@powersync/service-module-mongodb 0.6.0 → 0.6.1

package/src/replication/ChangeStream.ts

@@ -14,7 +14,7 @@ import { MongoLSN } from '../common/MongoLSN.js';
 import { PostImagesOption } from '../types/types.js';
 import { escapeRegExp } from '../utils.js';
 import { MongoManager } from './MongoManager.js';
-import { constructAfterRecord, createCheckpoint, getMongoRelation } from './MongoRelation.js';
+import { constructAfterRecord, createCheckpoint, getCacheIdentifier, getMongoRelation } from './MongoRelation.js';
 import { CHECKPOINTS_COLLECTION } from './replication-utils.js';
 
 export interface ChangeStreamOptions {
@@ -89,6 +89,10 @@ export class ChangeStream {
     return this.connections.options.postImages == PostImagesOption.AUTO_CONFIGURE;
   }
 
+  private get logPrefix() {
+    return `[powersync_${this.group_id}]`;
+  }
+
   /**
    * This resolves a pattern, persists the related metadata, and returns
    * the resulting SourceTables.
@@ -124,18 +128,13 @@ export class ChangeStream {
       .toArray();
 
     if (!tablePattern.isWildcard && collections.length == 0) {
-      logger.warn(`Collection ${schema}.${tablePattern.name} not found`);
+      logger.warn(`${this.logPrefix} Collection ${schema}.${tablePattern.name} not found`);
     }
 
     for (let collection of collections) {
       const table = await this.handleRelation(
         batch,
-        {
-          name: collection.name,
-          schema,
-          objectId: collection.name,
-          replicationColumns: [{ name: '_id' }]
-        } as SourceEntityDescriptor,
+        getMongoRelation({ db: schema, coll: collection.name }),
         // This is done as part of the initial setup - snapshot is handled elsewhere
         { snapshot: false, collectionInfo: collection }
       );
@@ -149,7 +148,7 @@ export class ChangeStream {
   async initSlot(): Promise<InitResult> {
     const status = await this.storage.getStatus();
     if (status.snapshot_done && status.checkpoint_lsn) {
-      logger.info(`Initial replication already done`);
+      logger.info(`${this.logPrefix} Initial replication already done`);
       return { needsInitialSync: false };
     }
 
@@ -220,7 +219,7 @@ export class ChangeStream {
           }
 
           const { comparable: lsn } = new MongoLSN({ timestamp: snapshotTime });
-          logger.info(`Snapshot commit at ${snapshotTime.inspect()} / ${lsn}`);
+          logger.info(`${this.logPrefix} Snapshot commit at ${snapshotTime.inspect()} / ${lsn}`);
           await batch.commit(lsn);
         }
       );
@@ -289,7 +288,7 @@ export class ChangeStream {
     table: storage.SourceTable,
     session?: mongo.ClientSession
   ) {
-    logger.info(`Replicating ${table.qualifiedName}`);
+    logger.info(`${this.logPrefix} Replicating ${table.qualifiedName}`);
     const estimatedCount = await this.estimatedCount(table);
     let at = 0;
     let lastLogIndex = 0;
@@ -319,7 +318,7 @@ export class ChangeStream {
 
       at += 1;
       if (at - lastLogIndex >= 5000) {
-        logger.info(`[${this.group_id}] Replicating ${table.qualifiedName} ${at}/${estimatedCount}`);
+        logger.info(`${this.logPrefix}  Replicating ${table.qualifiedName} ${at}/${estimatedCount}`);
         lastLogIndex = at;
       }
       Metrics.getInstance().rows_replicated_total.add(1);
@@ -328,14 +327,16 @@ export class ChangeStream {
     }
 
     await batch.flush();
-    logger.info(`Replicated ${at} documents for ${table.qualifiedName}`);
+    logger.info(`${this.logPrefix} Replicated ${at} documents for ${table.qualifiedName}`);
   }
 
   private async getRelation(
     batch: storage.BucketStorageBatch,
-    descriptor: SourceEntityDescriptor
+    descriptor: SourceEntityDescriptor,
+    options: { snapshot: boolean }
   ): Promise<SourceTable> {
-    const existing = this.relation_cache.get(descriptor.objectId);
+    const cacheId = getCacheIdentifier(descriptor);
+    const existing = this.relation_cache.get(cacheId);
     if (existing != null) {
       return existing;
     }
@@ -344,7 +345,7 @@ export class ChangeStream {
     // missing values.
     const collection = await this.getCollectionInfo(descriptor.schema, descriptor.name);
 
