aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* [SPARK-12315][SQL] isnotnull operator not pushed down for JDBC datasource.hyukjinkwon2015-12-152-0/+3
| | | | | | | | | | | | | | | | https://issues.apache.org/jira/browse/SPARK-12315 `IsNotNull` filter is not being pushed down for JDBC datasource. It looks it is SQL standard according to [SQL-92](http://www.contrib.andrew.cmu.edu/~shadow/sql/sql1992.txt), SQL:1999, [SQL:2003](http://www.wiscorp.com/sql_2003_standard.zip) and [SQL:201x](http://www.wiscorp.com/sql20nn.zip) and I believe most databases support this. In this PR, I simply added the case for `IsNotNull` filter to produce a proper filter string. Author: hyukjinkwon <gurwls223@gmail.com> This patch had conflicts when merged, resolved by Committer: Reynold Xin <rxin@databricks.com> Closes #10287 from HyukjinKwon/SPARK-12315.
* [SPARK-12314][SQL] isnull operator not pushed down for JDBC datasource.hyukjinkwon2015-12-152-0/+2
| | | | | | | | | | | | | | | | https://issues.apache.org/jira/browse/SPARK-12314 `IsNull` filter is not being pushed down for JDBC datasource. It looks it is SQL standard according to [SQL-92](http://www.contrib.andrew.cmu.edu/~shadow/sql/sql1992.txt), SQL:1999, [SQL:2003](http://www.wiscorp.com/sql_2003_standard.zip) and [SQL:201x](http://www.wiscorp.com/sql20nn.zip) and I believe most databases support this. In this PR, I simply added the case for `IsNull` filter to produce a proper filter string. Author: hyukjinkwon <gurwls223@gmail.com> This patch had conflicts when merged, resolved by Committer: Reynold Xin <rxin@databricks.com> Closes #10286 from HyukjinKwon/SPARK-12314.
* [SPARK-12249][SQL] JDBC non-equality comparison operator not pushed down.hyukjinkwon2015-12-152-0/+3
| | | | | | | | | | https://issues.apache.org/jira/browse/SPARK-12249 Currently `!=` operator is not pushed down correctly. I simply added a case for this. Author: hyukjinkwon <gurwls223@gmail.com> Closes #10233 from HyukjinKwon/SPARK-12249.
* [SPARK-12304][STREAMING] Make Spark Streaming web UI display more fri…proflin2015-12-151-1/+7
| | | | | | | | | | | | | | | | | | | | …endly Receiver graphs Currently, the Spark Streaming web UI uses the same maxY when displays 'Input Rate Times& Histograms' and 'Per-Receiver Times& Histograms'. This may lead to somewhat un-friendly graphs: once we have tens of Receivers or more, every 'Per-Receiver Times' line almost hits the ground. This issue proposes to calculate a new maxY against the original one, which is shared among all the `Per-Receiver Times& Histograms' graphs. Before: ![before-5](https://cloud.githubusercontent.com/assets/15843379/11761362/d790c356-a0fa-11e5-860e-4b834603de1d.png) After: ![after-5](https://cloud.githubusercontent.com/assets/15843379/11761361/cfabf692-a0fa-11e5-97d0-4ad124aaca2a.png) Author: proflin <proflin.me@gmail.com> Closes #10318 from proflin/SPARK-12304.
* [SPARK-4117][YARN] Spark on Yarn handle AM being told command from RMDevaraj K2015-12-151-1/+8
| | | | | | | | | | | | Spark on Yarn handle AM being told command from RM When RM throws ApplicationAttemptNotFoundException for allocate invocation, making the ApplicationMaster to finish immediately without any retries. Author: Devaraj K <devaraj@apache.org> Closes #10129 from devaraj-kavali/SPARK-4117.
* [SPARK-10477][SQL] using DSL in ColumnPruningSuite to improve readabilityWenchen Fan2015-12-152-21/+27
| | | | | | Author: Wenchen Fan <cloud0fan@outlook.com> Closes #8645 from cloud-fan/test.
