||
|
||
#include "state.hh"
|
||
|
||
using namespace nix;
|
||
|
||
|
||
void State::makeRunnable(Step::ptr step)
|
||
{
|
||
printMsg(lvlChatty, format("step ‘%1%’ is now runnable") % step->drvPath);
|
||
|
||
{
|
||
auto step_(step->state.lock());
|
||
assert(step_->created);
|
||
assert(!step->finished);
|
||
assert(step_->deps.empty());
|
||
step_->runnableSince = std::chrono::system_clock::now();
|
||
}
|
||
|
||
{
|
||
auto runnable_(runnable.lock());
|
||
runnable_->push_back(step);
|
||
}
|
||
|
||
wakeDispatcher();
|
||
}
|
||
|
||
|
||
void State::dispatcher()
|
||
{
|
||
while (true) {
|
||
printMsg(lvlDebug, "dispatcher woken up");
|
||
|
||
auto sleepUntil = doDispatch();
|
||
|
||
/* Sleep until we're woken up (either because a runnable build
|
||
is added, or because a build finishes). */
|
||
{
|
||
auto dispatcherWakeup_(dispatcherWakeup.lock());
|
||
if (!*dispatcherWakeup_) {
|
||
printMsg(lvlDebug, format("dispatcher sleeping for %1%s") %
|
||
std::chrono::duration_cast<std::chrono::seconds>(sleepUntil - std::chrono::system_clock::now()).count());
|
||
dispatcherWakeup_.wait_until(dispatcherWakeupCV, sleepUntil);
|
||
}
|
||
nrDispatcherWakeups++;
|
||
*dispatcherWakeup_ = false;
|
||
}
|
||
}
|
||
|
||
printMsg(lvlError, "dispatcher exits");
|
||
}
|
||
|
||
|
||
system_time State::doDispatch()
|
||
{
|
||
/* Prune old historical build step info from the jobsets. */
|
||
{
|
||
auto jobsets_(jobsets.lock());
|
||
for (auto & jobset : *jobsets_) {
|
||
auto s1 = jobset.second->shareUsed();
|
||
jobset.second->pruneSteps();
|
||
auto s2 = jobset.second->shareUsed();
|
||
if (s1 != s2)
|
||
printMsg(lvlDebug, format("pruned scheduling window of ‘%1%:%2%’ from %3% to %4%")
|
||
% jobset.first.first % jobset.first.second % s1 % s2);
|
||
}
|
||
}
|
||
|
||
/* Start steps until we're out of steps or slots. */
|
||
auto sleepUntil = system_time::max();
|
||
bool keepGoing;
|
||
|
||
do {
|
||
system_time now = std::chrono::system_clock::now();
|
||
|
||
/* Copy the currentJobs field of each machine. This is
|
||
necessary to ensure that the sort comparator below is
|
||
an ordering. std::sort() can segfault if it isn't. Also
|
||
filter out temporarily disabled machines. */
|
||
struct MachineInfo
|
||
{
|
||
Machine::ptr machine;
|
||
unsigned int currentJobs;
|
||
};
|
||
std::vector<MachineInfo> machinesSorted;
|
||
{
|
||
auto machines_(machines.lock());
|
||
for (auto & m : *machines_) {
|
||
auto info(m.second->state->connectInfo.lock());
|
||
if (info->consecutiveFailures && info->disabledUntil > now) {
|
||
if (info->disabledUntil < sleepUntil)
|
||
sleepUntil = info->disabledUntil;
|
||
continue;
|
||
}
|
||
machinesSorted.push_back({m.second, m.second->state->currentJobs});
|
||
}
|
||
}
|
||
|
||
/* Sort the machines by a combination of speed factor and
|
||
available slots. Prioritise the available machines as
|
||
follows:
|
||
|
||
- First by load divided by speed factor, rounded to the
|
||
nearest integer. This causes fast machines to be
|
||
preferred over slow machines with similar loads.
|
||
|
||
- Then by speed factor.