-    return this.handleRelation(batch, descriptor, { snapshot: false, collectionInfo: collection });
+    return this.handleRelation(batch, descriptor, { snapshot: options.snapshot, collectionInfo: collection });
   }
 
   private async getCollectionInfo(db: string, name: string): Promise<mongo.CollectionInfo | undefined> {
@@ -375,7 +376,7 @@ export class ChangeStream {
         collMod: collectionInfo.name,
         changeStreamPreAndPostImages: { enabled: true }
       });
-      logger.info(`Enabled postImages on ${db}.${collectionInfo.name}`);
+      logger.info(`${this.logPrefix} Enabled postImages on ${db}.${collectionInfo.name}`);
     } else if (!enabled) {
       throw new ServiceError(ErrorCode.PSYNC_S1343, `postImages not enabled on ${db}.${collectionInfo.name}`);
     }
@@ -394,9 +395,6 @@ export class ChangeStream {
     }
 
     const snapshot = options.snapshot;
-    if (!descriptor.objectId && typeof descriptor.objectId != 'string') {
-      throw new ReplicationAssertionError('MongoDB replication - objectId expected');
-    }
     const result = await this.storage.resolveTable({
       group_id: this.group_id,
       connection_id: this.connection_id,
@@ -404,10 +402,16 @@ export class ChangeStream {
       entity_descriptor: descriptor,
       sync_rules: this.sync_rules
     });
-    this.relation_cache.set(descriptor.objectId, result.table);
+    this.relation_cache.set(getCacheIdentifier(descriptor), result.table);
 
-    // Drop conflicting tables. This includes for example renamed tables.
-    await batch.drop(result.dropTables);
+    // Drop conflicting collections.
+    // This is generally not expected for MongoDB source dbs, so we log an error.
+    if (result.dropTables.length > 0) {
+      logger.error(
+        `Conflicting collections found for ${JSON.stringify(descriptor)}. Dropping: ${result.dropTables.map((t) => t.id).join(', ')}`
+      );
+      await batch.drop(result.dropTables);
+    }
 
     // Snapshot if:
     // 1. Snapshot is requested (false for initial snapshot, since that process handles it elsewhere)
@@ -415,6 +419,7 @@ export class ChangeStream {
     // 3. The table is used in sync rules.
     const shouldSnapshot = snapshot && !result.table.snapshotComplete && result.table.syncAny;
     if (shouldSnapshot) {
+      logger.info(`${this.logPrefix} New collection: ${descriptor.schema}.${descriptor.name}`);
       // Truncate this table, in case a previous snapshot was interrupted.
       await batch.truncate([result.table]);
 
@@ -434,7 +439,7 @@ export class ChangeStream {
     change: mongo.ChangeStreamDocument
   ): Promise<storage.FlushedResult | null> {
     if (!table.syncAny) {
-      logger.debug(`Collection ${table.qualifiedName} not used in sync rules - skipping`);
+      logger.debug(`${this.logPrefix} Collection ${table.qualifiedName} not used in sync rules - skipping`);
       return null;
     }
 
@@ -528,7 +533,7 @@ export class ChangeStream {
         const startAfter = lastLsn?.timestamp;
         const resumeAfter = lastLsn?.resumeToken;
 
-        logger.info(`Resume streaming at ${startAfter?.inspect()} / ${lastLsn}`);
+        logger.info(`${this.logPrefix} Resume streaming at ${startAfter?.inspect()} / ${lastLsn}`);
 
         const filters = this.getSourceNamespaceFilters();
 
@@ -590,13 +595,14 @@ export class ChangeStream {
 
         let splitDocument: mongo.ChangeStreamDocument | null = null;
 
+        let flexDbNameWorkaroundLogged = false;
+
         while (true) {
           if (this.abort_signal.aborted) {
             break;
           }
 
           const originalChangeDocument = await stream.tryNext();
-
           // The stream was closed, we will only ever receive `null` from it
           if (!originalChangeDocument && stream.closed) {
             break;
@@ -636,6 +642,29 @@ export class ChangeStream {
             throw new ReplicationAssertionError(`Incomplete splitEvent: ${JSON.stringify(splitDocument.splitEvent)}`);
           }
 
