In the current implementation the mesos coarse scheduler does not wait for the mesos tasks to complete before ending the driver. This causes a race where the task has to finish cleaning up before the mesos driver terminates it with a SIGINT (and SIGKILL after 3 seconds if the SIGINT doesn't work).
This PR causes the mesos coarse scheduler to wait for the mesos tasks to finish (with a timeout defined by `spark.mesos.coarse.shutdown.ms`)
This PR also fixes a regression caused by [SPARK-10987] whereby submitting a shutdown causes a race between the local shutdown procedure and the notification of the scheduler driver disconnection. If the scheduler driver disconnection wins the race, the coarse executor incorrectly exits with status 1 (instead of the proper status 0)
With this patch the mesos coarse scheduler terminates properly, the executors clean up, and the tasks are reported as `FINISHED` in the Mesos console (as opposed to `KILLED` in < 1.6 or `FAILED` in 1.6 and later)
Author: Charles Allen <charles@allen-net.com>
Closes#10319 from drcrallen/SPARK-12330.
JIRA: https://issues.apache.org/jira/browse/SPARK-13113
As we shift bits right, looks like the bitwise AND operation is unnecessary.
Author: Liang-Chi Hsieh <viirya@gmail.com>
Closes#11002 from viirya/improve-decodepagenumber.
Make an internal non-deprecated version of incBytesRead and incRecordsRead so we don't have unecessary deprecation warnings in our build.
Right now incBytesRead and incRecordsRead are marked as deprecated and for internal use only. We should make private[spark] versions which are not deprecated and switch to those internally so as to not clutter up the warning messages when building.
cc andrewor14 who did the initial deprecation
Author: Holden Karau <holden@us.ibm.com>
Closes#11056 from holdenk/SPARK-13152-fix-task-metrics-deprecation-warnings.
Best time is stabler than average time, also added a column for nano seconds per row (which could be used to estimate contributions of each components in a query).
Having best time and average time together for more information (we can see kind of variance).
rate, time per row and relative are all calculated using best time.
The result looks like this:
```
Intel(R) Core(TM) i7-4558U CPU 2.80GHz
rang/filter/sum: Best/Avg Time(ms) Rate(M/s) Per Row(ns) Relative
-------------------------------------------------------------------------------------------
rang/filter/sum codegen=false 14332 / 16646 36.0 27.8 1.0X
rang/filter/sum codegen=true 845 / 940 620.0 1.6 17.0X
```
Author: Davies Liu <davies@databricks.com>
Closes#11018 from davies/gen_bench.
`rpcEnv.awaitTermination()` was not added in #10854 because some Streaming Python tests hung forever.
This patch fixed the hung issue and added rpcEnv.awaitTermination() back to SparkEnv.
Previously, Streaming Kafka Python tests shutdowns the zookeeper server before stopping StreamingContext. Then when stopping StreamingContext, KafkaReceiver may be hung due to https://issues.apache.org/jira/browse/KAFKA-601, hence, some thread of RpcEnv's Dispatcher cannot exit and rpcEnv.awaitTermination is hung.The patch just changed the shutdown order to fix it.
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#11031 from zsxwing/awaitTermination.
https://issues.apache.org/jira/browse/SPARK-13122
A race condition can occur in MemoryStore's unrollSafely() method if two threads that
return the same value for currentTaskAttemptId() execute this method concurrently. This
change makes the operation of reading the initial amount of unroll memory used, performing
the unroll, and updating the associated memory maps atomic in order to avoid this race
condition.
Initial proposed fix wraps all of unrollSafely() in a memoryManager.synchronized { } block. A cleaner approach might be introduce a mechanism that synchronizes based on task attempt ID. An alternative option might be to track unroll/pending unroll memory based on block ID rather than task attempt ID.
Author: Adam Budde <budde@amazon.com>
Closes#11012 from budde/master.
The driver filesystem is likely different from where the executors will run, so resolving paths (and symlinks, etc.) will lead to invalid paths on executors.
Author: Iulian Dragos <jaguarul@gmail.com>
Closes#10923 from dragos/issue/canonical-paths.
This takes over #10729 and makes sure that `spark-shell` fails with a proper error message. There is a slight behavioral change: before this change `spark-shell` would exit, while now the REPL is still there, but `sc` and `sqlContext` are not defined and the error is visible to the user.