* [SPARK-12062][CORE] Change Master to asyc rebuild UI when application completesBryan Cutler2015-12-152-29/+52
| | | | | | | | This change builds the event history of completed apps asynchronously so the RPC thread will not be blocked and allow new workers to register/remove if the event log history is very large and takes a long time to rebuild. Author: Bryan Cutler <bjcutler@us.ibm.com> Closes #10284 from BryanCutler/async-MasterUI-SPARK-12062.
* [SPARK-9886][CORE] Fix to use ShutdownHookManager inNaveen2015-12-151-11/+5
| | | | | | | | ExternalBlockStore.scala Author: Naveen <naveenminchu@gmail.com> Closes #10313 from naveenminchu/branch-fix-SPARK-9886.
* [SPARK-10123][DEPLOY] Support specifying deploy mode from configurationjerryshao2015-12-155-7/+64
| | | | | | | | Please help to review, thanks a lot. Author: jerryshao <sshao@hortonworks.com> Closes #10195 from jerryshao/SPARK-10123.
* [SPARK-9026][SPARK-4514] Modifications to JobWaiter, FutureAction, and ↵Richard W. Eggert II2015-12-157-158/+251
| | | | | | | | | | | | | | | AsyncRDDActions to support non-blocking operation These changes rework the implementations of `SimpleFutureAction`, `ComplexFutureAction`, `JobWaiter`, and `AsyncRDDActions` such that asynchronous callbacks on the generated `Futures` NEVER block waiting for a job to complete. A small amount of mutex synchronization is necessary to protect the internal fields that manage cancellation, but these locks are only held very briefly and in practice should almost never cause any blocking to occur. The existing blocking APIs of these classes are retained, but they simply delegate to the underlying non-blocking API and `Await` the results with indefinite timeouts. Associated JIRA ticket: https://issues.apache.org/jira/browse/SPARK-9026 Also fixes: https://issues.apache.org/jira/browse/SPARK-4514 This pull request contains all my own original work, which I release to the Spark project under its open source license. Author: Richard W. Eggert II <richard.eggert@gmail.com> Closes #9264 from reggert/fix-futureaction.
* [SPARK-9516][UI] Improvement of Thread Dump PageCodingCat2015-12-154-43/+118
| | | | | | | | | | | | | | | | | | https://issues.apache.org/jira/browse/SPARK-9516 - [x] new look of Thread Dump Page - [x] click column title to sort - [x] grep - [x] search as you type squito JoshRosen It's ready for the review now Author: CodingCat <zhunansjtu@gmail.com> Closes #7910 from CodingCat/SPARK-9516.
* [SPARK-12351][MESOS] Add documentation about submitting Spark with mesos ↵Timothy Chen2015-12-152-6/+35
| | | | | | | | | | cluster mode. Adding more documentation about submitting jobs with mesos cluster mode. Author: Timothy Chen <tnachen@gmail.com> Closes #10086 from tnachen/mesos_supervise_docs.
* [SPARK-12130] Replace shuffleManagerClass with shortShuffleMgrNames in ↵Lianhui Wang2015-12-158-12/+18
| | | | | | | | | | ExternalShuffleBlockResolver Replace shuffleManagerClassName with shortShuffleMgrName is to reduce time of string's comparison. and put sort's comparison on the front. cc JoshRosen andrewor14 Author: Lianhui Wang <lianhuiwang09@gmail.com> Closes #10131 from lianhuiwang/spark-12130.
* [SPARK-12056][CORE] Part 2 Create a TaskAttemptContext only after calling ↵tedyu2015-12-151-2/+2
| | | | | | | | | | | | | setConf This is continuation of SPARK-12056 where change is applied to SqlNewHadoopRDD.scala andrewor14 FYI Author: tedyu <yuzhihong@gmail.com> Closes #10164 from tedyu/master.
* [HOTFIX] Compile error from commit 31b3910Andrew Or2015-12-151-1/+1
|
* [SPARK-12105] [SQL] add convenient show functionsJean-Baptiste Onofré2015-12-151-9/+16
| | | | | | Author: Jean-Baptiste Onofré <jbonofre@apache.org> Closes #10130 from jbonofre/SPARK-12105.
