Several of our tests call System.setProperty (or test code which implicitly sets system properties) and don't always reset/clear the modified properties, which can create ordering dependencies between tests and cause hard-to-diagnose failures.
This patch removes most uses of System.setProperty from our tests, since in most cases we can use SparkConf to set these configurations (there are a few exceptions, including the tests of SparkConf itself).
For the cases where we continue to use System.setProperty, this patch introduces a `ResetSystemProperties` ScalaTest mixin class which snapshots the system properties before individual tests and to automatically restores them on test completion / failure. See the block comment at the top of the ResetSystemProperties class for more details.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#3739 from JoshRosen/cleanup-system-properties-in-tests and squashes the following commits:
0236d66 [Josh Rosen] Replace setProperty uses in two example programs / tools
3888fe3 [Josh Rosen] Remove setProperty use in LocalJavaStreamingContext
4f4031d [Josh Rosen] Add note on why SparkSubmitSuite needs ResetSystemProperties
4742a5b [Josh Rosen] Clarify ResetSystemProperties trait inheritance ordering.
0eaf0b6 [Josh Rosen] Remove setProperty call in TaskResultGetterSuite.
7a3d224 [Josh Rosen] Fix trait ordering
3fdb554 [Josh Rosen] Remove setProperty call in TaskSchedulerImplSuite
bee20df [Josh Rosen] Remove setProperty calls in SparkContextSchedulerCreationSuite
655587c [Josh Rosen] Remove setProperty calls in JobCancellationSuite
3f2f955 [Josh Rosen] Remove System.setProperty calls in DistributedSuite
cfe9cce [Josh Rosen] Remove use of system properties in SparkContextSuite
8783ab0 [Josh Rosen] Remove TestUtils.setSystemProperty, since it is subsumed by the ResetSystemProperties trait.
633a84a [Josh Rosen] Remove use of system properties in FileServerSuite
25bfce2 [Josh Rosen] Use ResetSystemProperties in UtilsSuite
1d1aa5a [Josh Rosen] Use ResetSystemProperties in SizeEstimatorSuite
dd9492b [Josh Rosen] Use ResetSystemProperties in AkkaUtilsSuite
b0daff2 [Josh Rosen] Use ResetSystemProperties in BlockManagerSuite
e9ded62 [Josh Rosen] Use ResetSystemProperties in TaskSchedulerImplSuite
5b3cb54 [Josh Rosen] Use ResetSystemProperties in SparkListenerSuite
0995c4b [Josh Rosen] Use ResetSystemProperties in SparkContextSchedulerCreationSuite
c83ded8 [Josh Rosen] Use ResetSystemProperties in SparkConfSuite
51aa870 [Josh Rosen] Use withSystemProperty in ShuffleSuite
60a63a1 [Josh Rosen] Use ResetSystemProperties in JobCancellationSuite
14a92e4 [Josh Rosen] Use withSystemProperty in FileServerSuite
628f46c [Josh Rosen] Use ResetSystemProperties in DistributedSuite
9e3e0dd [Josh Rosen] Add ResetSystemProperties test fixture mixin; use it in SparkSubmitSuite.
4dcea38 [Josh Rosen] Move withSystemProperty to TestUtils class.
Somehow maven shade plugin is set in infinite loop of creating effective pom.
Author: Prashant Sharma <prashant.s@imaginea.com>
Author: Prashant Sharma <scrapcodes@gmail.com>
Closes#2959 from ScrapCodes/SPARK-3962/scope-provided and squashes the following commits:
994d1d3 [Prashant Sharma] Fixed failing flume tests
270b4fb [Prashant Sharma] Removed most of the unused code.
bb3bbfd [Prashant Sharma] SPARK-3962 Marked scope as provided for external.
A leak of event loops may be causing test failures.
Author: Aaron Davidson <aaron@databricks.com>
Closes#3053 from aarondav/leak and squashes the following commits:
e676d18 [Aaron Davidson] Typo!
8f96475 [Aaron Davidson] Keep original ssc semantics
7e49f10 [Aaron Davidson] A leak of event loops may be causing test failures.
If classes implementing Serializable or Externalizable interfaces throw
exceptions other than IOException or ClassNotFoundException from their
(de)serialization methods, then this results in an unhelpful
"IOException: unexpected exception type" rather than the actual exception that
produced the (de)serialization error.
This patch fixes this by adding a utility method that re-wraps any uncaught
exceptions in IOException (unless they are already instances of IOException).
