Testing the S3A filesystem client and its features, including S3Guard

The submitter of any patch is required to run all the integration tests and declare which S3 region/implementation they used.

This is important: patches which do not include this declaration will be ignored

This policy has proven to be the only mechanism to guarantee full regression testing of code changes. Why the declaration of region? Two reasons

1. It helps us identify regressions which only surface against specific endpoints or third-party implementations of the S3 protocol.
2. It forces the submitters to be more honest about their testing. It’s easy to lie, “yes, I tested this”. To say “yes, I tested this against S3 US-west” 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. 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 don’t have the time to do these tests, especially as every failure will simply make for a slow iterative development.

Please: run the tests. And if you don’t, we are sorry for declining your patch, but we have to.

What if there’s an intermittent failure of a test?

Some of the tests do fail intermittently, especially in parallel runs. If this happens, try to run the test on its own to see if the test succeeds.

If it still fails, include this fact in your declaration. We know some tests are intermittently unreliable.

What if the tests are timing out or failing over my network connection?

The tests and the S3A client are designed to be configurable for different timeouts. If you are seeing problems and this configuration isn’t working, that’s a sign of the configuration mechanism isn’t complete. If it’s happening in the production code, that could be a sign of a problem which may surface over long-haul connections. Please help us identify and fix these problems — especially as you are the one best placed to verify the fixes work.

File core-site.xml

This file pre-exists and sources the configurations created under auth-keys.xml.

For most purposes you will not need to edit this file unless you need to apply a specific, non-default property change during the tests.

File auth-keys.xml

The presence of this file triggers the testing of the S3 classes.

Without this file, none of the integration tests in this module will be executed.

The XML file must contain all the ID/key information needed to connect each of the filesystem clients to the object stores, and a URL for each filesystem for its testing.

1. test.fs.s3a.name : the URL of the bucket for S3a tests
2. fs.contract.test.fs.s3a : the URL of the bucket for S3a filesystem contract tests

The contents of the bucket will be destroyed during the test process: do not use the bucket for any purpose other than testing. Furthermore, for s3a, all in-progress multi-part uploads to the bucket will be aborted at the start of a test (by forcing fs.s3a.multipart.purge=true) to clean up the temporary state of previously failed tests.

Example:

<configuration>

  <property>
    <name>test.fs.s3a.name</name>
    <value>s3a://test-aws-s3a/</value>
  </property>

  <property>
    <name>fs.contract.test.fs.s3a</name>
    <value>${test.fs.s3a.name}</value>
  </property>

  <property>
    <name>fs.s3a.access.key</name>
    <description>AWS access key ID. Omit for IAM role-based authentication.</description>
    <value>DONOTCOMMITTHISKEYTOSCM</value>
  </property>

  <property>
    <name>fs.s3a.secret.key</name>
    <description>AWS secret key. Omit for IAM role-based authentication.</description>
    <value>DONOTEVERSHARETHISSECRETKEY!</value>
  </property>

  <property>
    <name>test.sts.endpoint</name>
    <description>Specific endpoint to use for STS requests.</description>
    <value>sts.amazonaws.com</value>
  </property>

</configuration>

Configuring S3a Encryption

For S3a encryption tests to run correctly, the fs.s3a.server-side-encryption.key must be configured in the s3a contract xml file or auth-keys.xml file with a AWS KMS encryption key arn as this value is different for each AWS KMS. Please note this KMS key should be created in the same region as your S3 bucket. Otherwise, you may get KMS.NotFoundException.

Example:

<property>
  <name>fs.s3a.server-side-encryption.key</name>
  <value>arn:aws:kms:us-west-2:360379543683:key/071a86ff-8881-4ba0-9230-95af6d01ca01</value>
</property>

You can also force all the tests to run with a specific SSE encryption method by configuring the property fs.s3a.server-side-encryption-algorithm in the s3a contract file.

Default Encryption

Buckets can be configured with default encryption on the AWS side. Some S3AFileSystem tests are skipped when default encryption is enabled due to unpredictability in how ETags are generated.

Testing against different regions

S3A can connect to different regions —the tests support this. Simply define the target region in auth-keys.xml.

<property>
  <name>fs.s3a.endpoint</name>
  <value>s3.eu-central-1.amazonaws.com</value>
</property>

Alternatively you can use endpoints defined in core-site.xml.

<property>
  <name>fs.s3a.endpoint</name>
  <value>${frankfurt.endpoint}</value>
</property>

This is used for all tests expect for scale tests using a Public CSV.gz file (see below)

CSV Data Tests

The TestS3AInputStreamPerformance tests require read access to a multi-MB text file. The default file for these tests is one published by amazon, s3a://landsat-pds.s3.amazonaws.com/scene_list.gz. This is a gzipped CSV index of other files which amazon serves for open use.