Author: Nilanjan Raychaudhuri <nraychaudhuri@gmail.com>
Author: Iulian Dragos <jaguarul@gmail.com>
Closes#10921 from dragos/pr/10729.
Fix zookeeper dir configuration used in cluster mode, and also add documentation around these settings.
Author: Timothy Chen <tnachen@gmail.com>
Closes#10057 from tnachen/fix_mesos_dir.
Add a local property to indicate if checkpointing all RDDs that are marked with the checkpoint flag, and enable it in Streaming
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10934 from zsxwing/recursive-checkpoint.
This patch changes Spark's build to make Scala 2.11 the default Scala version. To be clear, this does not mean that Spark will stop supporting Scala 2.10: users will still be able to compile Spark for Scala 2.10 by following the instructions on the "Building Spark" page; however, it does mean that Scala 2.11 will be the default Scala version used by our CI builds (including pull request builds).
The Scala 2.11 compiler is faster than 2.10, so I think we'll be able to look forward to a slight speedup in our CI builds (it looks like it's about 2X faster for the Maven compile-only builds, for instance).
After this patch is merged, I'll update Jenkins to add new compile-only jobs to ensure that Scala 2.10 compilation doesn't break.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#10608 from JoshRosen/SPARK-6363.
This issue is causing tests to fail consistently in master with Hadoop 2.6 / 2.7. This is because for Hadoop 2.5+ we overwrite existing values of `InputMetrics#bytesRead` in each call to `HadoopRDD#compute`. In the case of coalesce, e.g.
```
sc.textFile(..., 4).coalesce(2).count()
```
we will call `compute` multiple times in the same task, overwriting `bytesRead` values from previous calls to `compute`.
For a regression test, see `InputOutputMetricsSuite.input metrics for old hadoop with coalesce`. I did not add a new regression test because it's impossible without significant refactoring; there's a lot of existing duplicate code in this corner of Spark.
This was caused by #10835.
Author: Andrew Or <andrew@databricks.com>
Closes#10973 from andrewor14/fix-input-metrics-coalesce.
Apparently chrome removed `SVGElement.prototype.getTransformToElement`, which is used by our JS library dagre-d3 when creating edges. The real diff can be found here: 7d6c0002e4, which is taken from the fix in the main repo: 1ef067f1c6
Upstream issue: https://github.com/cpettitt/dagre-d3/issues/202
Author: Andrew Or <andrew@databricks.com>
Closes#10986 from andrewor14/fix-dag-viz.
This is an existing issue uncovered recently by #10835. The reason for the exception was because the `SQLHistoryListener` gets all sorts of accumulators, not just the ones that represent SQL metrics. For example, the listener gets the `internal.metrics.shuffleRead.remoteBlocksFetched`, which is an Int, then it proceeds to cast the Int to a Long, which fails.
The fix is to mark accumulators representing SQL metrics using some internal metadata. Then we can identify which ones are SQL metrics and only process those in the `SQLHistoryListener`.
Author: Andrew Or <andrew@databricks.com>
Closes#10971 from andrewor14/fix-sql-history.
[SPARK-10873] Support column sort and search for History Server using jQuery DataTable and REST API. Before this commit, the history server was generated hard-coded html and can not support search, also, the sorting was disabled if there is any application that has more than one attempt. Supporting search and sort (over all applications rather than the 20 entries in the current page) in any case will greatly improve user experience.
1. Create the historypage-template.html for displaying application information in datables.
2. historypage.js uses jQuery to access the data from /api/v1/applications REST API, and use DataTable to display each application's information. For application that has more than one attempt, the RowsGroup is used to merge such entries while at the same time supporting sort and search.
3. "duration" and "lastUpdated" rest API are added to application's "attempts".
4. External javascirpt and css files for datatables, RowsGroup and jquery plugins are added with licenses clarified.
Snapshots for how it looks like now:
History page view:
Search:
Sort by started time:
Author: zhuol <zhuol@yahoo-inc.com>
Closes#10648 from zhuoliu/10873.
by explicitly marking annotated parameters as vals (SI-8813).
Caused by #10835.
Author: Andrew Or <andrew@databricks.com>
Closes#10955 from andrewor14/fix-scala211.
Spark's `Partition` and `RDD.partitions` APIs have a contract which requires custom implementations of `RDD.partitions` to ensure that for all `x`, `rdd.partitions(x).index == x`; in other words, the `index` reported by a repartition needs to match its position in the partitions array.
