SERVER-54727 Make resumable index builds work on time-series collections

5 files changed, 115 insertions, 18 deletions

diff --git a/jstests/noPassthrough/libs/index_build.js b/jstests/noPassthrough/libs/index_build.js
index b3ef206858d..7fe4ed625c8 100644
--- a/jstests/noPassthrough/libs/index_build.js
+++ b/jstests/noPassthrough/libs/index_build.js
@@ -137,6 +137,28 @@ var IndexBuildTest = class {
     }
 
     /**
+     * Returns true if the passed in collection is clustered.
+     */
+    static isCollectionClustered(coll) {
+        const res = assert.commandWorked(
+            coll.getDB().runCommand({listCollections: 1, filter: {name: coll.getName()}}));
+        assert.eq(1, res.cursor.firstBatch.length);
+
+        const collInfo = res.cursor.firstBatch[0];
+        return collInfo.options.hasOwnProperty("clusteredIndex");
+    }
+
+    static assertIndexesIdHelper(coll, numIndexes, readyIndexes, notReadyIndexes, options) {
+        if (!IndexBuildTest.isCollectionClustered(coll)) {
+            numIndexes++;
+            readyIndexes.concat("_id_");
+        }
+
+        return IndexBuildTest.assertIndexes(
+            coll, numIndexes, readyIndexes, notReadyIndexes, options);
+    }
+
+    /**
      * Runs listIndexes command on collection.
      * If 'options' is provided, these will be sent along with the command request.
      * Asserts that all the indexes on this collection fit within the first batch of results.
@@ -253,7 +275,7 @@ const ResumableIndexBuildTest = class {
             indexNamesFlatSinglePerBuild[i] = indexNames[i][0];
             buildUUIDs[i] = extractUUIDFromObject(
                 IndexBuildTest
-                    .assertIndexes(coll, indexNamesFlat.length + 1, ["_id_"], indexNamesFlat, {
+                    .assertIndexesIdHelper(coll, indexNamesFlat.length, [], indexNamesFlat, {
                         includeBuildUUIDs: true
                     })[indexNames[i][0]]
                     .buildUUID);
@@ -371,11 +393,9 @@ const ResumableIndexBuildTest = class {
         const indexNamesFlat = ResumableIndexBuildTest.flatten(indexNames);
         let indexes;
         assert.soonNoExcept(function() {
-            indexes = IndexBuildTest.assertIndexes(coll,
-                                                   indexNamesFlat.length + 1,
-                                                   ["_id_"],
-                                                   indexNamesFlat,
-                                                   {includeBuildUUIDs: true});
+            indexes = IndexBuildTest.assertIndexesIdHelper(
+                coll, indexNamesFlat.length, [], indexNamesFlat, {includeBuildUUIDs: true});
+
             return true;
         });
 
@@ -430,8 +450,7 @@ const ResumableIndexBuildTest = class {
                 namespace: coll.getFullName()
             });
         }
-
-        IndexBuildTest.assertIndexes(coll, indexNames.length + 1, indexNames.concat("_id_"));
+        IndexBuildTest.assertIndexesIdHelper(coll, indexNames.length, indexNames);
     }
 
     /**
@@ -712,7 +731,8 @@ const ResumableIndexBuildTest = class {
 
         const buildUUID = extractUUIDFromObject(
             IndexBuildTest
-                .assertIndexes(coll, 2, ["_id_"], [indexName], {includeBuildUUIDs: true})[indexName]
+                .assertIndexesIdHelper(
+                    coll, 1, [], [indexName], {includeBuildUUIDs: true})[indexName]
                 .buildUUID);
 