The path to this object is set in the option fs.s3a.scale.test.csvfile,

<property>
  <name>fs.s3a.scale.test.csvfile</name>
  <value>s3a://landsat-pds/scene_list.gz</value>
</property>

1. If the option is not overridden, the default value is used. This is hosted in Amazon’s US-east datacenter.
2. If fs.s3a.scale.test.csvfile is empty, tests which require it will be skipped.
3. If the data cannot be read for any reason then the test will fail.
4. If the property is set to a different path, then that data must be readable and “sufficiently” large.

(the reason the space or newline is needed is to add “an empty entry”; an empty <value/> would be considered undefined and pick up the default)

Of using a test file in an S3 region requiring a different endpoint value set in fs.s3a.endpoint, a bucket-specific endpoint must be defined. For the default test dataset, hosted in the landsat-pds bucket, this is:

<property>
  <name>fs.s3a.bucket.landsat-pds.endpoint</name>
  <value>s3.amazonaws.com</value>
  <description>The endpoint for s3a://landsat-pds URLs</description>
</property>

Enabling auditing of markers

To enable an audit of the output directory of every test suite, enable the option fs.s3a.directory.marker.audit

-Dfs.s3a.directory.marker.audit=true

When set, if the marker policy is to delete markers under the test output directory, then the marker tool audit command will be run. This will fail if a marker was found.

This adds extra overhead to every operation, but helps verify that the connector is not keeping markers where it needs to be deleting them -and hence backwards compatibility is maintained.

Enabling the Scale Tests

The tests are enabled if the scale property is set in the maven build this can be done regardless of whether or not the parallel test profile is used

mvn verify -Dscale

mvn verify -Dparallel-tests -Dscale -DtestsThreadCount=8

The most bandwidth intensive tests (those which upload data) always run sequentially; those which are slow due to HTTPS setup costs or server-side actions are included in the set of parallelized tests.

Tuning scale options from Maven

Some of the tests can be tuned from the maven build or from the configuration file used to run the tests.

mvn verify -Dparallel-tests -Dscale -DtestsThreadCount=8 -Dfs.s3a.scale.test.huge.filesize=128M

The algorithm is

1. The value is queried from the configuration file, using a default value if it is not set.
2. The value is queried from the JVM System Properties, where it is passed down by maven.
3. If the system property is null, an empty string, or it has the value unset, then the configuration value is used. The unset option is used to work round a quirk in maven property propagation.

Only a few properties can be set this way; more will be added.

The file and partition sizes are numeric values with a k/m/g/t/p suffix depending on the desired size. For example: 128M, 128m, 2G, 2G, 4T or even 1P.

Scale test configuration options

Some scale tests perform multiple operations (such as creating many directories).

The exact number of operations to perform is configurable in the option scale.test.operation.count

<property>
  <name>scale.test.operation.count</name>
  <value>10</value>
</property>

Larger values generate more load, and are recommended when testing locally, or in batch runs.

Smaller values results in faster test runs, especially when the object store is a long way away.

Operations which work on directories have a separate option: this controls the width and depth of tests creating recursive directories. Larger values create exponentially more directories, with consequent performance impact.

<property>
  <name>scale.test.directory.count</name>
  <value>2</value>
</property>

DistCp tests targeting S3A support a configurable file size. The default is 10 MB, but the configuration value is expressed in KB so that it can be tuned smaller to achieve faster test runs.

<property>
  <name>scale.test.distcp.file.size.kb</name>
  <value>10240</value>
</property>

S3A specific scale test properties are

fs.s3a.scale.test.huge.filesize: size in MB for “Huge file tests”.

The Huge File tests validate S3A’s ability to handle large files —the property fs.s3a.scale.test.huge.filesize declares the file size to use.

<property>
  <name>fs.s3a.scale.test.huge.filesize</name>
  <value>200M</value>
</property>

Amazon S3 handles files larger than 5GB differently than smaller ones. Setting the huge filesize to a number greater than that) validates support for huge files.

<property>
  <name>fs.s3a.scale.test.huge.filesize</name>
  <value>6G</value>
</property>

Tests at this scale are slow: they are best executed from hosts running in the cloud infrastructure where the S3 endpoint is based. Otherwise, set a large timeout in fs.s3a.scale.test.timeout

<property>
  <name>fs.s3a.scale.test.timeout</name>
  <value>432000</value>
</property>

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.

Disabling the encryption tests

If the endpoint doesn’t support server-side-encryption, these will fail. They can be turned off.

<property>
  <name>test.fs.s3a.encryption.enabled</name>
  <value>false</value>
</property>

Encryption is only used for those specific test suites with Encryption in their classname.

Configuring the CSV file read tests**

To test on alternate infrastructures supporting the same APIs, the option fs.s3a.scale.test.csvfile must either be set to " ", or an object of at least 10MB is uploaded to the object store, and the fs.s3a.scale.test.csvfile option set to its path.

<property>
  <name>fs.s3a.scale.test.csvfile</name>
  <value> </value>
</property>

(yes, the space is necessary. The Hadoop Configuration class treats an empty value as “do not override the default”).

Turning off S3 Select

The S3 select tests are skipped when the S3 endpoint doesn’t support S3 Select.