If a custom RDD implementation violates this contract, then Spark has the potential to become stuck in an infinite recomputation loop when recomputing a subset of an RDD's partitions, since the tasks that are actually run will not correspond to the missing output partitions that triggered the recomputation. Here's a link to a notebook which demonstrates this problem: 5e8a5aa8d2/Violating%2520RDD.partitions%2520contract.html
In order to guard against this infinite loop behavior, this patch modifies Spark so that it fails fast and refuses to compute RDDs' whose `partitions` violate the API contract.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#10932 from JoshRosen/SPARK-13021.
The high level idea is that instead of having the executors send both accumulator updates and TaskMetrics, we should have them send only accumulator updates. This eliminates the need to maintain both code paths since one can be implemented in terms of the other. This effort is split into two parts:
**SPARK-12895: Implement TaskMetrics using accumulators.** TaskMetrics is basically just a bunch of accumulable fields. This patch makes TaskMetrics a syntactic wrapper around a collection of accumulators so we don't need to send TaskMetrics from the executors to the driver.
**SPARK-12896: Send only accumulator updates to the driver.** Now that TaskMetrics are expressed in terms of accumulators, we can capture all TaskMetrics values if we just send accumulator updates from the executors to the driver. This completes the parent issue SPARK-10620.
While an effort has been made to preserve as much of the public API as possible, there were a few known breaking DeveloperApi changes that would be very awkward to maintain. I will gather the full list shortly and post it here.
Note: This was once part of #10717. This patch is split out into its own patch from there to make it easier for others to review. Other smaller pieces of already been merged into master.
Author: Andrew Or <andrew@databricks.com>
Closes#10835 from andrewor14/task-metrics-use-accums.
If there's an RPC issue while sparkContext is alive but stopped (which would happen only when executing SparkContext.stop), log a warning instead. This is a common occurrence.
vanzin
Author: Nishkam Ravi <nishkamravi@gmail.com>
Author: nishkamravi2 <nishkamravi@gmail.com>
Closes#10881 from nishkamravi2/master_netty.
Right now RpcEndpointRef.ask may throw exception in some corner cases, such as calling ask after stopping RpcEnv. It's better to avoid throwing exception from RpcEndpointRef.ask. We can send the exception to the future for `ask`.
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10568 from zsxwing/send-ask-fail.
Call system.exit explicitly to make sure non-daemon user threads terminate. Without this, user applications might live forever if the cluster manager does not appropriately kill them. E.g., YARN had this bug: HADOOP-12441.
Author: zhuol <zhuol@yahoo-inc.com>
Closes#9946 from zhuoliu/10911.
Fix Java function API methods for flatMap and mapPartitions to require producing only an Iterator, not Iterable. Also fix DStream.flatMap to require a function producing TraversableOnce only, not Traversable.
CC rxin pwendell for API change; tdas since it also touches streaming.
Author: Sean Owen <sowen@cloudera.com>
Closes#10413 from srowen/SPARK-3369.
JIRA: https://issues.apache.org/jira/browse/SPARK-12961
To prevent memory leak in snappy-java, just call the method once and cache the result. After the library releases new version, we can remove this object.
JoshRosen
Author: Liang-Chi Hsieh <viirya@gmail.com>
Closes#10875 from viirya/prevent-snappy-memory-leak.
Added color coding to the Executors page for Active Tasks, Failed Tasks, Completed Tasks and Task Time.
Active Tasks is shaded blue with it's range based on percentage of total cores used.
Failed Tasks is shaded red ranging over the first 10% of total tasks failed
Completed Tasks is shaded green ranging over 10% of total tasks including failed and active tasks, but only when there are active or failed tasks on that executor.
Task Time is shaded red when GC Time goes over 10% of total time with it's range directly corresponding to the percent of total time.
Author: Alex Bozarth <ajbozart@us.ibm.com>
Closes#10154 from ajbozarth/spark12149.
[SPARK-12755][CORE] Stop the event logger before the DAG scheduler to avoid a race condition where the standalone master attempts to build the app's history UI before the event log is stopped.
This contribution is my original work, and I license this work to the Spark project under the project's open source license.
Author: Michael Allman <michael@videoamp.com>
Closes#10700 from mallman/stop_event_logger_first.