+          if (
+            !filters.multipleDatabases &&
+            'ns' in changeDocument &&
+            changeDocument.ns.db != this.defaultDb.databaseName &&
+            changeDocument.ns.db.endsWith(`_${this.defaultDb.databaseName}`)
+          ) {
+            // When all of the following conditions are met:
+            // 1. We're replicating from an Atlas Flex instance.
+            // 2. There were changestream events recorded while the PowerSync service is paused.
+            // 3. We're only replicating from a single database.
+            // Then we've observed an ns with for example {db: '67b83e86cd20730f1e766dde_ps'},
+            // instead of the expected {db: 'ps'}.
+            // We correct this.
+            changeDocument.ns.db = this.defaultDb.databaseName;
+
+            if (!flexDbNameWorkaroundLogged) {
+              flexDbNameWorkaroundLogged = true;
+              logger.warn(
+                `${this.logPrefix} Incorrect DB name in change stream: ${changeDocument.ns.db}. Changed to ${this.defaultDb.databaseName}.`
+              );
+            }
+          }
+
           if (
             (changeDocument.operationType == 'insert' ||
               changeDocument.operationType == 'update' ||
@@ -682,28 +711,44 @@ export class ChangeStream {
               waitForCheckpointLsn = await createCheckpoint(this.client, this.defaultDb);
             }
             const rel = getMongoRelation(changeDocument.ns);
-            const table = await this.getRelation(batch, rel);
+            const table = await this.getRelation(batch, rel, {
+              // In most cases, we should not need to snapshot this. But if this is the first time we see the collection
+              // for whatever reason, then we do need to snapshot it.
+              // This may result in some duplicate operations when a collection is created for the first time after
+              // sync rules was deployed.
+              snapshot: true
+            });
             if (table.syncAny) {
               await this.writeChange(batch, table, changeDocument);
             }
           } else if (changeDocument.operationType == 'drop') {
             const rel = getMongoRelation(changeDocument.ns);
-            const table = await this.getRelation(batch, rel);
+            const table = await this.getRelation(batch, rel, {
+              // We're "dropping" this collection, so never snapshot it.
+              snapshot: false
+            });
             if (table.syncAny) {
               await batch.drop([table]);
-              this.relation_cache.delete(table.objectId);
+              this.relation_cache.delete(getCacheIdentifier(rel));
             }
           } else if (changeDocument.operationType == 'rename') {
             const relFrom = getMongoRelation(changeDocument.ns);
             const relTo = getMongoRelation(changeDocument.to);
-            const tableFrom = await this.getRelation(batch, relFrom);
+            const tableFrom = await this.getRelation(batch, relFrom, {
+              // We're "dropping" this collection, so never snapshot it.
+              snapshot: false
+            });
             if (tableFrom.syncAny) {
               await batch.drop([tableFrom]);
-              this.relation_cache.delete(tableFrom.objectId);
+              this.relation_cache.delete(getCacheIdentifier(relFrom));
             }
             // Here we do need to snapshot the new table
             const collection = await this.getCollectionInfo(relTo.schema, relTo.name);
-            await this.handleRelation(batch, relTo, { snapshot: true, collectionInfo: collection });
+            await this.handleRelation(batch, relTo, {
+              // This is a new (renamed) collection, so always snapshot it.
+              snapshot: true,
+              collectionInfo: collection
+            });
           }
         }
       }

package/src/replication/ChangeStreamReplicationJob.ts

@@ -40,8 +40,8 @@ export class ChangeStreamReplicationJob extends replication.AbstractReplicationJ
       this.logger.error(`Replication failed`, e);
 
       if (e instanceof ChangeStreamInvalidatedError) {
-        // This stops replication on this slot, and creates a new slot
-        await this.options.storage.factory.slotRemoved(this.slotName);
+        // This stops replication and restarts with a new instance
+        await this.options.storage.factory.restartReplication(this.storage.group_id);
       }
     } finally {
       this.abortController.abort();

package/src/replication/MongoRelation.ts

@@ -11,11 +11,19 @@ export function getMongoRelation(source: mongo.ChangeStreamNameSpace): storage.S
   return {
     name: source.coll,
     schema: source.db,
-    objectId: source.coll,
+    // Not relevant for MongoDB - we use db + coll name as the identifier
+    objectId: undefined,
     replicationColumns: [{ name: '_id' }]
   } satisfies storage.SourceEntityDescriptor;
 }
 
+/**
+ * For in-memory cache only.
+ */
+export function getCacheIdentifier(source: storage.SourceEntityDescriptor): string {
+  return `${source.schema}.${source.name}`;
+}
+
 export function constructAfterRecord(document: mongo.Document): SqliteRow {
   let record: SqliteRow = {};
   for (let key of Object.keys(document)) {

package/test/src/change_stream.test.ts

@@ -239,6 +239,7 @@ bucket_definitions:
             - SELECT _id as id, description FROM "test_DATA"
       `);
 
+    await db.createCollection('test_DATA');
     await context.replicateSnapshot();
 
     context.startStreaming();
@@ -261,6 +262,7 @@ bucket_definitions:
           data:
             - SELECT _id as id, name, description FROM "test_data"
       `);
+    await db.createCollection('test_data');
 
     await context.replicateSnapshot();
     context.startStreaming();
@@ -371,6 +373,8 @@ bucket_definitions:
           - SELECT _id as id, name, other FROM "test_data"`);
     const { db } = context;
 
+    await db.createCollection('test_data');
+
     await context.replicateSnapshot();
 
     const collection = db.collection('test_data');
@@ -451,6 +455,8 @@ bucket_definitions:
 
     const data = await context.getBucketData('global[]');
     expect(data).toMatchObject([
+      // An extra op here, since this triggers a snapshot in addition to getting the event.
+      test_utils.putOp('test_data', { id: test_id!.toHexString(), description: 'test2' }),
       test_utils.putOp('test_data', { id: test_id!.toHexString(), description: 'test1' }),
       test_utils.putOp('test_data', { id: test_id!.toHexString(), description: 'test2' })
     ]);