         clearRawMongoProgramOutput();
diff --git a/jstests/noPassthrough/resumable_timeseries_index_build_collection_scan_phase.js b/jstests/noPassthrough/resumable_timeseries_index_build_collection_scan_phase.js
new file mode 100644
index 00000000000..a6f9bf3e2d9
--- /dev/null
+++ b/jstests/noPassthrough/resumable_timeseries_index_build_collection_scan_phase.js
@@ -0,0 +1,64 @@
+/**
+ * Tests that resumable index builds on time-series collections in the collection scan phase write
+ * their state to disk upon clean shutdown and are resumed from the same phase to completion when
+ * the node is started back up.
+ *
+ * @tags: [
+ *   requires_majority_read_concern,
+ *   requires_persistence,
+ *   requires_replication,
+ *   requires_wiredtiger,
+ *   sbe_incompatible,
+ * ]
+ */
+(function() {
+"use strict";
+
+load("jstests/noPassthrough/libs/index_build.js");
+load("jstests/core/timeseries/libs/timeseries.js");
+
+const rst = new ReplSetTest({nodes: 1});
+rst.startSet();
+rst.initiate();
+
+const dbName = "test";
+const db = rst.getPrimary().getDB(dbName);
+
+if (!TimeseriesTest.timeseriesCollectionsEnabled(db.getMongo())) {
+    jsTestLog("Skipping test because the time-series collection feature flag is disabled");
+    rst.stopSet();
+    return;
+}
+
+const coll = db.timeseries_resumable_index_build;
+coll.drop();
+
+const timeFieldName = "time";
+const metaFieldName = 'meta';
+assert.commandWorked(db.createCollection(
+    coll.getName(), {timeseries: {timeField: timeFieldName, metaField: metaFieldName}}));
+
+// Use different metadata fields to guarantee creating three individual buckets in the buckets
+// collection.
+for (let i = 0; i < 3; i++) {
+    assert.commandWorked(coll.insert({
+        _id: i,
+        measurement: "measurement",
+        time: ISODate(),
+        meta: i,
+    }));
+}
+
+const bucketColl = db.getCollection("system.buckets." + coll.getName());
+ResumableIndexBuildTest.run(
+    rst,
+    dbName,
+    bucketColl.getName(),
+    [[{"control.min.time": 1}, {"control.max.time": 1}]],
+    [{name: "hangIndexBuildDuringCollectionScanPhaseAfterInsertion", logIdWithBuildUUID: 20386}],
+    /*iteration=*/0,
+    ["collection scan"],
+    [{numScannedAferResume: 2}]);
+
+rst.stopSet();
+})();
\ No newline at end of file
diff --git a/src/mongo/db/catalog/multi_index_block.cpp b/src/mongo/db/catalog/multi_index_block.cpp
index 506136851e4..6076b502305 100644
--- a/src/mongo/db/catalog/multi_index_block.cpp
+++ b/src/mongo/db/catalog/multi_index_block.cpp
@@ -897,8 +897,15 @@ BSONObj MultiIndexBlock::_constructStateObject(OperationContext* opCtx,
 
     // We can be interrupted by shutdown before inserting the first document from the collection
     // scan, in which case there is no _lastRecordIdInserted.
-    if (_phase == IndexBuildPhaseEnum::kCollectionScan && _lastRecordIdInserted)
-        builder.append("collectionScanPosition", _lastRecordIdInserted->asLong());
+    if (_phase == IndexBuildPhaseEnum::kCollectionScan && _lastRecordIdInserted) {
+        _lastRecordIdInserted->withFormat(
+            [](RecordId::Null n) { invariant(false); },
+            [&](int64_t rid) { builder.append("collectionScanPosition", rid); },
+            [&](const char* str, int size) {
+                OID oid = OID::from(str);
+                builder.appendOID("collectionScanPosition", &oid);
+            });
+    }
 
     BSONArrayBuilder indexesArray(builder.subarrayStart("indexes"));
     for (const auto& index : _indexes) {
diff --git a/src/mongo/db/index_builds_coordinator.cpp b/src/mongo/db/index_builds_coordinator.cpp
index 63632382041..4fa1338c73e 100644
--- a/src/mongo/db/index_builds_coordinator.cpp
+++ b/src/mongo/db/index_builds_coordinator.cpp
@@ -2178,12 +2178,17 @@ void IndexBuildsCoordinator::_resumeIndexBuildFromPhase(
 
     if (resumeInfo.getPhase() == IndexBuildPhaseEnum::kInitialized ||
         resumeInfo.getPhase() == IndexBuildPhaseEnum::kCollectionScan) {
-        _scanCollectionAndInsertSortedKeysIntoIndex(
-            opCtx,
-            replState,
-            resumeInfo.getCollectionScanPosition()
-                ? boost::make_optional<RecordId>(RecordId(*resumeInfo.getCollectionScanPosition()))
-                : boost::none);
+        boost::optional<RecordId> resumeAfterRecordId;
+        if (resumeInfo.getCollectionScanPosition()) {
+            auto scanPosition = *resumeInfo.getCollectionScanPosition();
+            if (auto recordIdOIDPtr = stdx::get_if<OID>(&scanPosition)) {
+                resumeAfterRecordId.emplace(recordIdOIDPtr->view().view(), OID::kOIDSize);
+            } else if (auto recordIdLongPtr = stdx::get_if<int64_t>(&scanPosition)) {
+                resumeAfterRecordId.emplace(RecordId(*recordIdLongPtr));
+            }
+        }
+
+        _scanCollectionAndInsertSortedKeysIntoIndex(opCtx, replState, resumeAfterRecordId);
     } else if (resumeInfo.getPhase() == IndexBuildPhaseEnum::kBulkLoad) {
         _insertSortedKeysIntoIndexForResume(opCtx, replState);
     }
diff --git a/src/mongo/db/resumable_index_builds.idl b/src/mongo/db/resumable_index_builds.idl
index 8ae9bb0844e..6e940dc3c02 100644
--- a/src/mongo/db/resumable_index_builds.idl
+++ b/src/mongo/db/resumable_index_builds.idl
@@ -116,7 +116,8 @@ structs:
                 type: uuid
             collectionScanPosition:
                 description: "The last record id inserted into the sorter before shutdown"
-                type: long
+                type:
+                    variant: [objectid, long]
                 optional: true
             indexes:
                 description: "The information needed to resume each specific index in this build"