HADOOP-18508. S3A: Support parallel integration test runs on same bucket (#5081)

It is now possible to provide a job ID in the maven "job.id" property
hadoop-aws test runs to isolate paths under a the test bucket
under which all tests will be executed.

This will allow independent builds *in different source trees*
to test against the same bucket in parallel, and is designed for
CI testing.

Example:

mvn verify -Dparallel-tests -Droot.tests.enabled=false -Djob.id=1
mvn verify -Droot.tests.enabled=false -Djob.id=2

- Root tests must be be disabled to stop them cleaning up
  the test paths of other test runs.
- Do still regularly run the root tests just to force cleanup
  of the output of any interrupted test suites.

Contributed by Steve Loughran

Author: steveloughran

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/fs/FSMainOperationsBaseTest.java

@@ -102,7 +102,9 @@ public void setUp() throws Exception {
 
   @After
   public void tearDown() throws Exception {
-    fSys.delete(new Path(getAbsoluteTestRootPath(fSys), new Path("test")), true);
+    if (fSys != null) {
+      fSys.delete(new Path(getAbsoluteTestRootPath(fSys), new Path("test")), true);
+    }
   }
 
 
@@ -192,7 +194,7 @@ public void testWorkingDirectory() throws Exception {
 
   @Test
   public void testWDAbsolute() throws IOException {
-    Path absoluteDir = new Path(fSys.getUri() + "/test/existingDir");
+    Path absoluteDir = getTestRootPath(fSys, "test/existingDir");
     fSys.mkdirs(absoluteDir);
     fSys.setWorkingDirectory(absoluteDir);
     Assert.assertEquals(absoluteDir, fSys.getWorkingDirectory());

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/fs/FileContextMainOperationsBaseTest.java

@@ -81,6 +81,12 @@ public abstract class FileContextMainOperationsBaseTest  {
   protected final FileContextTestHelper fileContextTestHelper =
     createFileContextHelper();
 
+  /**
+   * Create the test helper.
+   * Important: this is invoked during the construction of the base class,
+   * so is very brittle.
+   * @return a test helper.
+   */
   protected FileContextTestHelper createFileContextHelper() {
     return new FileContextTestHelper();
   }
@@ -107,7 +113,7 @@ public boolean accept(Path file) {
 
   private static final byte[] data = getFileData(numBlocks,
       getDefaultBlockSize());
-  
+
   @Before
   public void setUp() throws Exception {
     File testBuildData = GenericTestUtils.getRandomizedTestDir();

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/fs/TestFSMainOperationsLocalFileSystem.java

@@ -21,23 +21,13 @@
 import java.io.IOException;
 
 import org.apache.hadoop.conf.Configuration;
-import org.junit.After;
-import org.junit.Assert;
-import org.junit.Before;
-import org.junit.Test;
 
 public class TestFSMainOperationsLocalFileSystem extends FSMainOperationsBaseTest {
 
   @Override
   protected FileSystem createFileSystem() throws IOException {
     return FileSystem.getLocal(new Configuration());
   }
-    
-  @Override
-  @Before
-  public void setUp() throws Exception {
-    super.setUp();
-  }
 
   static Path wd = null;
   @Override
@@ -46,19 +36,5 @@ protected Path getDefaultWorkingDirectory() throws IOException {
       wd = FileSystem.getLocal(new Configuration()).getWorkingDirectory();
     return wd;
   }
-  
-  @Override
-  @After
-  public void tearDown() throws Exception {
-    super.tearDown();
-  }
-  
-  @Test
-  @Override
-  public void testWDAbsolute() throws IOException {
-    Path absoluteDir = getTestRootPath(fSys, "test/existingDir");
-    fSys.mkdirs(absoluteDir);
-    fSys.setWorkingDirectory(absoluteDir);
-    Assert.assertEquals(absoluteDir, fSys.getWorkingDirectory());
-  }
+
 }

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/fs/viewfs/TestFSMainOperationsLocalFileSystem.java

@@ -53,14 +53,5 @@ public void tearDown() throws Exception {
     super.tearDown();
     ViewFileSystemTestSetup.tearDown(this, fcTarget);
   }
-  
-  @Test
-  @Override
-  public void testWDAbsolute() throws IOException {
-    Path absoluteDir = getTestRootPath(fSys, "test/existingDir");
-    fSys.mkdirs(absoluteDir);
-    fSys.setWorkingDirectory(absoluteDir);
-    Assert.assertEquals(absoluteDir, fSys.getWorkingDirectory());
 
-  }
 }

hadoop-tools/hadoop-aws/pom.xml

@@ -56,6 +56,11 @@
 
     <!-- Is prefetch enabled? -->
     <fs.s3a.prefetch.enabled>unset</fs.s3a.prefetch.enabled>
+    <!-- Job ID; allows for parallel jobs on same bucket -->
+    <!-- job.id is used to build the path for tests; default is 00.-->
+    <job.id>00</job.id>
+    <!-- are root tests enabled. Set to false when running parallel jobs on same bucket -->
+    <root.tests.enabled>unset</root.tests.enabled>
   </properties>
 