Author: Josh Rosen <joshrosen@databricks.com>
Closes#2932 from JoshRosen/SPARK-4080 and squashes the following commits:
cd3a9be [Josh Rosen] [SPARK-4080] Only throw IOException from [write|read][Object|External].
@harishreedharan @pwendell
See JIRA for diagnosis of the problem
https://issues.apache.org/jira/browse/SPARK-3912
The solution was to reimplement it.
1. Find a free port (by binding and releasing a server-scoket), and then use that port
2. Remove thread.sleep()s, instead repeatedly try to create a sender and send data and check whether data was sent. Use eventually() to minimize waiting time.
3. Check whether all the data was received, without caring about batches.
Author: Tathagata Das <tathagata.das1565@gmail.com>
Closes#2773 from tdas/flume-test-fix and squashes the following commits:
93cd7f6 [Tathagata Das] Reimplimented FlumeStreamSuite to be more robust.
Thread names are useful for correlating failures.
Author: Reynold Xin <rxin@apache.org>
Closes#2600 from rxin/log4j and squashes the following commits:
83ffe88 [Reynold Xin] [SPARK-3748] Log thread name in unit test logs
Since it looked quite easy, I took the liberty of making a quick PR that just uses `Utils.startServiceOnPort` to fix this. It works locally for me.
Author: Sean Owen <sowen@cloudera.com>
Closes#2601 from srowen/SPARK-3744 and squashes the following commits:
ddc9319 [Sean Owen] Avoid port contention in tests by retrying several ports for Flume stream
Author: GuoQiang Li <witgo@qq.com>
Closes#2265 from witgo/FlumeStreamSuite and squashes the following commits:
6c99e6e [GuoQiang Li] Remove extra semicolon in FlumeStreamSuite.scala
Currently lot of errors get thrown from Avro IPC layer when the dstream
or sink is shutdown. This PR cleans it up. Some refactoring is done in the
receiver code to put all of the RPC code into a single Try and just recover
from that. The sink code has also been cleaned up.
Author: Hari Shreedharan <hshreedharan@apache.org>
Closes#2065 from harishreedharan/clean-flume-shutdown and squashes the following commits:
f93a07c [Hari Shreedharan] Formatting fixes.
d7427cc [Hari Shreedharan] More fixes!
a0a8852 [Hari Shreedharan] Fix race condition, hopefully! Minor other changes.
4c9ed02 [Hari Shreedharan] Remove unneeded list in Callback handler. Other misc changes.
8fee36f [Hari Shreedharan] Scala-library is required, else maven build fails. Also catch InterruptedException in TxnProcessor.
445e700 [Hari Shreedharan] Merge remote-tracking branch 'asf/master' into clean-flume-shutdown
87232e0 [Hari Shreedharan] Refactor Flume Input Stream. Clean up code, better error handling.
9001d26 [Hari Shreedharan] Change log level to debug in TransactionProcessor#shutdown method
e7b8d82 [Hari Shreedharan] Incorporate review feedback
598efa7 [Hari Shreedharan] Clean up some exception handling code
e1027c6 [Hari Shreedharan] Merge remote-tracking branch 'asf/master' into clean-flume-shutdown
ed608c8 [Hari Shreedharan] [SPARK-3154][STREAMING] Make FlumePollingInputDStream shutdown cleaner.
This patch adds unit tests for Spark Sink.
It also removes the private[flume] for Spark Sink,
since the sink is instantiated from Flume configuration (looks like this is ignored by reflection which is used by
Flume, but we should still remove it anyway).
Author: Hari Shreedharan <hshreedharan@apache.org>
Author: Hari Shreedharan <hshreedharan@cloudera.com>
Closes#1958 from harishreedharan/spark-sink-test and squashes the following commits:
e3110b9 [Hari Shreedharan] Add a sleep to allow sink to commit the transactions
120b81e [Hari Shreedharan] Fix complexity in threading model in test
4df5be6 [Hari Shreedharan] Merge remote-tracking branch 'asf/master' into spark-sink-test
c9190d1 [Hari Shreedharan] Indentation and spaces changes
7fedc5a [Hari Shreedharan] Merge remote-tracking branch 'asf/master' into spark-sink-test
abc20cb [Hari Shreedharan] Minor test changes
7b9b649 [Hari Shreedharan] Merge branch 'master' into spark-sink-test
f2c56c9 [Hari Shreedharan] Update SparkSinkSuite.scala
a24aac8 [Hari Shreedharan] Remove unused var
c86d615 [Hari Shreedharan] [SPARK-3054][STREAMING] Add unit tests for Spark Sink.