* [SPARK-12236][SQL] JDBC filter tests all pass if filters are not really ↵hyukjinkwon2015-12-153-21/+19
| | | | | | | | | | | | | | | | pushed down https://issues.apache.org/jira/browse/SPARK-12236 Currently JDBC filters are not tested properly. All the tests pass even if the filters are not pushed down due to Spark-side filtering. In this PR, Firstly, I corrected the tests to properly check the pushed down filters by removing Spark-side filtering. Also, `!=` was being tested which is actually not pushed down. So I removed them. Lastly, I moved the `stripSparkFilter()` function to `SQLTestUtils` as this functions would be shared for all tests for pushed down filters. This function would be also shared with ORC datasource as the filters for that are also not being tested properly. Author: hyukjinkwon <gurwls223@gmail.com> Closes #10221 from HyukjinKwon/SPARK-12236.
* [SPARK-12271][SQL] Improve error message when Dataset.as[ ] has incompatible ↵Nong Li2015-12-154-7/+18
| | | | | | | | schemas. Author: Nong Li <nong@databricks.com> Closes #10260 from nongli/spark-11271.
* [MINOR][ML] Rename weights to coefficients for examples/DeveloperApiExampleYanbo Liang2015-12-152-19/+19
| | | | | | | | | | Rename ```weights``` to ```coefficients``` for examples/DeveloperApiExample. cc mengxr jkbradley Author: Yanbo Liang <ybliang8@gmail.com> Closes #10280 from yanboliang/spark-coefficients.
* [STREAMING][MINOR] Fix typo in function name of StateImpljerryshao2015-12-153-3/+3
| | | | | | | | cc\ tdas zsxwing , please review. Thanks a lot. Author: jerryshao <sshao@hortonworks.com> Closes #10305 from jerryshao/fix-typo-state-impl.
* [SPARK-12332][TRIVIAL][TEST] Fix minor typo in ResetSystemPropertiesHolden Karau2015-12-151-1/+1
| | | | | | | | Fix a minor typo (unbalanced bracket) in ResetSystemProperties. Author: Holden Karau <holden@us.ibm.com> Closes #10303 from holdenk/SPARK-12332-trivial-typo-in-ResetSystemProperties-comment.
* [SPARK-12288] [SQL] Support UnsafeRow in Coalesce/Except/Intersect.gatorsmile2015-12-142-1/+46
| | | | | | | | | | Support UnsafeRow for the Coalesce/Except/Intersect. Could you review if my code changes are ok? davies Thank you! Author: gatorsmile <gatorsmile@gmail.com> Closes #10285 from gatorsmile/unsafeSupportCIE.
* [SPARK-12188][SQL][FOLLOW-UP] Code refactoring and comment correction in ↵gatorsmile2015-12-141-1/+1
| | | | | | | | | | Dataset APIs marmbrus This PR is to address your comment. Thanks for your review! Author: gatorsmile <gatorsmile@gmail.com> Closes #10214 from gatorsmile/followup12188.
* [SPARK-12274][SQL] WrapOption should not have type constraint for childWenchen Fan2015-12-141-4/+1
| | | | | | | | I think it was a mistake, and we have not catched it so far until https://github.com/apache/spark/pull/10260 which begin to check if the `fromRowExpression` is resolved. Author: Wenchen Fan <wenchen@databricks.com> Closes #10263 from cloud-fan/encoder.
* [SPARK-12327] Disable commented code lintr temporarilyShivaram Venkataraman2015-12-141-1/+1
| | | | | | | | cc yhuai felixcheung shaneknapp Author: Shivaram Venkataraman <shivaram@cs.berkeley.edu> Closes #10300 from shivaram/comment-lintr-disable.
* [SPARK-12016] [MLLIB] [PYSPARK] Wrap Word2VecModel when loading it in pysparkLiang-Chi Hsieh2015-12-143-34/+67
| | | | | | | | | | JIRA: https://issues.apache.org/jira/browse/SPARK-12016 We should not directly use Word2VecModel in pyspark. We need to wrap it in a Word2VecModelWrapper when loading it in pyspark. Author: Liang-Chi Hsieh <viirya@appier.com> Closes #10100 from viirya/fix-load-py-wordvecmodel.
* [MINOR][DOC] Fix broken word2vec linkBenFradet2015-12-141-1/+1
| | | | | | | | Follow-up of [SPARK-12199](https://issues.apache.org/jira/browse/SPARK-12199) and #10193 where a broken link has been left as is. Author: BenFradet <benjamin.fradet@gmail.com> Closes #10282 from BenFradet/SPARK-12199.
