Merge pull request #20740 from vespa-engine/geirst/prepare-sampling-of-current-transient-disk-usage

Prepare for tracking the current transient disk usage across all docu…

2 files changed, 16 insertions, 26 deletions

diff --git a/searchlib/src/tests/util/searchable_stats/searchable_stats_test.cpp b/searchlib/src/tests/util/searchable_stats/searchable_stats_test.cpp
index b89324c71f5..ed857d5776b 100644
--- a/searchlib/src/tests/util/searchable_stats/searchable_stats_test.cpp
+++ b/searchlib/src/tests/util/searchable_stats/searchable_stats_test.cpp
@@ -7,43 +7,35 @@ LOG_SETUP("searchable_stats_test");
 
 using namespace search;
 
-TEST(SearchableStatsTest, merge_also_tracks_max_size_on_disk_for_component)
+TEST(SearchableStatsTest, stats_can_be_merged)
 {
     SearchableStats stats;
     EXPECT_EQ(0u, stats.memoryUsage().allocatedBytes());
     EXPECT_EQ(0u, stats.docsInMemory());
     EXPECT_EQ(0u, stats.sizeOnDisk());
-    EXPECT_EQ(0u, stats.max_component_size_on_disk());
+    EXPECT_EQ(0u, stats.fusion_size_on_disk());
     {
         SearchableStats rhs;
         EXPECT_EQ(&rhs.memoryUsage(vespalib::MemoryUsage(100,0,0,0)), &rhs);
         EXPECT_EQ(&rhs.docsInMemory(10), &rhs);
         EXPECT_EQ(&rhs.sizeOnDisk(1000), &rhs);
-        EXPECT_EQ(1000u, rhs.max_component_size_on_disk());
+        EXPECT_EQ(&rhs.fusion_size_on_disk(500), &rhs);
         EXPECT_EQ(&stats.merge(rhs), &stats);
     }
     EXPECT_EQ(100u, stats.memoryUsage().allocatedBytes());
     EXPECT_EQ(10u, stats.docsInMemory());
     EXPECT_EQ(1000u, stats.sizeOnDisk());
-    EXPECT_EQ(1000u, stats.max_component_size_on_disk());
+    EXPECT_EQ(500u, stats.fusion_size_on_disk());
 
     stats.merge(SearchableStats()
                         .memoryUsage(vespalib::MemoryUsage(150,0,0,0))
                         .docsInMemory(15)
-                        .sizeOnDisk(1500));
+                        .sizeOnDisk(1500)
+                        .fusion_size_on_disk(800));
     EXPECT_EQ(250u, stats.memoryUsage().allocatedBytes());
     EXPECT_EQ(25u, stats.docsInMemory());
     EXPECT_EQ(2500u, stats.sizeOnDisk());
-    EXPECT_EQ(1500u, stats.max_component_size_on_disk());
-
-    stats.merge(SearchableStats()
-                        .memoryUsage(vespalib::MemoryUsage(120,0,0,0))
-                        .docsInMemory(12)
-                        .sizeOnDisk(1200));
-    EXPECT_EQ(370u, stats.memoryUsage().allocatedBytes());
-    EXPECT_EQ(37u, stats.docsInMemory());
-    EXPECT_EQ(3700u, stats.sizeOnDisk());
-    EXPECT_EQ(1500u, stats.max_component_size_on_disk());
+    EXPECT_EQ(1300u, stats.fusion_size_on_disk());
 }
 
 GTEST_MAIN_RUN_ALL_TESTS()
diff --git a/searchlib/src/vespa/searchlib/util/searchable_stats.h b/searchlib/src/vespa/searchlib/util/searchable_stats.h
index 970076c12d0..e785a4c4483 100644
--- a/searchlib/src/vespa/searchlib/util/searchable_stats.h
+++ b/searchlib/src/vespa/searchlib/util/searchable_stats.h
@@ -15,11 +15,11 @@ class SearchableStats
 private:
     vespalib::MemoryUsage _memoryUsage;
     size_t _docsInMemory;
-    size_t _sizeOnDisk;
-    size_t _max_component_size_on_disk;
+    size_t _sizeOnDisk; // in bytes
+    size_t _fusion_size_on_disk; // in bytes
 
 public:
-    SearchableStats() : _memoryUsage(), _docsInMemory(0), _sizeOnDisk(0), _max_component_size_on_disk(0) {}
+    SearchableStats() : _memoryUsage(), _docsInMemory(0), _sizeOnDisk(0), _fusion_size_on_disk(0) {}
     SearchableStats &memoryUsage(const vespalib::MemoryUsage &usage) {
         _memoryUsage = usage;
         return *this;
@@ -32,22 +32,20 @@ public:
     size_t docsInMemory() const { return _docsInMemory; }
     SearchableStats &sizeOnDisk(size_t value) {
         _sizeOnDisk = value;
-        _max_component_size_on_disk = value;
         return *this;
     }
     size_t sizeOnDisk() const { return _sizeOnDisk; }
-
-    /**
-     * Returns the max disk size used by a single Searchable component,
-     * e.g. among the components that are merged into a SearchableStats instance via merge().
-     */
-    size_t max_component_size_on_disk() const { return _max_component_size_on_disk; }
+    SearchableStats& fusion_size_on_disk(size_t value) {
+        _fusion_size_on_disk = value;
+        return *this;
+    }
+    size_t fusion_size_on_disk() const { return _fusion_size_on_disk; }
 
     SearchableStats &merge(const SearchableStats &rhs) {
         _memoryUsage.merge(rhs._memoryUsage);
         _docsInMemory += rhs._docsInMemory;
         _sizeOnDisk += rhs._sizeOnDisk;
-        _max_component_size_on_disk = std::max(_max_component_size_on_disk, rhs._sizeOnDisk);
+        _fusion_size_on_disk += rhs._fusion_size_on_disk;
         return *this;
     }
 };