Author: Hari Shreedharan <harishreedharan@gmail.com>
Closes#1820 from harishreedharan/use-free-ports and squashes the following commits:
b939067 [Hari Shreedharan] Remove unused import.
67856a8 [Hari Shreedharan] Remove findFreePort.
0ea51d1 [Hari Shreedharan] Make some changes to getPort to use map on the serverOpt.
1fb0283 [Hari Shreedharan] Merge branch 'master' of https://github.com/apache/spark into use-free-ports
b351651 [Hari Shreedharan] Allow Netty to choose port, and query it to decide the port to bind to. Leaving findFreePort as is, if other tests want to use it at some point.
e6c9620 [Hari Shreedharan] Making sure the second sink uses the correct port.
11c340d [Hari Shreedharan] Add info about race condition to scaladoc.
e89d135 [Hari Shreedharan] Adding Scaladoc.
6013bb0 [Hari Shreedharan] [STREAMING] Find free ports to use before attempting to create Flume Sink in Flume Polling Suite
This is failing my tests in #1777. @tdas
Author: Andrew Or <andrewor14@gmail.com>
Closes#1803 from andrewor14/fix-flaky-streaming-test and squashes the following commits:
ea11a03 [Andrew Or] Catch all exceptions caused by BindExceptions
54a0ca0 [Andrew Or] Merge branch 'master' of github.com:apache/spark into fix-flaky-streaming-test
664095c [Andrew Or] Tone down bind exception message
af3ddc9 [Andrew Or] Handle port collisions in flume polling test
This has been failing on master. One possible cause is that the port
gets contended if multiple test runs happen concurrently and they
hit this test at the same time. Since this test takes a long time
(60 seconds) that's very plausible. This patch randomizes the port
used in this test to avoid contention.
...sh model
Currently Spark uses Flume's internal Avro Protocol to ingest data from Flume. If the executor running the
receiver fails, it currently has to be restarted on the same node to be able to receive data.
This commit adds a new Sink which can be deployed to a Flume agent. This sink can be polled by a new
DStream that is also included in this commit. This model ensures that data can be pulled into Spark from
Flume even if the receiver is restarted on a new node. This also allows the receiver to receive data on
multiple threads for better performance.
Author: Hari Shreedharan <harishreedharan@gmail.com>
Author: Hari Shreedharan <hshreedharan@apache.org>
Author: Tathagata Das <tathagata.das1565@gmail.com>
Author: harishreedharan <hshreedharan@cloudera.com>
Closes#807 from harishreedharan/master and squashes the following commits:
e7f70a3 [Hari Shreedharan] Merge remote-tracking branch 'asf-git/master'
96cfb6f [Hari Shreedharan] Merge remote-tracking branch 'asf/master'
e48d785 [Hari Shreedharan] Documenting flume-sink being ignored for Mima checks.
5f212ce [Hari Shreedharan] Ignore Spark Sink from mima.
981bf62 [Hari Shreedharan] Merge remote-tracking branch 'asf/master'
7a1bc6e [Hari Shreedharan] Fix SparkBuild.scala
a082eb3 [Hari Shreedharan] Merge remote-tracking branch 'asf/master'
1f47364 [Hari Shreedharan] Minor fixes.
73d6f6d [Hari Shreedharan] Cleaned up tests a bit. Added some docs in multiple places.
65b76b4 [Hari Shreedharan] Fixing the unit test.
e59cc20 [Hari Shreedharan] Use SparkFlumeEvent instead of the new type. Also, Flume Polling Receiver now uses the store(ArrayBuffer) method.
f3c99d1 [Hari Shreedharan] Merge remote-tracking branch 'asf/master'
3572180 [Hari Shreedharan] Adding a license header, making Jenkins happy.
799509f [Hari Shreedharan] Fix a compile issue.
3c5194c [Hari Shreedharan] Merge remote-tracking branch 'asf/master'
d248d22 [harishreedharan] Merge pull request #1 from tdas/flume-polling
10b6214 [Tathagata Das] Changed public API, changed sink package, and added java unit test to make sure Java API is callable from Java.
1edc806 [Hari Shreedharan] SPARK-1729. Update logging in Spark Sink.
8c00289 [Hari Shreedharan] More debug messages
393bd94 [Hari Shreedharan] SPARK-1729. Use LinkedBlockingQueue instead of ArrayBuffer to keep track of connections.