* [SPARK-12275][SQL] No plan for BroadcastHint in some conditionyucai2015-12-132-1/+8
| | | | | | | | | | When SparkStrategies.BasicOperators's "case BroadcastHint(child) => apply(child)" is hit, it only recursively invokes BasicOperators.apply with this "child". It makes many strategies have no change to process this plan, which probably leads to "No plan" issue, so we use planLater to go through all strategies. https://issues.apache.org/jira/browse/SPARK-12275 Author: yucai <yucai.yu@intel.com> Closes #10265 from yucai/broadcast_hint.
* [SPARK-12213][SQL] use multiple partitions for single distinct queryDavies Liu2015-12-1310-990/+422
| | | | | | | | | | | | | | | | | | | | | | | | | Currently, we could generate different plans for query with single distinct (depends on spark.sql.specializeSingleDistinctAggPlanning), one works better on low cardinality columns, the other works better for high cardinality column (default one). This PR change to generate a single plan (three aggregations and two exchanges), which work better in both cases, then we could safely remove the flag `spark.sql.specializeSingleDistinctAggPlanning` (introduced in 1.6). For a query like `SELECT COUNT(DISTINCT a) FROM table` will be ``` AGG-4 (count distinct) Shuffle to a single reducer Partial-AGG-3 (count distinct, no grouping) Partial-AGG-2 (grouping on a) Shuffle by a Partial-AGG-1 (grouping on a) ``` This PR also includes large refactor for aggregation (reduce 500+ lines of code) cc yhuai nongli marmbrus Author: Davies Liu <davies@databricks.com> Closes #10228 from davies/single_distinct.
* [SPARK-12281][CORE] Fix a race condition when reporting ExecutorState in the ↵Shixiong Zhu2015-12-133-3/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | shutdown hook 1. Make sure workers and masters exit so that no worker or master will still be running when triggering the shutdown hook. 2. Set ExecutorState to FAILED if it's still RUNNING when executing the shutdown hook. This should fix the potential exceptions when exiting a local cluster ``` java.lang.AssertionError: assertion failed: executor 4 state transfer from RUNNING to RUNNING is illegal at scala.Predef$.assert(Predef.scala:179) at org.apache.spark.deploy.master.Master$$anonfun$receive$1.applyOrElse(Master.scala:260) at org.apache.spark.rpc.netty.Inbox$$anonfun$process$1.apply$mcV$sp(Inbox.scala:116) at org.apache.spark.rpc.netty.Inbox.safelyCall(Inbox.scala:204) at org.apache.spark.rpc.netty.Inbox.process(Inbox.scala:100) at org.apache.spark.rpc.netty.Dispatcher$MessageLoop.run(Dispatcher.scala:215) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) at java.lang.Thread.run(Thread.java:745) java.lang.IllegalStateException: Shutdown hooks cannot be modified during shutdown. at org.apache.spark.util.SparkShutdownHookManager.add(ShutdownHookManager.scala:246) at org.apache.spark.util.ShutdownHookManager$.addShutdownHook(ShutdownHookManager.scala:191) at org.apache.spark.util.ShutdownHookManager$.addShutdownHook(ShutdownHookManager.scala:180) at org.apache.spark.deploy.worker.ExecutorRunner.start(ExecutorRunner.scala:73) at org.apache.spark.deploy.worker.Worker$$anonfun$receive$1.applyOrElse(Worker.scala:474) at org.apache.spark.rpc.netty.Inbox$$anonfun$process$1.apply$mcV$sp(Inbox.scala:116) at org.apache.spark.rpc.netty.Inbox.safelyCall(Inbox.scala:204) at org.apache.spark.rpc.netty.Inbox.process(Inbox.scala:100) at org.apache.spark.rpc.netty.Dispatcher$MessageLoop.run(Dispatcher.scala:215) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) at java.lang.Thread.run(Thread.java:745) ``` Author: Shixiong Zhu <shixiong@databricks.com> Closes #10269 from zsxwing/executor-state.
