fix: fix error handling in segment reader and add an integration test

sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigtable/BigtableServiceImpl.java

@@ -387,6 +387,11 @@ public void onResponse(Row response) {
                   currentByteSize += response.getSerializedSize();
                   rows.add(response);
                   if (currentByteSize > maxSegmentByteSize) {
+                    LOG.debug(
+                        "Reached maxSegmentByteSize, cancelling the stream. currentByteSize is {}, maxSegmentByteSize is {}, read rows {}",
+                        currentByteSize,
+                        maxSegmentByteSize,
+                        rows.size());
                     byteLimitReached = true;
                     controller.cancel();
                     return;
@@ -395,14 +400,25 @@ public void onResponse(Row response) {
 
                 @Override
                 public void onError(Throwable t) {
-                  future.setException(t);
+                  if (byteLimitReached) {
+                    // When the byte limit is reached we cancel the stream in onResponse.
+                    // In this case we don't want to fail the request with cancellation
+                    // exception. Instead, we construct the next request.
+                    onComplete();
+                  } else {
+                    future.setException(t);
+                  }
                 }
 
                 @Override
                 public void onComplete() {
                   ReadRowsRequest nextNextRequest = null;
 
-                  // When requested rows < limit, the current request will be the last
+                  // Only schedule the next segment fetch when there's a possibility of more
+                  // data to read. We know there might be more data when the current segment
+                  // ended with the artificial byte limit or the row limit.
+                  // If the RPC ended without hitting the byte limit or row limit, we know
+                  // there's no more data to read and nextNextRequest would be null.
                   if (byteLimitReached || rows.size() == nextRequest.getRowsLimit()) {
                     nextNextRequest =
                         truncateRequest(nextRequest, rows.get(rows.size() - 1).getKey());

sdks/java/io/google-cloud-platform/src/test/java/org/apache/beam/sdk/io/gcp/bigtable/BigtableReadIT.java

@@ -17,33 +17,88 @@
  */
 package org.apache.beam.sdk.io.gcp.bigtable;
 
+import com.google.cloud.bigtable.admin.v2.BigtableTableAdminClient;
+import com.google.cloud.bigtable.admin.v2.BigtableTableAdminSettings;
+import com.google.cloud.bigtable.admin.v2.models.CreateTableRequest;
 import com.google.cloud.bigtable.config.BigtableOptions;
+import com.google.cloud.bigtable.data.v2.BigtableDataClient;
+import com.google.cloud.bigtable.data.v2.BigtableDataSettings;
+import com.google.cloud.bigtable.data.v2.models.RowMutation;
+import java.io.IOException;
+import java.util.Date;
+import org.apache.beam.repackaged.core.org.apache.commons.lang3.StringUtils;
 import org.apache.beam.sdk.Pipeline;
 import org.apache.beam.sdk.extensions.gcp.options.GcpOptions;
 import org.apache.beam.sdk.options.PipelineOptionsFactory;
 import org.apache.beam.sdk.testing.PAssert;
 import org.apache.beam.sdk.testing.TestPipeline;
 import org.apache.beam.sdk.transforms.Count;
 import org.apache.beam.sdk.values.PCollection;
+import org.junit.After;
+import org.junit.Before;
 import org.junit.Test;
 import org.junit.runner.RunWith;
 import org.junit.runners.JUnit4;
 
 /** End-to-end tests of BigtableRead. */
 @RunWith(JUnit4.class)
 public class BigtableReadIT {
+  private static final String COLUMN_FAMILY_NAME = "cf";
 
-  @Test
-  public void testE2EBigtableRead() throws Exception {
+  private String project;
+
+  private BigtableTestOptions options;
+  private String tableId = String.format("BigtableReadIT-%tF-%<tH-%<tM-%<tS-%<tL", new Date());
+
+  private BigtableDataClient client;
+  private BigtableTableAdminClient tableAdminClient;
+
+  @Before
+  public void setup() throws IOException {
     PipelineOptionsFactory.register(BigtableTestOptions.class);
-    BigtableTestOptions options =
-        TestPipeline.testingPipelineOptions().as(BigtableTestOptions.class);
 
-    String project = options.getBigtableProject();
+    options = TestPipeline.testingPipelineOptions().as(BigtableTestOptions.class);
+
+    project = options.getBigtableProject();
+
     if (project.equals("")) {
       project = options.as(GcpOptions.class).getProject();
     }
 
+    BigtableDataSettings veneerSettings =
+        BigtableDataSettings.newBuilder()
+            .setProjectId(project)
+            .setInstanceId(options.getInstanceId())
+            .build();
+
+    BigtableTableAdminSettings adminSettings =
+        BigtableTableAdminSettings.newBuilder()
+            .setProjectId(project)
+            .setInstanceId(options.getInstanceId())
+            .build();
+
+    client = BigtableDataClient.create(veneerSettings);
+    tableAdminClient = BigtableTableAdminClient.create(adminSettings);
+  }
+
+  @After
+  public void tearDown() {
+    if (tableAdminClient != null) {
+      try {
+        tableAdminClient.deleteTable(tableId);
+      } catch (Exception e) {
+        // ignore exceptions
+      }
+      tableAdminClient.close();
+    }
+
+    if (client != null) {
+      client.close();
+    }
+  }
+
+  @Test
+  public void testE2EBigtableRead() {
     BigtableOptions.Builder bigtableOptionsBuilder =
         new BigtableOptions.Builder().setProjectId(project).setInstanceId(options.getInstanceId());
 
@@ -57,4 +112,32 @@ public void testE2EBigtableRead() throws Exception {
     PAssert.thatSingleton(count).isEqualTo(numRows);
     p.run();
   }
+
+  @Test
+  public void testE2EBigtableSegmentRead() {
+    tableAdminClient.createTable(CreateTableRequest.of(tableId).addFamily(COLUMN_FAMILY_NAME));
+
+    final long numRows = 20L;
+    String value = StringUtils.repeat("v", 100 * 1000 * 1000);
+    // populate a table with large rows, each row is 100 MB. This will make the pipeline reach
+    // segment reader memory limit with 8 rows when running on ec2-standard.
+    for (int i = 0; i < numRows; i++) {
+      client.mutateRow(
+          RowMutation.create(tableId, "key-" + i).setCell(COLUMN_FAMILY_NAME, "q", value));
+    }
+
+    BigtableOptions.Builder bigtableOptionsBuilder =
+        new BigtableOptions.Builder().setProjectId(project).setInstanceId(options.getInstanceId());
+
+    Pipeline p = Pipeline.create(options);
+    PCollection<Long> count =
+        p.apply(
+                BigtableIO.read()
+                    .withBigtableOptions(bigtableOptionsBuilder)
+                    .withTableId(tableId)
+                    .withMaxBufferElementCount(10))
+            .apply(Count.globally());
+    PAssert.thatSingleton(count).isEqualTo(numRows);
+    p.run();
+  }
 }