120e2a1 [Hari Shreedharan] SPARK-1729. Some test changes and changes to utils classes.
9fd0da7 [Hari Shreedharan] SPARK-1729. Use foreach instead of map for all Options.
8136aa6 [Hari Shreedharan] Adding TransactionProcessor to map on returning batch of data
86aa274 [Hari Shreedharan] Merge remote-tracking branch 'asf/master'
205034d [Hari Shreedharan] Merging master in
4b0c7fc [Hari Shreedharan] FLUME-1729. New Flume-Spark integration.
bda01fc [Hari Shreedharan] FLUME-1729. Flume-Spark integration.
0d69604 [Hari Shreedharan] FLUME-1729. Better Flume-Spark integration.
3c23c18 [Hari Shreedharan] SPARK-1729. New Spark-Flume integration.
70bcc2a [Hari Shreedharan] SPARK-1729. New Flume-Spark integration.
d6fa3aa [Hari Shreedharan] SPARK-1729. New Flume-Spark integration.
e7da512 [Hari Shreedharan] SPARK-1729. Fixing import order
9741683 [Hari Shreedharan] SPARK-1729. Fixes based on review.
c604a3c [Hari Shreedharan] SPARK-1729. Optimize imports.
0f10788 [Hari Shreedharan] SPARK-1729. Make Flume pull data from source, rather than the current push model
87775aa [Hari Shreedharan] SPARK-1729. Make Flume pull data from source, rather than the current push model
8df37e4 [Hari Shreedharan] SPARK-1729. Make Flume pull data from source, rather than the current push model
03d6c1c [Hari Shreedharan] SPARK-1729. Make Flume pull data from source, rather than the current push model
08176ad [Hari Shreedharan] SPARK-1729. Make Flume pull data from source, rather than the current push model
d24d9d4 [Hari Shreedharan] SPARK-1729. Make Flume pull data from source, rather than the current push model
6d6776a [Hari Shreedharan] SPARK-1729. Make Flume pull data from source, rather than the current push model
The line break inadvertently means this was interpreted as a call to the no-arg constructor. This doesn't exist in older Netty even. (Also fixed a val name typo.)
Author: Sean Owen <srowen@gmail.com>
Closes#1466 from srowen/SPARK-1478.2 and squashes the following commits:
59c3501 [Sean Owen] Line break caused Scala to interpret NioServerSocketChannelFactory constructor as the no-arg version, which is not even present in some versions of Netty
This is a modified version of this PR https://github.com/apache/spark/pull/1168 done by @tmalaska
Adds MIMA binary check exclusions.
Author: tmalaska <ted.malaska@cloudera.com>
Author: Tathagata Das <tathagata.das1565@gmail.com>
Closes#1347 from tdas/FLUME-1915 and squashes the following commits:
96065df [Tathagata Das] Added Mima exclusion for FlumeReceiver.
41d5338 [tmalaska] Address line 57 that was too long
12617e5 [tmalaska] SPARK-1478: Upgrade FlumeInputDStream's Flume...
flume event sent to Spark will fail if the body is too large and numHeaders is greater than zero
Author: joyyoj <sunshch@gmail.com>
Closes#951 from joyyoj/master and squashes the following commits:
f4660c5 [joyyoj] [SPARK-1998] SparkFlumeEvent with body bigger than 1020 bytes are not read properly
The changes could be ported back to 0.9 as well.
Changing in.read to in.readFully to read the whole input stream rather than the first 1020 bytes.
This should ok considering that Flume caps the body size to 32K by default.
Author: David Lemieux <david.lemieux@radialpoint.com>
Closes#865 from lemieud/SPARK-1916 and squashes the following commits:
a265673 [David Lemieux] Updated SparkFlumeEvent to read the whole stream rather than the first X bytes.
(cherry picked from commit 0b769b73fb)
Signed-off-by: Patrick Wendell <pwendell@gmail.com>
This is a few changes based on the original patch by @scrapcodes.
Author: Prashant Sharma <prashant.s@imaginea.com>
Author: Patrick Wendell <pwendell@gmail.com>
Closes#785 from pwendell/package-docs and squashes the following commits:
c32b731 [Patrick Wendell] Changes based on Prashant's patch
c0463d3 [Prashant Sharma] added eof new line
ce8bf73 [Prashant Sharma] Added eof new line to all files.