* [SPARK-12267][CORE] Store the remote RpcEnv address to send the correct ↵Shixiong Zhu2015-12-124-1/+65
| | | | | | | | disconnetion message Author: Shixiong Zhu <shixiong@databricks.com> Closes #10261 from zsxwing/SPARK-12267.
* [SPARK-12199][DOC] Follow-up: Refine example code in ml-features.mdXusen Yin2015-12-124-15/+15
| | | | | | | | | | | | https://issues.apache.org/jira/browse/SPARK-12199 Follow-up PR of SPARK-11551. Fix some errors in ml-features.md mengxr Author: Xusen Yin <yinxusen@gmail.com> Closes #10193 from yinxusen/SPARK-12199.
* [SPARK-11193] Use Java ConcurrentHashMap instead of SynchronizedMap trait in ↵Jean-Baptiste Onofré2015-12-121-8/+8
| | | | | | | | order to avoid ClassCastException due to KryoSerializer in KinesisReceiver Author: Jean-Baptiste Onofré <jbonofre@apache.org> Closes #10203 from jbonofre/SPARK-11193.
* [SPARK-12158][SPARKR][SQL] Fix 'sample' functions that break R unit test casesgatorsmile2015-12-112-6/+15
| | | | | | | | | | | The existing sample functions miss the parameter `seed`, however, the corresponding function interface in `generics` has such a parameter. Thus, although the function caller can call the function with the 'seed', we are not using the value. This could cause SparkR unit tests failed. For example, I hit it in another PR: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/47213/consoleFull Author: gatorsmile <gatorsmile@gmail.com> Closes #10160 from gatorsmile/sampleR.
* [SPARK-12298][SQL] Fix infinite loop in DataFrame.sortWithinPartitionsAnkur Dave2015-12-112-3/+3
| | | | | | | | Modifies the String overload to call the Column overload and ensures this is called in a test. Author: Ankur Dave <ankurdave@gmail.com> Closes #10271 from ankurdave/SPARK-12298.
* [SPARK-11978][ML] Move dataset_example.py to examples/ml and rename to ↵Yanbo Liang2015-12-112-26/+38
| | | | | | | | | | | | | | dataframe_example.py Since ```Dataset``` has a new meaning in Spark 1.6, we should rename it to avoid confusion. #9873 finished the work of Scala example, here we focus on the Python one. Move dataset_example.py to ```examples/ml``` and rename to ```dataframe_example.py```. BTW, fix minor missing issues of #9873. cc mengxr Author: Yanbo Liang <ybliang8@gmail.com> Closes #9957 from yanboliang/SPARK-11978.
* [SPARK-12217][ML] Document invalid handling for StringIndexerBenFradet2015-12-111-0/+36
| | | | | | | | | | Added a paragraph regarding StringIndexer#setHandleInvalid to the ml-features documentation. I wonder if I should also add a snippet to the code example, input welcome. Author: BenFradet <benjamin.fradet@gmail.com> Closes #10257 from BenFradet/SPARK-12217.
* [SPARK-11497][MLLIB][PYTHON] PySpark RowMatrix Constructor Has Type Erasure ↵Mike Dusenberry2015-12-111-1/+1
| | | | | | | | | | | | | | Issue As noted in PR #9441, implementing `tallSkinnyQR` uncovered a bug with our PySpark `RowMatrix` constructor. As discussed on the dev list [here](http://apache-spark-developers-list.1001551.n3.nabble.com/K-Means-And-Class-Tags-td10038.html), there appears to be an issue with type erasure with RDDs coming from Java, and by extension from PySpark. Although we are attempting to construct a `RowMatrix` from an `RDD[Vector]` in [PythonMLlibAPI](https://github.com/apache/spark/blob/master/mllib/src/main/scala/org/apache/spark/mllib/api/python/PythonMLLibAPI.scala#L1115), the `Vector` type is erased, resulting in an `RDD[Object]`. Thus, when calling Scala's `tallSkinnyQR` from PySpark, we get a Java `ClassCastException` in which an `Object` cannot be cast to a Spark `Vector`. As noted in the aforementioned dev list thread, this issue was also encountered with `DecisionTrees`, and the fix involved an explicit `retag` of the RDD with a `Vector` type. `IndexedRowMatrix` and `CoordinateMatrix` do not appear to have this issue likely due to their related helper functions in `PythonMLlibAPI` creating the RDDs explicitly from DataFrames with pattern matching, thus preserving the types. This PR currently contains that retagging fix applied to the `createRowMatrix` helper function in `PythonMLlibAPI`. This PR blocks #9441, so once this is merged, the other can be rebased. cc holdenk Author: Mike Dusenberry <mwdusenb@us.ibm.com> Closes #9458 from dusenberrymw/SPARK-11497_PySpark_RowMatrix_Constructor_Has_Type_Erasure_Issue.
