[SPARK-25692][TEST] Increase timeout in fetchBothChunks test
authorDongjoon Hyun <dongjoon@apache.org>
Sat, 12 Jan 2019 01:22:05 +0000 (17:22 -0800)
committerDongjoon Hyun <dongjoon@apache.org>
Sat, 12 Jan 2019 01:22:05 +0000 (17:22 -0800)
## What changes were proposed in this pull request?

`ChunkFetchIntegrationSuite.fetchBothChunks` fails frequently due to timeout in Apache Spark Jenkins environments.

```scala
org.apache.spark.network.ChunkFetchIntegrationSuite
[ERROR] fetchBothChunks(org.apache.spark.network.ChunkFetchIntegrationSuite)
Time elapsed: 5.015 s  <<< FAILURE!
java.lang.AssertionError: Timeout getting response from the server
at org.apache.spark.network.ChunkFetchIntegrationSuite.fetchChunks(ChunkFetchIntegrationSuite.java:176)
at org.apache.spark.network.ChunkFetchIntegrationSuite.fetchBothChunks(ChunkFetchIntegrationSuite.java:210)
```

The followings are the recent failures on `amp-jenkins-worker-05`. The timeout seems to be too sensitive in low-end machines. This PR increases the timeout from 5 seconds to 60 seconds in order to be more robust.

- [master 5856](https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-2.7/5856/)
- [master 5837](https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-2.7/5837/testReport)
- [master 5835](https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-2.7/5835/testReport)
- [master 5829](https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-2.7/5829/testReport)
- [master 5828](https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-2.7/5828/testReport)
- [master 5822](https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-2.7/5822/testReport)
- [master 5814](https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-2.7/5814/testReport)

- [SparkPullRequestBuilder 100784](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/100784/consoleFull)

- [SparkPullRequestBuilder 100785](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/100785/consoleFull)

- [SparkPullRequestBuilder 100787](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/100787/consoleFull)

- [SparkPullRequestBuilder 100788](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/100788/consoleFull)

## How was this patch tested?

N/A (Monitor the Jenkins on `amp-jenkins-worker-05` machine)

Closes #23522 from dongjoon-hyun/SPARK-25692.

Authored-by: Dongjoon Hyun <dongjoon@apache.org>
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
common/network-common/src/test/java/org/apache/spark/network/ChunkFetchIntegrationSuite.java

index 37a8664..ab4dd04 100644 (file)
@@ -172,7 +172,7 @@ public class ChunkFetchIntegrationSuite {
       for (int chunkIndex : chunkIndices) {
         client.fetchChunk(STREAM_ID, chunkIndex, callback);
       }
-      if (!sem.tryAcquire(chunkIndices.size(), 5, TimeUnit.SECONDS)) {
+      if (!sem.tryAcquire(chunkIndices.size(), 60, TimeUnit.SECONDS)) {
         fail("Timeout getting response from the server");
       }
     }