4c35f2e [Prashant Sharma] SPARK-1563 Add package-info.java and package.scala files for all packages that appear in docs
Three issues related to temp files that tests generate – these should be touched up for hygiene but are not urgent.
Modules have a log4j.properties which directs the unit-test.log output file to a directory like `[module]/target/unit-test.log`. But this ends up creating `[module]/[module]/target/unit-test.log` instead of former.
The `work/` directory is not deleted by "mvn clean", in the parent and in modules. Neither is the `checkpoint/` directory created under the various external modules.
Many tests create a temp directory, which is not usually deleted. This can be largely resolved by calling `deleteOnExit()` at creation and trying to call `Utils.deleteRecursively` consistently to clean up, sometimes in an `@After` method.
_If anyone seconds the motion, I can create a more significant change that introduces a new test trait along the lines of `LocalSparkContext`, which provides management of temp directories for subclasses to take advantage of._
Author: Sean Owen <sowen@cloudera.com>
Closes#732 from srowen/SPARK-1798 and squashes the following commits:
5af578e [Sean Owen] Try to consistently delete test temp dirs and files, and set deleteOnExit() for each
b21b356 [Sean Owen] Remove work/ and checkpoint/ dirs with mvn clean
bdd0f41 [Sean Owen] Remove duplicate module dir in log4j.properties output path for tests
Unfortunately, this is not exhaustive - particularly hive tests still fail due to path issues.
Author: Mridul Muralidharan <mridulm80@apache.org>
This patch had conflicts when merged, resolved by
Committer: Matei Zaharia <matei@databricks.com>
Closes#505 from mridulm/windows_fixes and squashes the following commits:
ef12283 [Mridul Muralidharan] Move to org.apache.commons.lang3 for StringEscapeUtils. Earlier version was buggy appparently
cdae406 [Mridul Muralidharan] Remove leaked changes from > 2G fix branch
3267f4b [Mridul Muralidharan] Fix build failures
35b277a [Mridul Muralidharan] Fix Scalastyle failures
bc69d14 [Mridul Muralidharan] Change from hardcoded path separator
10c4d78 [Mridul Muralidharan] Use explicit encoding while using getBytes
1337abd [Mridul Muralidharan] fix classpath while running in windows
The current Network Receiver API makes it slightly complicated to right a new receiver as one needs to create an instance of BlockGenerator as shown in SocketReceiver
https://github.com/apache/spark/blob/master/streaming/src/main/scala/org/apache/spark/streaming/dstream/SocketInputDStream.scala#L51
Exposing the BlockGenerator interface has made it harder to improve the receiving process. The API of NetworkReceiver (which was not a very stable API anyways) needs to be change if we are to ensure future stability.
Additionally, the functions like streamingContext.socketStream that create input streams, return DStream objects. That makes it hard to expose functionality (say, rate limits) unique to input dstreams. They should return InputDStream or NetworkInputDStream. This is still not yet implemented.
This PR is blocked on the graceful shutdown PR #247
Author: Tathagata Das <tathagata.das1565@gmail.com>
Closes#300 from tdas/network-receiver-api and squashes the following commits:
ea27b38 [Tathagata Das] Merge remote-tracking branch 'apache-github/master' into network-receiver-api
3a4777c [Tathagata Das] Renamed NetworkInputDStream to ReceiverInputDStream, and ActorReceiver related stuff.
838dd39 [Tathagata Das] Added more events to the StreamingListener to report errors and stopped receivers.
a75c7a6 [Tathagata Das] Address some PR comments and fixed other issues.
91bfa72 [Tathagata Das] Fixed bugs.
8533094 [Tathagata Das] Scala style fixes.
028bde6 [Tathagata Das] Further refactored receiver to allow restarting of a receiver.
43f5290 [Tathagata Das] Made functions that create input streams return InputDStream and NetworkInputDStream, for both Scala and Java.
2c94579 [Tathagata Das] Fixed graceful shutdown by removing interrupts on receiving thread.
9e37a0b [Tathagata Das] Merge remote-tracking branch 'apache-github/master' into network-receiver-api
3223e95 [Tathagata Das] Refactored the code that runs the NetworkReceiver into further classes and traits to make them more testable.
a36cc48 [Tathagata Das] Refactored the NetworkReceiver API for future stability.
Author: Prashant Sharma <prashant.s@imaginea.com>
Closes#261 from ScrapCodes/comment-style-check2 and squashes the following commits:
6cde61e [Prashant Sharma] comment style space before ending */ check.