<property>
  <name>fs.s3a.select.enabled</name>
  <value>false</value>
</property>

If your endpoint doesn’t support that feature, this option should be in your core-site.xml file, so that trying to use S3 select fails fast with a meaningful error (“S3 Select not supported”) rather than a generic Bad Request exception.

Testing Session Credentials

Some tests requests a session credentials and assumed role credentials from the AWS Secure Token Service, then use them to authenticate with S3 either directly or via delegation tokens.

If an S3 implementation does not support STS, then these functional test cases must be disabled:

<property>
  <name>test.fs.s3a.sts.enabled</name>
  <value>false</value>
</property>

These tests request a temporary set of credentials from the STS service endpoint. An alternate endpoint may be defined in fs.s3a.assumed.role.sts.endpoint. If this is set, a delegation token region must also be defined: in fs.s3a.assumed.role.sts.endpoint.region. This is useful not just for testing alternative infrastructures, but to reduce latency on tests executed away from the central service.

<property>
  <name>fs.s3a.delegation.token.endpoint</name>
  <value>fs.s3a.assumed.role.sts.endpoint</value>
</property>
<property>
  <name>fs.s3a.assumed.role.sts.endpoint.region</name>
  <value>eu-west-2</value>
</property>

The default is ""; meaning “use the amazon default endpoint” (sts.amazonaws.com).

Consult the AWS documentation for the full list of locations.

Subclasses Existing Shared Base Classes

Extend AbstractS3ATestBase or AbstractSTestS3AHugeFiles unless justifiable. These set things up for testing against the object stores, provide good threadnames, help generate isolated paths, and for AbstractSTestS3AHugeFiles subclasses, only run if -Dscale is set.

Key features of AbstractS3ATestBase

• getFileSystem() returns the S3A Filesystem bonded to the contract test Filesystem defined in fs.s3a.contract.test
• will automatically skip all tests if that URL is unset.
• Extends AbstractFSContractTestBase and Assert for all their methods.

Having shared base classes may help reduce future maintenance too. Please use them/

Secure

Don’t ever log credentials. The credential tests go out of their way to not provide meaningful logs or assertion messages precisely to avoid this.

Efficient of Time and Money

This means efficient in test setup/teardown, and, ideally, making use of existing public datasets to save setup time and tester cost.

Strategies of particular note are:

1. ITestS3ADirectoryPerformance: a single test case sets up the directory tree then performs different list operations, measuring the time taken.
2. AbstractSTestS3AHugeFiles: marks the test suite as @FixMethodOrder(MethodSorters.NAME_ASCENDING) then orders the test cases such that each test case expects the previous test to have completed (here: uploaded a file, renamed a file, …). This provides for independent tests in the reports, yet still permits an ordered sequence of operations. Do note the use of Assume.assume() to detect when the preconditions for a single test case are not met, hence, the tests become skipped, rather than fail with a trace which is really a false alarm.

The ordered test case mechanism of AbstractSTestS3AHugeFiles is probably the most elegant way of chaining test setup/teardown.

Regarding reusing existing data, we tend to use the landsat archive of AWS US-East for our testing of input stream operations. This doesn’t work against other regions, or with third party S3 implementations. Thus the URL can be overridden for testing elsewhere.

Works With Other S3 Endpoints

Don’t assume AWS S3 US-East only, do allow for working with external S3 implementations. Those may be behind the latest S3 API features, not support encryption, session APIs, etc.

They won’t have the same CSV test files as some of the input tests rely on. Look at ITestS3AInputStreamPerformance to see how tests can be written to support the declaration of a specific large test file on alternate filesystems.

Works Over Long-haul Links

As well as making file size and operation counts scalable, this includes making test timeouts adequate. The Scale tests make this configurable; it’s hard coded to ten minutes in AbstractS3ATestBase(); subclasses can change this by overriding getTestTimeoutMillis().

Equally importantly: support proxies, as some testers need them.

Provides Diagnostics and timing information

1. Give threads useful names.
2. Create logs, log things. Know that the S3AFileSystem and its input and output streams all provide useful statistics in their {{toString()}} calls; logging them is useful on its own.
3. you can use AbstractS3ATestBase.describe(format-stringm, args) here.; it adds some newlines so as to be easier to spot.
4. Use ContractTestUtils.NanoTimer to measure the duration of operations, and log the output.

Fails Meaningfully

The ContractTestUtils class contains a whole set of assertions for making statements about the expected state of a filesystem, e.g. assertPathExists(FS, path), assertPathDoesNotExists(FS, path), and others. These do their best to provide meaningful diagnostics on failures (e.g. directory listings, file status, …), so help make failures easier to understand.

At the very least, do not use assertTrue() or assertFalse() without including error messages.

Sets up its filesystem and checks for those settings

Tests can overrun createConfiguration() to add new options to the configuration file for the S3A Filesystem instance used in their tests.

However, filesystem caching may mean that a test suite may get a cached instance created with an different configuration. For tests which don’t need specific configurations caching is good: it reduces test setup time.