|
||
|
||
- Finally by load. */
|
||
sort(machinesSorted.begin(), machinesSorted.end(),
|
||
[](const MachineInfo & a, const MachineInfo & b) -> bool
|
||
{
|
||
float ta = roundf(a.currentJobs / a.machine->speedFactor);
|
||
float tb = roundf(b.currentJobs / b.machine->speedFactor);
|
||
return
|
||
ta != tb ? ta < tb :
|
||
a.machine->speedFactor != b.machine->speedFactor ? a.machine->speedFactor > b.machine->speedFactor :
|
||
a.currentJobs > b.currentJobs;
|
||
});
|
||
|
||
/* Sort the runnable steps by priority. Priority is establised
|
||
as follows (in order of precedence):
|
||
|
||
- The global priority of the builds that depend on the
|
||
step. This allows admins to bump a build to the front of
|
||
the queue.
|
||
|
||
- The lowest used scheduling share of the jobsets depending
|
||
on the step.
|
||
|
||
- The local priority of the build, as set via the build's
|
||
meta.schedulingPriority field. Note that this is not
|
||
quite correct: the local priority should only be used to
|
||
establish priority between builds in the same jobset, but
|
||
here it's used between steps in different jobsets if they
|
||
happen to have the same lowest used scheduling share. But
|
||
that's not every likely.
|
||
|
||
- The lowest ID of the builds depending on the step;
|
||
i.e. older builds take priority over new ones.
|
||
|
||
FIXME: O(n lg n); obviously, it would be better to keep a
|
||
runnable queue sorted by priority. */
|
||
std::vector<Step::ptr> runnableSorted;
|
||
struct RunnablePerType
|
||
{
|
||
unsigned int count{0};
|
||
std::chrono::seconds waitTime{0};
|
||
};
|
||
std::unordered_map<std::string, RunnablePerType> runnablePerType;
|
||
{
|
||
auto runnable_(runnable.lock());
|
||
runnableSorted.reserve(runnable_->size());
|
||
for (auto i = runnable_->begin(); i != runnable_->end(); ) {
|
||
auto step = i->lock();
|
||
|
||
/* Remove dead steps. */
|
||
if (!step) {
|
||
i = runnable_->erase(i);
|
||
continue;
|
||
}
|
||
|
||
++i;
|
||
|
||
auto & r = runnablePerType[step->systemType];
|
||
r.count++;
|
||
|
||
/* Skip previously failed steps that aren't ready
|
||
to be retried. */
|
||
{
|
||
auto step_(step->state.lock());
|
||
r.waitTime += std::chrono::duration_cast<std::chrono::seconds>(now - step_->runnableSince);
|
||
if (step_->tries > 0 && step_->after > now) {
|
||
if (step_->after < sleepUntil)
|
||
sleepUntil = step_->after;
|
||
continue;
|
||
}
|
||
}
|
||
|
||
runnableSorted.push_back(step);
|
||
}
|
||
}
|
||
|
||
for (auto & step : runnableSorted) {
|
||
auto step_(step->state.lock());
|
||
step_->lowestShareUsed = 1e9;
|
||
for (auto & jobset : step_->jobsets)
|
||
step_->lowestShareUsed = std::min(step_->lowestShareUsed, jobset->shareUsed());
|
||
}
|
||
|
||
sort(runnableSorted.begin(), runnableSorted.end(),
|
||
[](const Step::ptr & a, const Step::ptr & b)
|
||
{
|
||
auto a_(a->state.lock());
|
||
auto b_(b->state.lock()); // FIXME: deadlock?