- Remove Akka dependency from core. Note: the streaming-akka project still uses Akka.
- Remove HttpFileServer
- Remove Akka configs from SparkConf and SSLOptions
- Rename `spark.akka.frameSize` to `spark.rpc.message.maxSize`. I think it's still worth to keep this config because using `DirectTaskResult` or `IndirectTaskResult` depends on it.
- Update comments and docs
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10854 from zsxwing/remove-akka.
Including the following changes:
1. Add StreamingListenerForwardingBus to WrappedStreamingListenerEvent process events in `onOtherEvent` to StreamingListener
2. Remove StreamingListenerBus
3. Merge AsynchronousListenerBus and LiveListenerBus to the same class LiveListenerBus
4. Add `logEvent` method to SparkListenerEvent so that EventLoggingListener can use it to ignore WrappedStreamingListenerEvents
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10779 from zsxwing/streaming-listener.
This is a step in implementing SPARK-10620, which migrates TaskMetrics to accumulators.
TaskMetrics has a bunch of var's, some are fully public, some are `private[spark]`. This is bad coding style that makes it easy to accidentally overwrite previously set metrics. This has happened a few times in the past and caused bugs that were difficult to debug.
Instead, we should have get-or-create semantics, which are more readily understandable. This makes sense in the case of TaskMetrics because these are just aggregated metrics that we want to collect throughout the task, so it doesn't matter who's incrementing them.
Parent PR: #10717
Author: Andrew Or <andrew@databricks.com>
Author: Josh Rosen <joshrosen@databricks.com>
Author: andrewor14 <andrew@databricks.com>
Closes#10815 from andrewor14/get-or-create-metrics.
This is a small step in implementing SPARK-10620, which migrates TaskMetrics to accumulators. This patch is strictly a cleanup patch and introduces no change in functionality. It literally just renames 3 fields for consistency. Today we have:
```
inputMetrics.recordsRead
outputMetrics.bytesWritten
shuffleReadMetrics.localBlocksFetched
...
shuffleWriteMetrics.shuffleRecordsWritten
shuffleWriteMetrics.shuffleBytesWritten
shuffleWriteMetrics.shuffleWriteTime
```
The shuffle write ones are kind of redundant. We can drop the `shuffle` part in the method names. I added backward compatible (but deprecated) methods with the old names.
Parent PR: #10717
Author: Andrew Or <andrew@databricks.com>
Closes#10811 from andrewor14/rename-things.
This patch refactors portions of the BlockManager and CacheManager in order to avoid having to pass `evictedBlocks` lists throughout the code. It appears that these lists were only consumed by `TaskContext.taskMetrics`, so the new code now directly updates the metrics from the lower-level BlockManager methods.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#10776 from JoshRosen/SPARK-10985.
This is a small step in implementing SPARK-10620, which migrates `TaskMetrics` to accumulators. This patch is strictly a cleanup patch and introduces no change in functionality. It literally just moves classes to their own files to avoid having single monolithic ones that contain 10 different classes.
Parent PR: #10717
Author: Andrew Or <andrew@databricks.com>
Closes#10810 from andrewor14/move-things.
This inlines a few of the Parquet decoders and adds vectorized APIs to support decoding in batch.
There are a few particulars in the Parquet encodings that make this much more efficient. In
particular, RLE encodings are very well suited for batch decoding. The Parquet 2.0 encodings are
also very suited for this.
This is a work in progress and does not affect the current execution. In subsequent patches, we will
support more encodings and types before enabling this.
Simple benchmarks indicate this can decode single ints about > 3x faster.
Author: Nong Li <nong@databricks.com>
Author: Nong <nongli@gmail.com>
Closes#10593 from nongli/spark-12644.
Added a Totals table to the top of the page to display the totals of each applicable column in the executors table.
Old Description:
~~Created a TOTALS row containing the totals of each column in the executors UI. By default the TOTALS row appears at the top of the table. When a column is sorted the TOTALS row will always sort to either the top or bottom of the table.~~
Author: Alex Bozarth <ajbozart@us.ibm.com>
Closes#10668 from ajbozarth/spark12716.
This pull request removes the external block store API. This is rarely used, and the file system interface is actually a better, more standard way to interact with external storage systems.
There are some other things to remove also, as pointed out by JoshRosen. We will do those as follow-up pull requests.
