Merge pull request #8299 from urbas/max-substitution-jobs

`max-substitution-jobs` setting

src/libstore/build/derivation-goal.hh

@@ -333,6 +333,8 @@ struct DerivationGoal : public Goal
     void waiteeDone(GoalPtr waitee, ExitCode result) override;
 
     StorePathSet exportReferences(const StorePathSet & storePaths);
+
+    JobCategory jobCategory() override { return JobCategory::Build; };
 };
 
 MakeError(NotDeterministic, BuildError);

src/libstore/build/drv-output-substitution-goal.hh

@@ -21,7 +21,7 @@ class Worker;
 class DrvOutputSubstitutionGoal : public Goal {
 
     /**
-     * The drv output we're trying to substitue
+     * The drv output we're trying to substitute
      */
     DrvOutput id;
 
@@ -72,6 +72,8 @@ public:
 
     void work() override;
     void handleEOF(int fd) override;
+
+    JobCategory jobCategory() override { return JobCategory::Substitution; };
 };
 
 }

src/libstore/build/goal.hh

@@ -34,6 +34,17 @@ typedef std::set<WeakGoalPtr, std::owner_less<WeakGoalPtr>> WeakGoals;
  */
 typedef std::map<StorePath, WeakGoalPtr> WeakGoalMap;
 
+/**
+ * Used as a hint to the worker on how to schedule a particular goal. For example,
+ * builds are typically CPU- and memory-bound, while substitutions are I/O bound.
+ * Using this information, the worker might decide to schedule more or fewer goals
+ * of each category in parallel.
+ */
+enum struct JobCategory {
+    Build,
+    Substitution,
+};
+
 struct Goal : public std::enable_shared_from_this<Goal>
 {
     typedef enum {ecBusy, ecSuccess, ecFailed, ecNoSubstituters, ecIncompleteClosure} ExitCode;
@@ -150,6 +161,8 @@ public:
     void amDone(ExitCode result, std::optional<Error> ex = {});
 
     virtual void cleanup() { }
+
+    virtual JobCategory jobCategory() = 0;
 };
 
 void addToWeakGoals(WeakGoals & goals, GoalPtr p);

src/libstore/build/substitution-goal.cc

@@ -204,7 +204,7 @@ void PathSubstitutionGoal::tryToRun()
        if maxBuildJobs == 0 (no local builds allowed), we still allow
        a substituter to run.  This is because substitutions cannot be
        distributed to another machine via the build hook. */
-    if (worker.getNrLocalBuilds() >= std::max(1U, (unsigned int) settings.maxBuildJobs)) {
+    if (worker.getNrSubstitutions() >= std::max(1U, (unsigned int) settings.maxSubstitutionJobs)) {
         worker.waitForBuildSlot(shared_from_this());
         return;
     }

src/libstore/build/substitution-goal.hh

@@ -115,6 +115,8 @@ public:
     void handleEOF(int fd) override;
 
     void cleanup() override;
+
+    JobCategory jobCategory() override { return JobCategory::Substitution; };
 };
 
 }

src/libstore/build/worker.cc

@@ -18,6 +18,7 @@ Worker::Worker(Store & store, Store & evalStore)
 {
     /* Debugging: prevent recursive workers. */
     nrLocalBuilds = 0;
+    nrSubstitutions = 0;
     lastWokenUp = steady_time_point::min();
     permanentFailure = false;
     timedOut = false;
@@ -176,6 +177,12 @@ unsigned Worker::getNrLocalBuilds()
 }
 
 
+unsigned Worker::getNrSubstitutions()
+{
+    return nrSubstitutions;
+}
+
+
 void Worker::childStarted(GoalPtr goal, const std::set<int> & fds,
     bool inBuildSlot, bool respectTimeouts)
 {
@@ -187,7 +194,10 @@ void Worker::childStarted(GoalPtr goal, const std::set<int> & fds,
     child.inBuildSlot = inBuildSlot;
     child.respectTimeouts = respectTimeouts;
     children.emplace_back(child);
-    if (inBuildSlot) nrLocalBuilds++;
+    if (inBuildSlot) {
+        if (goal->jobCategory() == JobCategory::Substitution) nrSubstitutions++;
+        else nrLocalBuilds++;
+    }
 }
 
 
@@ -198,8 +208,13 @@ void Worker::childTerminated(Goal * goal, bool wakeSleepers)
     if (i == children.end()) return;
 
     if (i->inBuildSlot) {
-        assert(nrLocalBuilds > 0);
-        nrLocalBuilds--;
+        if (goal->jobCategory() == JobCategory::Substitution) {
+            assert(nrSubstitutions > 0);
+            nrSubstitutions--;
+        } else {
+            assert(nrLocalBuilds > 0);
+            nrLocalBuilds--;
+        }
     }
 
     children.erase(i);
@@ -220,7 +235,9 @@ void Worker::childTerminated(Goal * goal, bool wakeSleepers)
 void Worker::waitForBuildSlot(GoalPtr goal)
 {
     debug("wait for build slot");
-    if (getNrLocalBuilds() < settings.maxBuildJobs)
+    bool isSubstitutionGoal = goal->jobCategory() == JobCategory::Substitution;
+    if ((!isSubstitutionGoal && getNrLocalBuilds() < settings.maxBuildJobs) ||
+        (isSubstitutionGoal && getNrSubstitutions() < settings.maxSubstitutionJobs))
         wakeUp(goal); /* we can do it right away */
     else
         addToWeakGoals(wantingToBuild, goal);

src/libstore/build/worker.hh

@@ -88,11 +88,16 @@ private:
     std::list<Child> children;
 
     /**
-     * Number of build slots occupied.  This includes local builds and
-     * substitutions but not remote builds via the build hook.
+     * Number of build slots occupied.  This includes local builds but does not
+     * include substitutions or remote builds via the build hook.
      */
     unsigned int nrLocalBuilds;
 
+    /**
+     * Number of substitution slots occupied.
+     */
+    unsigned int nrSubstitutions;
+
     /**
      * Maps used to prevent multiple instantiations of a goal for the
      * same derivation / path.
@@ -220,12 +225,16 @@ public:
     void wakeUp(GoalPtr goal);
 
     /**
-     * Return the number of local build and substitution processes
-     * currently running (but not remote builds via the build
-     * hook).
+     * Return the number of local build processes currently running (but not
+     * remote builds via the build hook).
      */
     unsigned int getNrLocalBuilds();
 
+    /**
+     * Return the number of substitution processes currently running.
+     */
+    unsigned int getNrSubstitutions();
+
     /**
      * Registers a running child process.  `inBuildSlot` means that
      * the process counts towards the jobs limit.

src/libstore/globals.hh

@@ -159,6 +159,15 @@ public:
         )",
         {"build-max-jobs"}};
 
+    Setting<unsigned int> maxSubstitutionJobs{
+        this, 16, "max-substitution-jobs",
+        R"(
+          This option defines the maximum number of substitution jobs that Nix
+          will try to run in parallel. The default is `16`. The minimum value
+          one can choose is `1` and lower values will be interpreted as `1`.
+        )",
+        {"substitution-max-jobs"}};
+
     Setting<unsigned int> buildCores{
         this,
         getDefaultCores(),