   <profiles>
@@ -115,14 +120,8 @@
                 <test.build.data>${test.build.data}/${surefire.forkNumber}</test.build.data>
                 <test.build.dir>${test.build.dir}/${surefire.forkNumber}</test.build.dir>
                 <hadoop.tmp.dir>${hadoop.tmp.dir}/${surefire.forkNumber}</hadoop.tmp.dir>
+                <test.unique.fork.id>job-${job.id}-fork-000${surefire.forkNumber}</test.unique.fork.id>
 
-                <!-- Due to a Maven quirk, setting this to just -->
-                <!-- surefire.forkNumber won't do the parameter -->
-                <!-- substitution.  Putting a prefix in front of it like -->
-                <!-- "fork-" makes it work. -->
-                <!-- Important: Those leading 0s are needed to guarantee that -->
-                <!-- trailing three chars are always numeric and unique -->
-                <test.unique.fork.id>fork-000${surefire.forkNumber}</test.unique.fork.id>
                 <!-- Propagate scale parameters -->
                 <fs.s3a.scale.test.enabled>${fs.s3a.scale.test.enabled}</fs.s3a.scale.test.enabled>
                 <fs.s3a.scale.test.huge.filesize>${fs.s3a.scale.test.huge.filesize}</fs.s3a.scale.test.huge.filesize>
@@ -163,7 +162,7 @@
                     <!-- surefire.forkNumber won't do the parameter -->
                     <!-- substitution.  Putting a prefix in front of it like -->
                     <!-- "fork-" makes it work. -->
-                    <test.unique.fork.id>fork-000${surefire.forkNumber}</test.unique.fork.id>
+                    <test.unique.fork.id>job-${job.id}-fork-000${surefire.forkNumber}</test.unique.fork.id>
                     <!-- Propagate scale parameters -->
                     <fs.s3a.scale.test.enabled>${fs.s3a.scale.test.enabled}</fs.s3a.scale.test.enabled>
                     <fs.s3a.scale.test.huge.filesize>${fs.s3a.scale.test.huge.filesize}</fs.s3a.scale.test.huge.filesize>
@@ -174,14 +173,14 @@
                     <test.default.timeout>${test.integration.timeout}</test.default.timeout>
                     <!-- Prefetch -->
                     <fs.s3a.prefetch.enabled>${fs.s3a.prefetch.enabled}</fs.s3a.prefetch.enabled>
+                    <!-- are root tests enabled. Set to false when running parallel jobs on same bucket -->
+                    <fs.s3a.root.tests.enabled>${root.tests.enabled}</fs.s3a.root.tests.enabled>
+
                   </systemPropertyVariables>
                   <!-- Some tests cannot run in parallel.  Tests that cover -->
                   <!-- access to the root directory must run in isolation -->
                   <!-- from anything else that could modify the bucket. -->
-                  <!-- S3A tests that cover multi-part upload must run in -->
-                  <!-- isolation, because the file system is configured to -->
-                  <!-- purge existing multi-part upload data on -->
-                  <!-- initialization.  MiniYARNCluster has not yet been -->
+                  <!-- MiniYARNCluster has not yet been -->
                   <!-- changed to handle parallel test execution gracefully. -->
                   <!-- Exclude all of these tests from parallel execution, -->
                   <!-- and instead run them sequentially in a separate -->
@@ -228,6 +227,9 @@
                     <fs.s3a.directory.marker.audit>${fs.s3a.directory.marker.audit}</fs.s3a.directory.marker.audit>
                     <!-- Prefetch -->
                     <fs.s3a.prefetch.enabled>${fs.s3a.prefetch.enabled}</fs.s3a.prefetch.enabled>
+                    <!-- are root tests enabled. Set to false when running parallel jobs on same bucket -->
+                    <fs.s3a.root.tests.enabled>${root.tests.enabled}</fs.s3a.root.tests.enabled>
+                    <test.unique.fork.id>job-${job.id}</test.unique.fork.id>
                   </systemPropertyVariables>
                   <!-- Do a sequential run for tests that cannot handle -->
                   <!-- parallel execution. -->
@@ -289,6 +291,7 @@
                     <fs.s3a.directory.marker.audit>${fs.s3a.directory.marker.audit}</fs.s3a.directory.marker.audit>
                     <!-- Prefetch -->
                     <fs.s3a.prefetch.enabled>${fs.s3a.prefetch.enabled}</fs.s3a.prefetch.enabled>
+                    <test.unique.fork.id>job-${job.id}</test.unique.fork.id>
                   </systemPropertyVariables>
                   <forkedProcessTimeoutInSeconds>${fs.s3a.scale.test.timeout}</forkedProcessTimeoutInSeconds>
                 </configuration>

hadoop-tools/hadoop-aws/src/site/markdown/tools/hadoop-aws/testing.md

@@ -43,7 +43,7 @@ is a more specific lie and harder to make. And, if you get caught out: you
 lose all credibility with the project.
 
