SPARK-998: Support Launching Driver Inside of Standalone Mode
[NOTE: I need to bring the tests up to date with new changes, so for now they will fail]
This patch provides support for launching driver programs inside of a standalone cluster manager. It also supports monitoring and re-launching of driver programs which is useful for long running, recoverable applications such as Spark Streaming jobs. For those jobs, this patch allows a deployment mode which is resilient to the failure of any worker node, failure of a master node (provided a multi-master setup), and even failures of the applicaiton itself, provided they are recoverable on a restart. Driver information, such as the status and logs from a driver, is displayed in the UI
There are a few small TODO's here, but the code is generally feature-complete. They are:
- Bring tests up to date and add test coverage
- Restarting on failure should be optional and maybe off by default.
- See if we can re-use akka connections to facilitate clients behind a firewall
A sensible place to start for review would be to look at the `DriverClient` class which presents users the ability to launch their driver program. I've also added an example program (`DriverSubmissionTest`) that allows you to test this locally and play around with killing workers, etc. Most of the code is devoted to persisting driver state in the cluster manger, exposing it in the UI, and dealing correctly with various types of failures.
Instructions to test locally:
- `sbt/sbt assembly/assembly examples/assembly`
- start a local version of the standalone cluster manager
```
./spark-class org.apache.spark.deploy.client.DriverClient \
-j -Dspark.test.property=something \
-e SPARK_TEST_KEY=SOMEVALUE \
launch spark://10.99.1.14:7077 \
../path-to-examples-assembly-jar \
org.apache.spark.examples.DriverSubmissionTest 1000 some extra options --some-option-here -X 13
```
- Go in the UI and make sure it started correctly, look at the output etc
- Kill workers, the driver program, masters, etc.
Refactored the streaming project to separate external libraries like Twitter, Kafka, Flume, etc.
At a high level, these are the following changes.
1. All the external code was put in `SPARK_HOME/external/` as separate SBT projects and Maven modules. Their artifact names are `spark-streaming-twitter`, `spark-streaming-kafka`, etc. Both SparkBuild.scala and pom.xml files have been updated. References to external libraries and repositories have been removed from the settings of root and streaming projects/modules.
2. To avail the external functionality (say, creating a Twitter stream), the developer has to `import org.apache.spark.streaming.twitter._` . For Scala API, the developer has to call `TwitterUtils.createStream(streamingContext, ...)`. For the Java API, the developer has to call `TwitterUtils.createStream(javaStreamingContext, ...)`.
3. Each external project has its own scala and java unit tests. Note the unit tests of each external library use classes of the streaming unit tests (`TestSuiteBase`, `LocalJavaStreamingContext`, etc.). To enable this code sharing among test classes, `dependsOn(streaming % "compile->compile,test->test")` was used in the SparkBuild.scala . In the streaming/pom.xml, an additional `maven-jar-plugin` was necessary to capture this dependency (see comment inside the pom.xml for more information).
4. Jars of the external projects have been added to examples project but not to the assembly project.
5. In some files, imports have been rearrange to conform to the Spark coding guidelines.
Approximate distinct count
Added countApproxDistinct() to RDD and countApproxDistinctByKey() to PairRDDFunctions to approximately count distinct number of elements and distinct number of values per key, respectively. Both functions use HyperLogLog from stream-lib for counting. Both functions take a parameter that controls the trade-off between accuracy and memory consumption. Also added Scala docs and test suites for both methods.
The correct format is to not have a trailing slash.
For me this caused non-deterministic failures due to issues fetching
certain artifacts. The issue was that some of the maven caches would
fail to fetch the artifact (due to the way that the artifact
path was concatenated with the repository) and this short-circuited
the download process in a silent way. Here is what the log output
looked like:
Downloading: http://repo.maven.apache.org/maven2/org/spark-project/akka/akka-remote_2.10/2.2.3-shaded-protobuf/akka-remote_2.10-2.2.3-shaded-protobuf.pom
[WARNING] The POM for org.spark-project.akka:akka-remote_2.10🫙2.2.3-shaded-protobuf is missing, no dependency information available
This was pretty brutal to debug since there was no error message
anywhere and the path *looks* correct as reported by the Maven log.
Standalone Scheduler fault tolerance using ZooKeeper
This patch implements full distributed fault tolerance for standalone scheduler Masters.
There is only one master Leader at a time, which is actively serving scheduling
requests. If this Leader crashes, another master will eventually be elected, reconstruct
the state from the first Master, and continue serving scheduling requests.
Leader election is performed using the ZooKeeper leader election pattern. We try to minimize
the use of ZooKeeper and the assumptions about ZooKeeper's behavior, so there is a layer of
retries and session monitoring on top of the ZooKeeper client.
Master failover follows directly from the single-node Master recovery via the file
system (patch d5a96fe), save that the Master state is stored in ZooKeeper instead.
Configuration:
By default, no recovery mechanism is enabled (spark.deploy.recoveryMode = NONE).
By setting spark.deploy.recoveryMode to ZOOKEEPER and setting spark.deploy.zookeeper.url
to an appropriate ZooKeeper URL, ZooKeeper recovery mode is enabled.
By setting spark.deploy.recoveryMode to FILESYSTEM and setting spark.deploy.recoveryDirectory
to an appropriate directory accessible by the Master, we will keep the behavior of from d5a96fe.
Additionally, places where a Master could be specificied by a spark:// url can now take
comma-delimited lists to specify backup masters. Note that this is only used for registration
of NEW Workers and application Clients. Once a Worker or Client has registered with the
Master Leader, it is "in the system" and will never need to register again.
Conflicts:
bagel/pom.xml
core/pom.xml
core/src/test/scala/org/apache/spark/ui/UISuite.scala
examples/pom.xml
mllib/pom.xml
pom.xml
project/SparkBuild.scala
repl/pom.xml
streaming/pom.xml
tools/pom.xml
In scala 2.10, a shorter representation is used for naming artifacts
so changed to shorter scala version for artifacts and made it a property in pom.
