Charles Reiss
feadaf72f4
Mark key as not loading in CacheTracker even when compute() fails
2013-01-01 07:57:20 -08:00
Josh Rosen
f803953998
Raise exception when hashing Java arrays (SPARK-597)
2012-12-31 20:20:11 -08:00
Tathagata Das
7e0271b438
Refactored a whole lot to push all DStreams into the spark.streaming.dstream package.
2012-12-30 15:19:55 -08:00
Tathagata Das
9e644402c1
Improved jekyll and scala docs. Made many classes and method private to remove them from scala docs.
2012-12-29 18:31:51 -08:00
Josh Rosen
59195c68ec
Update PySpark for compatibility with TaskContext.
2012-12-29 16:01:03 -08:00
Josh Rosen
c5cee53f20
Merge remote-tracking branch 'origin/master' into python-api
...
Conflicts:
docs/quick-start.md
2012-12-29 16:00:51 -08:00
Josh Rosen
7ec3595de2
Fix bug (introduced by batching) in PySpark take()
2012-12-28 22:21:16 -08:00
Josh Rosen
397e67103c
Change Utils.fetchFile() warning to SparkException.
2012-12-28 17:37:13 -08:00
Josh Rosen
d64fa72d2e
Add addFile() and addJar() to JavaSparkContext.
2012-12-28 17:00:57 -08:00
Josh Rosen
bd237d4a9d
Add synchronization to LocalScheduler.updateDependencies().
2012-12-28 17:00:57 -08:00
Josh Rosen
f1bf4f0385
Skip deletion of files in clearFiles().
...
This fixes an issue where Spark could delete
original files in the current working directory
that were added to the job using addFile().
There was also the potential for addFile() to
overwrite local files, which is addressed by
changing Utils.fetchFile() to log a warning
instead of overwriting a file with new contents.
This is a short-term fix; a better long-term
solution would be to remove the dependence on
storing files in the current working directory,
since we can't change the cwd from Java.
2012-12-28 17:00:57 -08:00
Josh Rosen
fbadb1cda5
Mark api.python classes as private; echo Java output to stderr.
2012-12-28 09:06:11 -08:00
Tathagata Das
0bc0a60d30
Modifications to make sure LocalScheduler terminate cleanly without errors when SparkContext is shutdown, to minimize spurious exception during master failure tests.
2012-12-27 15:37:33 -08:00
Tathagata Das
7c33f76291
Merge branch 'mesos' into dev-merge
2012-12-26 19:19:07 -08:00
Tathagata Das
836042bb9f
Merge branch 'dev-checkpoint' of github.com:radlab/spark into dev-merge
...
Conflicts:
core/src/main/scala/spark/ParallelCollection.scala
core/src/main/scala/spark/RDD.scala
core/src/main/scala/spark/rdd/BlockRDD.scala
core/src/main/scala/spark/rdd/CartesianRDD.scala
core/src/main/scala/spark/rdd/CoGroupedRDD.scala
core/src/main/scala/spark/rdd/CoalescedRDD.scala
core/src/main/scala/spark/rdd/FilteredRDD.scala
core/src/main/scala/spark/rdd/FlatMappedRDD.scala
core/src/main/scala/spark/rdd/GlommedRDD.scala
core/src/main/scala/spark/rdd/HadoopRDD.scala
core/src/main/scala/spark/rdd/MapPartitionsRDD.scala
core/src/main/scala/spark/rdd/MapPartitionsWithSplitRDD.scala
core/src/main/scala/spark/rdd/MappedRDD.scala
core/src/main/scala/spark/rdd/PipedRDD.scala
core/src/main/scala/spark/rdd/SampledRDD.scala
core/src/main/scala/spark/rdd/ShuffledRDD.scala
core/src/main/scala/spark/rdd/UnionRDD.scala
core/src/main/scala/spark/scheduler/ResultTask.scala
core/src/test/scala/spark/CheckpointSuite.scala
2012-12-26 19:09:01 -08:00
Josh Rosen
1dca0c5180
Remove debug output from PythonPartitioner.
2012-12-26 18:23:06 -08:00
Josh Rosen
4608902fb8
Use filesystem to collect RDDs in PySpark.
...
Passing large volumes of data through Py4J seems
to be slow. It appears to be faster to write the
data to the local filesystem and read it back from
Python.
2012-12-24 17:20:10 -08:00
Mark Hamstra
903f3518df
fall back to filter-map-collect when calling lookup() on an RDD without a partitioner
2012-12-24 13:18:45 -08:00
Mark Hamstra
61be8566e2
Allow distinct() to be called without parentheses when using the default number of splits.
2012-12-24 02:36:47 -08:00
Reynold Xin
60f7338092
Remove the call to close input stream in Kryo serializer.
2012-12-21 15:49:33 -08:00
Matei Zaharia
3334b7c6b5
Merge pull request #341 from rxin/4a3fb06ac2d11125feb08acbbd4df76d1e91b677
...