|
||
return
|
||
a_->highestGlobalPriority != b_->highestGlobalPriority ? a_->highestGlobalPriority > b_->highestGlobalPriority :
|
||
a_->lowestShareUsed != b_->lowestShareUsed ? a_->lowestShareUsed < b_->lowestShareUsed :
|
||
a_->highestLocalPriority != b_->highestLocalPriority ? a_->highestLocalPriority > b_->highestLocalPriority :
|
||
a_->lowestBuildID < b_->lowestBuildID;
|
||
});
|
||
|
||
/* Find a machine with a free slot and find a step to run
|
||
on it. Once we find such a pair, we restart the outer
|
||
loop because the machine sorting will have changed. */
|
||
keepGoing = false;
|
||
|
||
for (auto & mi : machinesSorted) {
|
||
if (mi.machine->state->currentJobs >= mi.machine->maxJobs) continue;
|
||
|
||
for (auto & step : runnableSorted) {
|
||
|
||
/* Can this machine do this step? */
|
||
if (!mi.machine->supportsStep(step)) continue;
|
||
|
||
/* Let's do this step. Remove it from the runnable
|
||
list. FIXME: O(n). */
|
||
{
|
||
auto runnable_(runnable.lock());
|
||
bool removed = false;
|
||
for (auto i = runnable_->begin(); i != runnable_->end(); )
|
||
if (i->lock() == step) {
|
||
i = runnable_->erase(i);
|
||
removed = true;
|
||
break;
|
||
} else ++i;
|
||
assert(removed);
|
||
auto & r = runnablePerType[step->systemType];
|
||
assert(r.count);
|
||
r.count--;
|
||
}
|
||
|
||
/* Make a slot reservation and start a thread to
|
||
do the build. */
|
||
auto builderThread = std::thread(&State::builder, this,
|
||
std::make_shared<MachineReservation>(*this, step, mi.machine));
|
||
builderThread.detach(); // FIXME?
|
||
|
||
keepGoing = true;
|
||
break;
|
||
}
|
||
|
||
if (keepGoing) break;
|
||
}
|
||
|
||
/* Update the stats for the auto-scaler. */
|
||
{
|
||
auto machineTypes_(machineTypes.lock());
|
||
|
||
for (auto & i : *machineTypes_)
|
||
i.second.runnable = 0;
|
||
|
||
for (auto & i : runnablePerType) {
|
||
auto & j = (*machineTypes_)[i.first];
|
||
j.runnable = i.second.count;
|
||
j.waitTime = i.second.waitTime;
|
||
}
|
||
}
|
||
|
||
lastDispatcherCheck = std::chrono::system_clock::to_time_t(now);
|
||
|
||
} while (keepGoing);
|
||
|
||
return sleepUntil;
|
||
}
|
||
|
||
|
||
void State::wakeDispatcher()
|
||
{
|
||
{
|
||
auto dispatcherWakeup_(dispatcherWakeup.lock());
|
||
*dispatcherWakeup_ = true;
|
||
}
|
||
dispatcherWakeupCV.notify_one();
|
||
}
|
||
|
||
|
||
void Jobset::addStep(time_t startTime, time_t duration)
|
||
{
|
||
auto steps_(steps.lock());
|
||
(*steps_)[startTime] = duration;
|
||
seconds += duration;
|
||
}
|
||
|
||
|
||
void Jobset::pruneSteps()
|
||
{
|
||
time_t now = time(0);
|
||
auto steps_(steps.lock());
|
||
while (!steps_->empty()) {
|
||
auto i = steps_->begin();
|
||
if (i->first > now - schedulingWindow) break;
|
||
seconds -= i->second;
|
||
steps_->erase(i);
|
||
}
|
||
}
|
||
|
||
|
||
State::MachineReservation::MachineReservation(State & state, Step::ptr step, Machine::ptr machine)
|
||
: state(state), step(step), machine(machine)
|
||
{
|
||
machine->state->currentJobs++;
|
||
|
||
{
|
||
auto machineTypes_(state.machineTypes.lock());
|
||
(*machineTypes_)[step->systemType].running++;
|
||
}
|
||
}
|
||
|
||
|
||
State::MachineReservation::~MachineReservation()
|
||
{
|
||
auto prev = machine->state->currentJobs--;
|
||
assert(prev);
|
||
if (prev == 1)
|
||
machine->state->idleSince = time(0);
|
||
|
||
{
|
||
auto machineTypes_(state.machineTypes.lock());
|
||
auto & machineType = (*machineTypes_)[step->systemType];
|
||
assert(machineType.running);
|
||
machineType.running--;
|
||
if (machineType.running == 0)
|
||
machineType.lastActive = std::chrono::system_clock::now();
|
||
}
|
||
}
|