Skip to content

armgit / floci

Public Imported from GitHub
0

Light, fluffy, and always free - The AWS Local Emulator alternative

fix(compat): retry RedshiftTest's SDK calls past a brief connection-refused hiccup (#3216)

Olamide Kosile974d357f88ab
diff --git a/compatibility-tests/sdk-test-java/src/test/java/io/github/hectorvent/floci/compat/RedshiftTest.java b/compatibility-tests/sdk-test-java/src/test/java/io/github/hectorvent/floci/compat/RedshiftTest.java
index db816e09e1..69bd877dc2 100644
--- a/compatibility-tests/sdk-test-java/src/test/java/io/github/hectorvent/floci/compat/RedshiftTest.java
+++ b/compatibility-tests/sdk-test-java/src/test/java/io/github/hectorvent/floci/compat/RedshiftTest.java
@@ -5,6 +5,7 @@ import org.junit.jupiter.api.MethodOrderer;
import org.junit.jupiter.api.Order;
import org.junit.jupiter.api.Test;
import org.junit.jupiter.api.TestMethodOrder;
+import software.amazon.awssdk.core.exception.SdkClientException;
import software.amazon.awssdk.services.redshift.RedshiftClient;
import software.amazon.awssdk.services.redshift.model.Cluster;
import software.amazon.awssdk.services.redshift.model.CreateClusterRequest;
@@ -19,11 +20,16 @@ import software.amazon.awssdk.services.s3.model.GetObjectRequest;
import software.amazon.awssdk.services.s3.model.ListObjectsV2Request;
import software.amazon.awssdk.services.s3.model.ListObjectsV2Response;
+import java.net.ConnectException;
import java.nio.charset.StandardCharsets;
import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.Statement;
+import java.time.Duration;
+import java.time.Instant;
+import java.util.function.Supplier;
+
import static org.junit.jupiter.api.Assertions.assertEquals;
import static org.junit.jupiter.api.Assertions.assertNotNull;
import static org.junit.jupiter.api.Assertions.assertTrue;
@@ -35,23 +41,61 @@ public class RedshiftTest {
return TestFixtures.redshiftClient();
}
+ // github.com/floci-io/floci/issues/2789: this class's first request intermittently hits
+ // "Connection refused" even after the AWS SDK's own 4 built-in attempts (whose backoff totals
+ // under a second) and even after CI's "wait for floci to be ready" step reports success - a
+ // brief port-publish/network-alias hiccup between the test runner and floci containers that
+ // self-resolves almost immediately. Retrying the whole call, not just relying on the SDK's
+ // built-in attempts, gives it long enough to clear.
+ //
+ // Only a pre-handshake ConnectException is worth retrying (review, PR #3216): a broader catch
+ // of SdkClientException would also retry failures that can happen after floci already processed
+ // the request, e.g. a response reset mid-stream on createCluster/deleteCluster - a retry there
+ // resubmits and gets back a real ClusterAlreadyExists/ClusterNotFound, which masks the original
+ // failure behind an unrelated one instead of fixing the flake.
+ private static <T> T withRetry(Supplier<T> action) throws InterruptedException {
+ Instant deadline = Instant.now().plus(Duration.ofSeconds(30));
+ SdkClientException last = null;
+ while (Instant.now().isBefore(deadline)) {
+ try {
+ return action.get();
+ } catch (SdkClientException e) {
+ if (!isConnectionRefused(e)) {
+ throw e;
+ }
+ last = e;
+ Thread.sleep(1000);
+ }
+ }
+ throw last;
+ }
+
+ private static boolean isConnectionRefused(Throwable e) {
+ for (Throwable cause = e; cause != null; cause = cause.getCause()) {
+ if (cause instanceof ConnectException) {
+ return true;
+ }
+ }
+ return false;
+ }
+
@Test
@Order(1)
public void testCreateCluster() throws Exception {
RedshiftClient client = getClient();
- CreateClusterResponse res = client.createCluster(CreateClusterRequest.builder()
+ CreateClusterResponse res = withRetry(() -> client.createCluster(CreateClusterRequest.builder()
.clusterIdentifier("test-cluster")
.nodeType("dc2.large")
.masterUsername("admin")
.masterUserPassword("Password123")
- .build());
-
+ .build()));
+
assertEquals("test-cluster", res.cluster().clusterIdentifier());
-
- DescribeClustersResponse describeRes = client.describeClusters(DescribeClustersRequest.builder()
+
+ DescribeClustersResponse describeRes = withRetry(() -> client.describeClusters(DescribeClustersRequest.builder()
.clusterIdentifier("test-cluster")
- .build());
-
+ .build()));
+
Cluster cluster = describeRes.clusters().get(0);
assertEquals("test-cluster", cluster.clusterIdentifier());
// Terraform's AWS provider polls these two on create and validates them on read (issue #3098).
@@ -108,11 +152,11 @@ public class RedshiftTest {
@Test
@Order(3)
- public void testDeleteCluster() {
+ public void testDeleteCluster() throws Exception {
RedshiftClient client = getClient();
- DeleteClusterResponse res = client.deleteCluster(DeleteClusterRequest.builder()
+ DeleteClusterResponse res = withRetry(() -> client.deleteCluster(DeleteClusterRequest.builder()
.clusterIdentifier("test-cluster")
- .build());
+ .build()));
assertEquals("test-cluster", res.cluster().clusterIdentifier());
assertEquals("deleting", res.cluster().clusterStatus());
}