For those tests which do need unique options (encryption, magic files), things can break, and they will do so in hard-to-replicate ways.

Use S3ATestUtils.disableFilesystemCaching(conf) to disable caching when modifying the config. As an example from AbstractTestS3AEncryption:

@Override
protected Configuration createConfiguration() {
  Configuration conf = super.createConfiguration();
  S3ATestUtils.disableFilesystemCaching(conf);
  conf.set(Constants.SERVER_SIDE_ENCRYPTION_ALGORITHM,
          getSSEAlgorithm().getMethod());
  return conf;
}

Then verify in the setup method or test cases that their filesystem actually has the desired feature (fs.getConf().getProperty(...)). This not only catches filesystem reuse problems, it catches the situation where the filesystem configuration in auth-keys.xml has explicit per-bucket settings which override the test suite’s general option settings.

Cleans Up Afterwards

Keeps costs down.

1. Do not only cleanup if a test case completes successfully; test suite teardown must do it.
2. That teardown code must check for the filesystem and other fields being null before the cleanup. Why? If test setup fails, the teardown methods still get called.

Works Reliably

We really appreciate this — you will too.

Runs in parallel unless this is unworkable.

Tests must be designed to run in parallel with other tests, all working with the same shared S3 bucket. This means

• Uses relative and JVM-fork-unique paths provided by the method AbstractFSContractTestBase.path(String filepath).
• Doesn’t manipulate the root directory or make assertions about its contents (for example: delete its contents and assert that it is now empty).
• Doesn’t have a specific requirement of all active clients of the bucket (example: SSE-C tests which require all files, even directory markers, to be encrypted with the same key).
• Doesn’t use so much bandwidth that all other tests will be starved of IO and start timing out (e.g. the scale tests).

Tests such as these can only be run as sequential tests. When adding one, exclude it in the POM file. from the parallel failsafe run and add to the sequential one afterwards. The IO heavy ones must also be subclasses of S3AScaleTestBase and so only run if the system/maven property fs.s3a.scale.test.enabled is true.

Keeping AWS Costs down

Most of the base S3 tests are designed to use public AWS data (the landsat-pds bucket) for read IO, so you don’t have to pay for bytes downloaded or long term storage costs. The scale tests do work with more data so will cost more as well as generally take more time to execute.

You are however billed for

1. Data left in S3 after test runs.
2. DynamoDB capacity reserved by S3Guard tables.
3. HTTP operations on files (HEAD, LIST, GET).
4. In-progress multipart uploads from bulk IO or S3A committer tests.
5. Encryption/decryption using AWS KMS keys.

The GET/decrypt costs are incurred on each partial read of a file, so random IO can cost more than sequential IO; the speedup of queries with columnar data usually justifies this.

The DynamoDB costs come from the number of entries stores and the allocated capacity.

How to keep costs down

• Don’t run the scale tests with large datasets; keep fs.s3a.scale.test.huge.filesize unset, or a few MB (minimum: 5).
• Remove all files in the filesystem. The root tests usually do this, but it can be manually done:

  *hadoop fs -rm -r -f -skipTrash s3a://test-bucket/

• Abort all outstanding uploads:

  hadoop s3guard uploads -abort -force s3a://test-bucket/

• If you don’t need it, destroy the S3Guard DDB table.

  hadoop s3guard destroy s3a://test-bucket/

The S3Guard tests will automatically create the Dynamo DB table in runs with -Ds3guard -Ddynamo set; default capacity of these buckets tests is very small; it keeps costs down at the expense of IO performance and, for test runs in or near the S3/DDB stores, throttling events.

If you want to manage capacity, use s3guard set-capacity to increase it (performance) or decrease it (costs). For remote hadoop-aws test runs, the read/write capacities of “0” each should suffice; increase it if parallel test run logs warn of throttling.

How to keep your credentials really safe

Although the auth-keys.xml file is marked as ignored in git and subversion, it is still in your source tree, and there’s always that risk that it may creep out.

You can avoid this by keeping your keys outside the source tree and using an absolute XInclude reference to it.

<configuration>

  <include xmlns="http://www.w3.org/2001/XInclude"
    href="file:///users/ubuntu/.auth-keys.xml" />

</configuration>

Enabling the InconsistentAmazonS3CClient

There are two ways of enabling the InconsistentAmazonS3Client: at config-time, or programmatically. For an example of programmatic test usage, see ITestS3GuardListConsistency.

To enable the fault-injecting client via configuration, switch the S3A client to use the “Inconsistent S3 Client Factory” when connecting to S3:

<property>
  <name>fs.s3a.s3.client.factory.impl</name>
  <value>org.apache.hadoop.fs.s3a.InconsistentS3ClientFactory</value>
</property>

The inconsistent client works by:

1. Choosing which objects will be “inconsistent” at the time the object is created or deleted.
2. When listObjects() is called, any keys that we have marked as inconsistent above will not be returned in the results (until the configured delay has elapsed). Similarly, deleted items may be added to missing results to delay the visibility of the delete.