* [SPARK-12273][STREAMING] Make Spark Streaming web UI list Receivers in orderproflin2015-12-111-2/+3
| | | | | | | | | | Currently the Streaming web UI does NOT list Receivers in order; however, it seems more convenient for the users if Receivers are listed in order. ![spark-12273](https://cloud.githubusercontent.com/assets/15843379/11736602/0bb7f7a8-a00b-11e5-8e86-96ba9297fb12.png) Author: proflin <proflin.me@gmail.com> Closes #10264 from proflin/Spark-12273.
* [SPARK-11964][DOCS][ML] Add in Pipeline Import/Export Documentationanabranch2015-12-111-0/+13
| | | | | | | | | Adding in Pipeline Import and Export Documentation. Author: anabranch <wac.chambers@gmail.com> Author: Bill Chambers <wchambers@ischool.berkeley.edu> Closes #10179 from anabranch/master.
* [SPARK-12146][SPARKR] SparkR jsonFile should support multiple input filesYanbo Liang2015-12-115-116/+138
| | | | | | | | | | | | | | | | | * ```jsonFile``` should support multiple input files, such as: ```R jsonFile(sqlContext, c(“path1”, “path2”)) # character vector as arguments jsonFile(sqlContext, “path1,path2”) ``` * Meanwhile, ```jsonFile``` has been deprecated by Spark SQL and will be removed at Spark 2.0. So we mark ```jsonFile``` deprecated and use ```read.json``` at SparkR side. * Replace all ```jsonFile``` with ```read.json``` at test_sparkSQL.R, but still keep jsonFile test case. * If this PR is accepted, we should also make almost the same change for ```parquetFile```. cc felixcheung sun-rui shivaram Author: Yanbo Liang <ybliang8@gmail.com> Closes #10145 from yanboliang/spark-12146.
* [SPARK-12258] [SQL] passing null into ScalaUDF (follow-up)Davies Liu2015-12-112-16/+23
| | | | | | | | This is a follow-up PR for #10259 Author: Davies Liu <davies@databricks.com> Closes #10266 from davies/null_udf2.
* [SPARK-10991][ML] logistic regression training summary handle empty ↵Holden Karau2015-12-112-2/+29
| | | | | | | | | | | prediction col LogisticRegression training summary should still function if the predictionCol is set to an empty string or otherwise unset (related too https://issues.apache.org/jira/browse/SPARK-9718 ) Author: Holden Karau <holden@pigscanfly.ca> Author: Holden Karau <holden@us.ibm.com> Closes #9037 from holdenk/SPARK-10991-LogisticRegressionTrainingSummary-handle-empty-prediction-col.
* [SPARK-12258][SQL] passing null into ScalaUDFDavies Liu2015-12-102-6/+10
| | | | | | | | | | Check nullability and passing them into ScalaUDF. Closes #10249 Author: Davies Liu <davies@databricks.com> Closes #10259 from davies/udf_null.
* [STREAMING][DOC][MINOR] Update the description of direct Kafka stream docjerryshao2015-12-101-1/+1
| | | | | | | | | | With the merge of [SPARK-8337](https://issues.apache.org/jira/browse/SPARK-8337), now the Python API has the same functionalities compared to Scala/Java, so here changing the description to make it more precise. zsxwing tdas , please review, thanks a lot. Author: jerryshao <sshao@hortonworks.com> Closes #10246 from jerryshao/direct-kafka-doc-update.