Author: Reynold Xin <rxin@databricks.com>
Closes#10752 from rxin/remove-offheap.
This patch significantly speeds up the BlockManagerSuite's "SPARK-9591: getRemoteBytes from another location when Exception throw" test, reducing the test time from 45s to ~250ms. The key change was to set `spark.shuffle.io.maxRetries` to 0 (the code previously set `spark.network.timeout` to `2s`, but this didn't make a difference because the slowdown was not due to this timeout).
Along the way, I also cleaned up the way that we handle SparkConf in BlockManagerSuite: previously, each test would mutate a shared SparkConf instance, while now each test gets a fresh SparkConf.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#10759 from JoshRosen/SPARK-12174.
Add `listener.synchronized` to get `storageStatusList` and `execInfo` atomically.
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10728 from zsxwing/SPARK-12784.
When an Executor process is destroyed, the FileAppender that is asynchronously reading the stderr stream of the process can throw an IOException during read because the stream is closed. Before the ExecutorRunner destroys the process, the FileAppender thread is flagged to stop. This PR wraps the inputStream.read call of the FileAppender in a try/catch block so that if an IOException is thrown and the thread has been flagged to stop, it will safely ignore the exception. Additionally, the FileAppender thread was changed to use Utils.tryWithSafeFinally to better log any exception that do occur. Added unit tests to verify a IOException is thrown and logged if FileAppender is not flagged to stop, and that no IOException when the flag is set.
Author: Bryan Cutler <cutlerb@gmail.com>
Closes#10714 from BryanCutler/file-appender-read-ioexception-SPARK-9844.
We've already removed local execution but didn't deprecate `TaskContext.isRunningLocally()`; we should deprecate it for 2.0.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#10751 from JoshRosen/remove-local-exec-from-taskcontext.
This problem lies in `BypassMergeSortShuffleWriter`, empty partition will also generate a temp shuffle file with several bytes. So here change to only create file when partition is not empty.
This problem only lies in here, no such issue in `HashShuffleWriter`.
Please help to review, thanks a lot.
Author: jerryshao <sshao@hortonworks.com>
Closes#10376 from jerryshao/SPARK-12400.
I hit the exception below. The `UnsafeKVExternalSorter` does pass `null` as the consumer when creating an `UnsafeInMemorySorter`. Normally the NPE doesn't occur because the `inMemSorter` is set to null later and the `free()` method is not called. It happens when there is another exception like OOM thrown before setting `inMemSorter` to null. Anyway, we can add the null check to avoid it.
```
ERROR spark.TaskContextImpl: Error in TaskCompletionListener
java.lang.NullPointerException
at org.apache.spark.util.collection.unsafe.sort.UnsafeInMemorySorter.free(UnsafeInMemorySorter.java:110)
at org.apache.spark.util.collection.unsafe.sort.UnsafeExternalSorter.cleanupResources(UnsafeExternalSorter.java:288)
at org.apache.spark.util.collection.unsafe.sort.UnsafeExternalSorter$1.onTaskCompletion(UnsafeExternalSorter.java:141)
at org.apache.spark.TaskContextImpl$$anonfun$markTaskCompleted$1.apply(TaskContextImpl.scala:79)
at org.apache.spark.TaskContextImpl$$anonfun$markTaskCompleted$1.apply(TaskContextImpl.scala:77)
at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59)
at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47)
at org.apache.spark.TaskContextImpl.markTaskCompleted(TaskContextImpl.scala:77)
at org.apache.spark.scheduler.Task.run(Task.scala:91)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:213)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1110)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:603)
at java.lang.Thread.run(Thread.java:722)
```
Author: Carson Wang <carson.wang@intel.com>
Closes#10637 from carsonwang/FixNPE.
Fix the style violation (space before , and :).
This PR is a followup for #10643
Author: Kousuke Saruta <sarutak@oss.nttdata.co.jp>
Closes#10719 from sarutak/SPARK-12692-followup-core.
- [x] Upgrade Py4J to 0.9.1
- [x] SPARK-12657: Revert SPARK-12617
- [x] SPARK-12658: Revert SPARK-12511
- Still keep the change that only reading checkpoint once. This is a manual change and worth to take a look carefully. bfd4b5c040
- [x] Verify no leak any more after reverting our workarounds
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10692 from zsxwing/py4j-0.9.1.