There are two ways of choosing which keys (filenames) will be affected: By substring, and by random probability.

<property>
  <name>fs.s3a.failinject.inconsistency.key.substring</name>
  <value>DELAY_LISTING_ME</value>
</property>

<property>
  <name>fs.s3a.failinject.inconsistency.probability</name>
  <value>1.0</value>
</property>

By default, any object which has the substring “DELAY_LISTING_ME” in its key will subject to delayed visibility. For example, the path s3a://my-bucket/test/DELAY_LISTING_ME/file.txt would match this condition. To match all keys use the value “*” (a single asterisk). This is a special value: We don’t support arbitrary wildcards.

The default probability of delaying an object is 1.0. This means that all keys that match the substring will get delayed visibility. Note that we take the logical and of the two conditions (substring matches and probability random chance occurs). Here are some example configurations:

| substring | probability |  behavior                                  |
|-----------|-------------|--------------------------------------------|
|           | 0.001       | An empty <value> tag in .xml config will   |
|           |             | be interpreted as unset and revert to the  |
|           |             | default value, "DELAY_LISTING_ME"          |
|           |             |                                            |
| *         | 0.001       | 1/1000 chance of *any* key being delayed.  |
|           |             |                                            |
| delay     | 0.01        | 1/100 chance of any key containing "delay" |
|           |             |                                            |
| delay     | 1.0         | All keys containing substring "delay" ..   |

You can also configure how long you want the delay in visibility to last. The default is 5000 milliseconds (five seconds).

<property>
  <name>fs.s3a.failinject.inconsistency.msec</name>
  <value>5000</value>
</property>

Limitations of Inconsistency Injection

Although InconsistentAmazonS3Client can delay the visibility of an object or parent directory, it does not prevent the key of that object from appearing in all prefix searches. For example, if we create the following object with the default configuration above, in an otherwise empty bucket:

s3a://bucket/a/b/c/DELAY_LISTING_ME

Then the following paths will still be visible as directories (ignoring possible real-world inconsistencies):

s3a://bucket/a
s3a://bucket/a/b

Whereas getFileStatus() on the following will be subject to delayed visibility (FileNotFoundException until delay has elapsed):

s3a://bucket/a/b/c
s3a://bucket/a/b/c/DELAY_LISTING_ME

In real-life S3 inconsistency, however, we expect that all the above paths (including a and b) will be subject to delayed visibility.

Using the InconsistentAmazonS3CClient in downstream integration tests

The inconsistent client is shipped in the hadoop-aws JAR, so it can be used in applications which work with S3 to see how they handle inconsistent directory listings.

Testing S3A with S3Guard Enabled

All the S3A tests which work with a private repository can be configured to run with S3Guard by using the s3guard profile. When set, this will run all the tests with local memory for the metadata set to “non-authoritative” mode.

mvn -T 1C verify -Dparallel-tests -DtestsThreadCount=6 -Ds3guard

When the s3guard profile is enabled, following profiles can be specified:

• dynamo: use an AWS-hosted DynamoDB table; creating the table if it does not exist. You will have to pay the bills for DynamoDB web service.
• auth: treat the S3Guard metadata as authoritative.

mvn -T 1C verify -Dparallel-tests -DtestsThreadCount=6 -Ds3guard -Ddynamo -Dauth

When experimenting with options, it is usually best to run a single test suite at a time until the operations appear to be working.

mvn -T 1C verify -Dtest=skip -Dit.test=ITestS3AMiscOperations -Ds3guard -Ddynamo

Notes

1. If the s3guard profile is not set, then the S3Guard properties are those of the test configuration set in s3a contract xml file or auth-keys.xml

If the s3guard profile is set: 1. The S3Guard options from maven (the dynamo and authoritative flags) overwrite any previously set in the configuration files. 1. DynamoDB will be configured to create any missing tables. 1. When using DynamoDB and running ITestDynamoDBMetadataStore, the fs.s3a.s3guard.ddb.test.table property MUST be configured, and the name of that table MUST be different than what is used for fs.s3a.s3guard.ddb.table. The test table is destroyed and modified multiple times during the test. 1. Several of the tests create and destroy DynamoDB tables. The table names are prefixed with the value defined by fs.s3a.s3guard.test.dynamo.table.prefix (default=“s3guard.test.”). The user executing the tests will need sufficient privilege to create and destroy such tables. If the tests abort uncleanly, these tables may be left behind, incurring AWS charges.