 You don't need to test from a VM within the AWS infrastructure; with the
-`-Dparallel=tests` option the non-scale tests complete in under ten minutes.
+`-Dparallel-tests` option the non-scale tests complete in under twenty minutes.
 Because the tests clean up after themselves, they are also designed to be low
 cost. It's neither hard nor expensive to run the tests; if you can't,
 there's no guarantee your patch works. The reviewers have enough to do, and
@@ -539,12 +539,51 @@ Otherwise, set a large timeout in `fs.s3a.scale.test.timeout`
 The tests are executed in an order to only clean up created files after
 the end of all the tests. If the tests are interrupted, the test data will remain.
 
+## <a name="CI"/> Testing through continuous integration
+
+### Parallel CI builds.
+For CI testing of the module, including the integration tests,
+it is generally necessary to support testing multiple PRs simultaneously.
+
+To do this
+1. A job ID must be supplied in the `job.id` property, so each job works on an isolated directory
+   tree. This should be a number or unique string, which will be used within a path element, so
+   must only contain characters valid in an S3/hadoop path element.
+2. Root directory tests need to be disabled by setting `fs.s3a.root.tests.enabled` to
+   `false`, either in the command line to maven or in the XML configurations.
+
+```
+mvn verify -T 1C -Dparallel-tests -DtestsThreadCount=14 -Dscale -Dfs.s3a.root.tests.enabled=false -Djob.id=001
+```
+
+This parallel execution feature is only for isolated builds sharing a single S3 bucket; it does
+not support parallel builds and tests from the same local source tree.
+
+Without the root tests being executed, set up a scheduled job to purge the test bucket of all
+data on a regular basis, to keep costs down.
+The easiest way to do this is to have a bucket lifecycle rule for the bucket to delete all files more than a few days old,
+alongside one to abort all pending uploads more than 24h old.
+
+
+### Securing CI builds
+
+It's clearly unsafe to have CI infrastructure testing PRs submitted to apache github account
+with AWS credentials -which is why it isn't done by the Yetus-initiated builds.
+
+Anyone doing this privately should:
+* Review incoming patches before triggering the tests.
+* Have a dedicated IAM role with restricted access to the test bucket, any KMS keys used, and the
+  external bucket containing the CSV test file.
+* Have a build process which generates short-lived session credentials for this role.
+* Run the tests in an EC2 VM/container which collects the restricted IAM credentials
+  from the IAM instance/container credentials provider.
+
 ## <a name="load"></a> Load tests.
 
-Some are designed to overload AWS services with more
+Some tests are designed to overload AWS services with more
 requests per second than an AWS account is permitted.
 
-The operation of these test maybe observable to other users of the same
+The operation of these tests may be observable to other users of the same
 account -especially if they are working in the AWS region to which the
 tests are targeted.
 
@@ -556,6 +595,10 @@ They do not run automatically: they must be explicitly run from the command line
 
 Look in the source for these and reads the Javadocs before executing.
 
+Note: one fear here was that asking for two many session/role credentials in a short period
+of time would actually lock an account out of a region. It doesn't: it simply triggers
+throttling of STS requests.
+
 ## <a name="alternate_s3"></a> Testing against non-AWS S3 Stores.
 
 The S3A filesystem is designed to work with S3 stores which implement

hadoop-tools/hadoop-aws/src/test/java/org/apache/hadoop/fs/contract/s3a/ITestS3AContractRootDir.java

@@ -27,6 +27,8 @@
 import org.apache.hadoop.fs.contract.AbstractFSContract;
 import org.apache.hadoop.fs.s3a.S3AFileSystem;
 
+import static org.apache.hadoop.fs.s3a.S3ATestUtils.maybeSkipRootTests;
+
 /**
  * root dir operations against an S3 bucket.
  */
@@ -36,6 +38,12 @@ public class ITestS3AContractRootDir extends
   private static final Logger LOG =
       LoggerFactory.getLogger(ITestS3AContractRootDir.class);
 
+  @Override
+  public void setup() throws Exception {
+    super.setup();
+    maybeSkipRootTests(getFileSystem().getConf());
+  }
+
   @Override
   protected AbstractFSContract createContract(Configuration conf) {
     return new S3AContract(conf);

hadoop-tools/hadoop-aws/src/test/java/org/apache/hadoop/fs/s3a/ITestS3AConfiguration.java

@@ -386,8 +386,10 @@ public void shouldBeAbleToSwitchOnS3PathStyleAccessViaConfigProperty()
           s3Configuration.pathStyleAccessEnabled());
       byte[] file = ContractTestUtils.toAsciiByteArray("test file");
       ContractTestUtils.writeAndRead(fs,
-          new Path("/path/style/access/testFile"), file, file.length,
-              (int) conf.getLongBytes(Constants.FS_S3A_BLOCK_SIZE, file.length), false, true);
+          createTestPath(new Path("/path/style/access/testFile")),
+          file, file.length,
+          (int) conf.getLongBytes(Constants.FS_S3A_BLOCK_SIZE, file.length),
+          false, true);
     } catch (final AWSRedirectException e) {
       LOG.error("Caught exception: ", e);
       // Catch/pass standard path style access behaviour when live bucket