* [SPARK-12155][SPARK-12253] Fix executor OOM in unified memory managementAndrew Or2015-12-104-31/+114
| | | | | | | | | | | | | | | | | | **Problem.** In unified memory management, acquiring execution memory may lead to eviction of storage memory. However, the space freed from evicting cached blocks is distributed among all active tasks. Thus, an incorrect upper bound on the execution memory per task can cause the acquisition to fail, leading to OOM's and premature spills. **Example.** Suppose total memory is 1000B, cached blocks occupy 900B, `spark.memory.storageFraction` is 0.4, and there are two active tasks. In this case, the cap on task execution memory is 100B / 2 = 50B. If task A tries to acquire 200B, it will evict 100B of storage but can only acquire 50B because of the incorrect cap. For another example, see this [regression test](https://github.com/andrewor14/spark/blob/fix-oom/core/src/test/scala/org/apache/spark/memory/UnifiedMemoryManagerSuite.scala#L233) that I stole from JoshRosen. **Solution.** Fix the cap on task execution memory. It should take into account the space that could have been freed by storage in addition to the current amount of memory available to execution. In the example above, the correct cap should have been 600B / 2 = 300B. This patch also guards against the race condition (SPARK-12253): (1) Existing tasks collectively occupy all execution memory (2) New task comes in and blocks while existing tasks spill (3) After tasks finish spilling, another task jumps in and puts in a large block, stealing the freed memory (4) New task still cannot acquire memory and goes back to sleep Author: Andrew Or <andrew@databricks.com> Closes #10240 from andrewor14/fix-oom.
* [SPARK-12251] Document and improve off-heap memory configurationsJosh Rosen2015-12-1015-22/+65
| | | | | | | | | | | | | This patch adds documentation for Spark configurations that affect off-heap memory and makes some naming and validation improvements for those configs. - Change `spark.memory.offHeapSize` to `spark.memory.offHeap.size`. This is fine because this configuration has not shipped in any Spark release yet (it's new in Spark 1.6). - Deprecated `spark.unsafe.offHeap` in favor of a new `spark.memory.offHeap.enabled` configuration. The motivation behind this change is to gather all memory-related configurations under the same prefix. - Add a check which prevents users from setting `spark.memory.offHeap.enabled=true` when `spark.memory.offHeap.size == 0`. After SPARK-11389 (#9344), which was committed in Spark 1.6, Spark enforces a hard limit on the amount of off-heap memory that it will allocate to tasks. As a result, enabling off-heap execution memory without setting `spark.memory.offHeap.size` will lead to immediate OOMs. The new configuration validation makes this scenario easier to diagnose, helping to avoid user confusion. - Document these configurations on the configuration page. Author: Josh Rosen <joshrosen@databricks.com> Closes #10237 from JoshRosen/SPARK-12251.
* [SPARK-11713] [PYSPARK] [STREAMING] Initial RDD updateStateByKey for PySparkBryan Cutler2015-12-104-5/+47
| | | | | | | | Adding ability to define an initial state RDD for use with updateStateByKey PySpark. Added unit test and changed stateful_network_wordcount example to use initial RDD. Author: Bryan Cutler <bjcutler@us.ibm.com> Closes #10082 from BryanCutler/initial-rdd-updateStateByKey-SPARK-11713.
* [SPARK-11563][CORE][REPL] Use RpcEnv to transfer REPL-generated classes.Marcelo Vanzin2015-12-1015-98/+183
| | | | | | | | | | | | | | | This avoids bringing up yet another HTTP server on the driver, and instead reuses the file server already managed by the driver's RpcEnv. As a bonus, the repl now inherits the security features of the network library. There's also a small change to create the directory for storing classes under the root temp dir for the application (instead of directly under java.io.tmpdir). Author: Marcelo Vanzin <vanzin@cloudera.com> Closes #9923 from vanzin/SPARK-11563.
* [SPARK-12212][ML][DOC] Clarifies the difference between spark.ml, ↵Timothy Hunter2015-12-1031-1793/+149
| | | | | | | | | | | | spark.mllib and mllib in the documentation. Replaces a number of occurences of `MLlib` in the documentation that were meant to refer to the `spark.mllib` package instead. It should clarify for new users the difference between `spark.mllib` (the package) and MLlib (the umbrella project for ML in spark). It also removes some files that I forgot to delete with #10207 Author: Timothy Hunter <timhunter@databricks.com> Closes #10234 from thunterdb/12212.
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-25 21:41:55 +0000