How to Dump the Table and Metastore State

There’s an unstable entry point to list the contents of a table and S3 filesystem ot a set of Tab Separated Value files:

hadoop org.apache.hadoop.fs.s3a.s3guard.DumpS3GuardDynamoTable s3a://bucket/ dir/out

This generates a set of files prefixed dir/out- with different views of the world which can then be viewed on the command line or editor:

"type" "deleted" "path" "is_auth_dir" "is_empty_dir" "len" "updated" "updated_s" "last_modified" "last_modified_s" "etag" "version"
"file" "true" "s3a://bucket/fork-0001/test/ITestS3AContractDistCp/testDirectWrite/remote" "false" "UNKNOWN" 0 1562171244451 "Wed Jul 03 17:27:24 BST 2019" 1562171244451 "Wed Jul 03 17:27:24 BST 2019" "" ""
"file" "true" "s3a://bucket/Users/stevel/Projects/hadoop-trunk/hadoop-tools/hadoop-aws/target/test-dir/1/5xlPpalRwv/test/new/newdir/file1" "false" "UNKNOWN" 0 1562171518435 "Wed Jul 03 17:31:58 BST 2019" 1562171518435 "Wed Jul 03 17:31:58 BST 2019" "" ""
"file" "true" "s3a://bucket/Users/stevel/Projects/hadoop-trunk/hadoop-tools/hadoop-aws/target/test-dir/1/5xlPpalRwv/test/new/newdir/subdir" "false" "UNKNOWN" 0 1562171518535 "Wed Jul 03 17:31:58 BST 2019" 1562171518535 "Wed Jul 03 17:31:58 BST 2019" "" ""
"file" "true" "s3a://bucket/test/DELAY_LISTING_ME/testMRJob" "false" "UNKNOWN" 0 1562172036299 "Wed Jul 03 17:40:36 BST 2019" 1562172036299 "Wed Jul 03 17:40:36 BST 2019" "" ""

This is unstable: the output format may change without warning. To understand the meaning of the fields, consult the documentation. They are, currently:

Files generated

Why the two scan entries? The S3A listing and treewalk operations may add new entries to the metastore/DynamoDB table.

Note 1: this is unstable; entry list and meaning may change, sorting of output, the listing algorithm, representation of types, etc. It’s expected uses are: diagnostics, support calls and helping us developers work out what we’ve just broken.

Note 2: This is safe to use against an active store; the tables may appear to be inconsistent due to changes taking place during the dump sequence.

Resetting the Metastore: PurgeS3GuardDynamoTable

The PurgeS3GuardDynamoTable entry point org.apache.hadoop.fs.s3a.s3guard.PurgeS3GuardDynamoTable can list all entries in a store for a specific filesystem, and delete them. It only deletes those entries in the store for that specific filesystem, even if the store is shared.

hadoop org.apache.hadoop.fs.s3a.s3guard.PurgeS3GuardDynamoTable \
  -force s3a://bucket/

Without the -force option the table is scanned, but no entries deleted; with it then all entries for that filesystem are deleted. No attempt is made to order the deletion; while the operation is under way the store is not fully connected (i.e. there may be entries whose parent has already been deleted).

Needless to say: it is not safe to use this against a table in active use.

Scale Testing MetadataStore Directly

There are some scale tests that exercise Metadata Store implementations directly. These ensure that S3Guard is are robust to things like DynamoDB throttling, and compare performance for different implementations. These are included in the scale tests executed when -Dscale is passed to the maven command line.

The two S3Guard scale tests are ITestDynamoDBMetadataStoreScale and ITestLocalMetadataStoreScale.

To run these tests, your DynamoDB table needs to be of limited capacity; the values in ITestDynamoDBMetadataStoreScale currently require a read capacity of 10 or less. a write capacity of 15 or more.

The following settings allow us to run ITestDynamoDBMetadataStoreScale with artificially low read and write capacity provisioned, so we can judge the effects of being throttled by the DynamoDB service:

<property>
  <name>scale.test.operation.count</name>
  <value>10</value>
</property>
<property>
  <name>scale.test.directory.count</name>
  <value>3</value>
</property>
<property>
  <name>fs.s3a.scale.test.enabled</name>
  <value>true</value>
</property>
<property>
  <name>fs.s3a.s3guard.ddb.table</name>
  <value>my-scale-test</value>
</property>
<property>
  <name>fs.s3a.s3guard.ddb.table.create</name>
  <value>true</value>
</property>
<property>
  <name>fs.s3a.s3guard.ddb.table.capacity.read</name>
  <value>5</value>
</property>
<property>
  <name>fs.s3a.s3guard.ddb.table.capacity.write</name>
  <value>5</value>
</property>

These tests verify that the invoked operations can trigger retries in the S3Guard code, rather than just in the AWS SDK level, so showing that if SDK operations fail, they get retried. They also verify that the filesystem statistics are updated to record that throttling took place.

Do not panic if these tests fail to detect throttling!

These tests are unreliable as they need certain conditions to be met to repeatedly fail:

1. You must have a low-enough latency connection to the DynamoDB store that, for the capacity allocated, you can overload it.
2. The AWS Console can give you a view of what is happening here.
3. Running a single test on its own is less likely to trigger an overload than trying to run the whole test suite.
4. And running the test suite more than once, back-to-back, can also help overload the cluster.
5. Stepping through with a debugger will reduce load, so may not trigger failures.

If the tests fail, it probably just means you aren’t putting enough load on the table.

These tests do not verify that the entire set of DynamoDB calls made during the use of a S3Guarded S3A filesystem are wrapped by retry logic.

*The best way to verify resilience is to run the entire hadoop-aws test suite, or even a real application, with throttling enabled.

Testing encrypted DynamoDB tables

By default, a DynamoDB table is encrypted using AWS owned customer master key (CMK). You can enable server side encryption (SSE) using AWS managed CMK or customer managed CMK in KMS before running the S3Guard tests. 1. To enable AWS managed CMK, set the config fs.s3a.s3guard.ddb.table.sse.enabled to true in auth-keys.xml. 1. To enable customer managed CMK, you need to create a KMS key and set the config in auth-keys.xml. The value can be the key ARN or alias. Example:

  <property>
    <name>fs.s3a.s3guard.ddb.table.sse.enabled</name>
    <value>true</value>
  </property>
  <property>
    <name>fs.s3a.s3guard.ddb.table.sse.cmk</name>
    <value>arn:aws:kms:us-west-2:360379543683:key/071a86ff-8881-4ba0-9230-95af6d01ca01</value>
  </property>

For more details about SSE on DynamoDB table, please see S3Guard doc.

Testing only: Local Metadata Store

There is an in-memory Metadata Store for testing.

<property>
  <name>fs.s3a.metadatastore.impl</name>
  <value>org.apache.hadoop.fs.s3a.s3guard.LocalMetadataStore</value>
</property>

This is not for use in production.

Basic command line regression testing

We need a run through of the CLI to see if there have been changes there which cause problems, especially whether new log messages have surfaced, or whether some packaging change breaks that CLI

From the root of the project, create a command line release mvn package -Pdist -DskipTests -Dmaven.javadoc.skip=true -DskipShade;

1. Change into the hadoop-dist/target/hadoop-x.y.z-SNAPSHOT dir.
2. Copy a core-site.xml file into etc/hadoop.
3. Set the HADOOP_OPTIONAL_TOOLS env var on the command line or ~/.hadoop-env.

export HADOOP_OPTIONAL_TOOLS="hadoop-aws"

Run some basic s3guard commands as well as file operations.

export BUCKETNAME=example-bucket-name
export BUCKET=s3a://$BUCKETNAME

bin/hadoop s3guard bucket-info $BUCKET

bin/hadoop s3guard uploads $BUCKET
# repeat twice, once with "no" and once with "yes" as responses
bin/hadoop s3guard uploads -abort $BUCKET

# ---------------------------------------------------
# assuming s3guard is enabled
# if on pay-by-request, expect an error message and exit code of -1
bin/hadoop s3guard set-capacity $BUCKET

# skip for PAY_PER_REQUEST
bin/hadoop s3guard set-capacity -read 15 -write 15 $BUCKET

bin/hadoop s3guard bucket-info -guarded $BUCKET 

bin/hadoop s3guard diff $BUCKET/
bin/hadoop s3guard prune -minutes 10 $BUCKET/
bin/hadoop s3guard import -verbose $BUCKET/
bin/hadoop s3guard authoritative -verbose $BUCKET

# ---------------------------------------------------
# root filesystem operatios
# ---------------------------------------------------

bin/hadoop fs -ls $BUCKET/
# assuming file is not yet created, expect error and status code of 1
bin/hadoop fs -ls $BUCKET/file

# exit code of 0 even when path doesn't exist
bin/hadoop fs -rm -R -f $BUCKET/dir-no-trailing
bin/hadoop fs -rm -R -f $BUCKET/dir-trailing/

# error because it is a directory
bin/hadoop fs -rm $BUCKET/

bin/hadoop fs -touchz $BUCKET/file
# expect I/O error as it is the root directory
bin/hadoop fs -rm -r $BUCKET/

# succeeds
bin/hadoop fs -rm -r $BUCKET/\*

# ---------------------------------------------------
# File operations
# ---------------------------------------------------

bin/hadoop fs -mkdir $BUCKET/dir-no-trailing
# used to fail with S3Guard
bin/hadoop fs -mkdir $BUCKET/dir-trailing/
bin/hadoop fs -touchz $BUCKET/file
bin/hadoop fs -ls $BUCKET/
bin/hadoop fs -mv $BUCKET/file $BUCKET/file2
# expect "No such file or directory"
bin/hadoop fs -stat $BUCKET/file

# expect success
bin/hadoop fs -stat $BUCKET/file2

# expect "file exists"
bin/hadoop fs -mkdir $BUCKET/dir-no-trailing
bin/hadoop fs -mv $BUCKET/file2 $BUCKET/dir-no-trailing
bin/hadoop fs -stat $BUCKET/dir-no-trailing/file2
# treated the same as the file stat
bin/hadoop fs -stat $BUCKET/dir-no-trailing/file2/
bin/hadoop fs -ls $BUCKET/dir-no-trailing/file2/
bin/hadoop fs -ls $BUCKET/dir-no-trailing
# expect a "0" here:
bin/hadoop fs -test -d  $BUCKET/dir-no-trailing ; echo $?
# expect a "1" here:
bin/hadoop fs -test -d  $BUCKET/dir-no-trailing/file2 ; echo $?
# will return NONE unless bucket has checksums enabled
bin/hadoop fs -checksum $BUCKET/dir-no-trailing/file2
# expect "etag" + a long string
bin/hadoop fs -D fs.s3a.etag.checksum.enabled=true -checksum $BUCKET/dir-no-trailing/file2
bin/hadoop fs -expunge -immediate -fs $BUCKET

# ---------------------------------------------------
# Delegation Token support
# ---------------------------------------------------

# failure unless delegation tokens are enabled
bin/hdfs fetchdt --webservice $BUCKET secrets.bin
# success
bin/hdfs fetchdt -D fs.s3a.delegation.token.binding=org.apache.hadoop.fs.s3a.auth.delegation.SessionTokenBinding --webservice $BUCKET secrets.bin
bin/hdfs fetchdt -print secrets.bin

# expect warning "No TokenRenewer defined for token kind S3ADelegationToken/Session"
bin/hdfs fetchdt -renew secrets.bin

# ---------------------------------------------------
# Directory markers
# ---------------------------------------------------

# require success
bin/hadoop s3guard bucket-info -markers aware $BUCKET
# expect failure unless bucket policy is keep
bin/hadoop s3guard bucket-info -markers keep $BUCKET/path

# you may need to set this on a per-bucket basis if you have already been
# playing with options
bin/hadoop s3guard -D fs.s3a.directory.marker.retention=keep bucket-info -markers keep $BUCKET/path
bin/hadoop s3guard -D fs.s3a.bucket.$BUCKETNAME.directory.marker.retention=keep bucket-info -markers keep $BUCKET/path

# expect to see "Directory markers will be kept" messages and status code of "46"
bin/hadoop fs -D fs.s3a.bucket.$BUCKETNAME.directory.marker.retention=keep -mkdir $BUCKET/p1
bin/hadoop fs -D fs.s3a.bucket.$BUCKETNAME.directory.marker.retention=keep -mkdir $BUCKET/p1/p2
bin/hadoop fs -D fs.s3a.bucket.$BUCKETNAME.directory.marker.retention=keep -touchz $BUCKET/p1/p2/file

# expect failure as markers will be found for /p1/ and /p1/p2/
bin/hadoop s3guard markers -audit -verbose $BUCKET

# clean will remove markers
bin/hadoop s3guard markers -clean -verbose $BUCKET

# expect success and exit code of 0
bin/hadoop s3guard markers -audit -verbose $BUCKET

# ---------------------------------------------------
# S3 Select on Landsat
# ---------------------------------------------------

export LANDSATGZ=s3a://landsat-pds/scene_list.gz

bin/hadoop s3guard select -header use -compression gzip $LANDSATGZ \
 "SELECT s.entityId,s.cloudCover FROM S3OBJECT s WHERE s.cloudCover < '0.0' LIMIT 100"

Other tests

• Whatever applications you have which use S3A: build and run them before the upgrade, Then see if complete successfully in roughly the same time once the upgrade is applied.
• Test any third-party endpoints you have access to.
• Try different regions (especially a v4 only region), and encryption settings.
• Any performance tests you have can identify slowdowns, which can be a sign of changed behavior in the SDK (especially on stream reads and writes).
• If you can, try to test in an environment where a proxy is needed to talk to AWS services.
• Try and get other people, especially anyone with their own endpoints, apps or different deployment environments, to run their own tests.
• Run the load tests, especially ILoadTestS3ABulkDeleteThrottling.

Dealing with Deprecated APIs and New Features

A Jenkins run should tell you if there are new deprecations. If so, you should think about how to deal with them.

Moving to methods and APIs which weren’t in the previous SDK release makes it harder to roll back if there is a problem; but there may be good reasons for the deprecation.

At the same time, there may be good reasons for staying with the old code.

• AWS have embraced the builder pattern for new operations; note that objects constructed this way often have their (existing) setter methods disabled; this may break existing code.
• New versions of S3 calls (list v2, bucket existence checks, bulk operations) may be better than the previous HTTP operations & APIs, but they may not work with third-party endpoints, so can only be adopted if made optional, which then adds a new configuration option (with docs, testing, …). A change like that must be done in its own patch, with its new tests which compare the old vs new operations.

Committing the patch

When the patch is committed: update the JIRA to the version number actually used; use that title in the commit message.

Be prepared to roll-back, re-iterate or code your way out of a regression.

There may be some problem which surfaces with wider use, which can get fixed in a new AWS release, rolling back to an older one, or just worked around HADOOP-14596.

Don’t be surprised if this happens, don’t worry too much, and, while that rollback option is there to be used, ideally try to work forwards.

If the problem is with the SDK, file issues with the AWS SDK Bug tracker. If the problem can be fixed or worked around in the Hadoop code, do it there too.