Kryo2 update against Spark master
2012-12-21 15:31:23 -08:00
Reynold Xin
eac566a7f4
Merge branch 'master' of github.com:mesos/spark into dev
...
Conflicts:
core/src/main/scala/spark/MapOutputTracker.scala
core/src/main/scala/spark/PairRDDFunctions.scala
core/src/main/scala/spark/ParallelCollection.scala
core/src/main/scala/spark/RDD.scala
core/src/main/scala/spark/rdd/BlockRDD.scala
core/src/main/scala/spark/rdd/CartesianRDD.scala
core/src/main/scala/spark/rdd/CoGroupedRDD.scala
core/src/main/scala/spark/rdd/CoalescedRDD.scala
core/src/main/scala/spark/rdd/FilteredRDD.scala
core/src/main/scala/spark/rdd/FlatMappedRDD.scala
core/src/main/scala/spark/rdd/GlommedRDD.scala
core/src/main/scala/spark/rdd/HadoopRDD.scala
core/src/main/scala/spark/rdd/MapPartitionsRDD.scala
core/src/main/scala/spark/rdd/MapPartitionsWithSplitRDD.scala
core/src/main/scala/spark/rdd/MappedRDD.scala
core/src/main/scala/spark/rdd/PipedRDD.scala
core/src/main/scala/spark/rdd/SampledRDD.scala
core/src/main/scala/spark/rdd/ShuffledRDD.scala
core/src/main/scala/spark/rdd/UnionRDD.scala
core/src/main/scala/spark/storage/BlockManager.scala
core/src/main/scala/spark/storage/BlockManagerId.scala
core/src/main/scala/spark/storage/BlockManagerMaster.scala
core/src/main/scala/spark/storage/StorageLevel.scala
core/src/main/scala/spark/util/MetadataCleaner.scala
core/src/main/scala/spark/util/TimeStampedHashMap.scala
core/src/test/scala/spark/storage/BlockManagerSuite.scala
run
2012-12-20 14:53:40 -08:00
Tathagata Das
8512dd3225
Merge branch 'dev' of github.com:radlab/spark into dev-checkpoint
...
Conflicts:
core/src/main/scala/spark/ParallelCollection.scala
core/src/test/scala/spark/CheckpointSuite.scala
streaming/src/main/scala/spark/streaming/DStream.scala
2012-12-20 14:24:19 -08:00
Tathagata Das
fe777eb77d
Fixed bugs in CheckpointRDD and spark.CheckpointSuite.
2012-12-20 13:39:27 -08:00
Tathagata Das
f9c5b0a6fe
Changed checkpoint writing and reading process.
2012-12-20 11:52:23 -08:00
Matei Zaharia
5e51b889fe
Merge pull request #327 from rxin/spark-633
...
Added the ability in block manager to remove blocks.
2012-12-20 11:33:38 -08:00
Reynold Xin
9397c5014e
Let the slave notify the master block removal.
2012-12-20 01:37:09 -08:00
Reynold Xin
68c52d80ec
Moved BlockManager's IdGenerator into BlockManager object. Removed some
...
excessive debug messages.
2012-12-19 15:27:23 -08:00
Tathagata Das
5184141936
Introduced getSpits, getDependencies, and getPreferredLocations in RDD and RDDCheckpointData.
2012-12-18 13:30:53 -08:00
Patrick Wendell
bfac06e1f6
SPARK-616: Logging dead workers in Web UI.
...
This patch keeps track of which workers have died and marks them
as such in the master web UI. It also handles workers which die and
re-register using different actor ID's.
2012-12-17 23:09:05 -08:00
Tathagata Das
72eed2b95e
Converted CheckpointState in RDDCheckpointData to use scala Enumeration.
2012-12-17 18:52:43 -08:00
Matei Zaharia
b82a6dd2c7
Merge pull request #332 from JoshRosen/spark-607
...
Add try-finally to handle MapOutputTracker timeouts
2012-12-14 11:41:16 -08:00
Reynold Xin
06f855c24d
Merge branch 'spark-633' of github.com:rxin/spark into spark-633
2012-12-14 00:27:24 -08:00
Reynold Xin
8c01295b85
Fixed conflicts from merging Charles' and TD's block manager changes.
2012-12-14 00:26:36 -08:00
Charles Reiss
c528932a41
Code review cleanup.
2012-12-13 22:37:16 -08:00
Charles Reiss
0aad42b5e7
Have standalone cluster report exit codes to clients. Addresses SPARK-639.
2012-12-13 22:37:16 -08:00
Reynold Xin
97434f49b8
Merged TD's block manager refactoring.
2012-12-13 22:32:19 -08:00
Reynold Xin
41e58a519a
Merge branch 'master' of github.com:mesos/spark into spark-633
2012-12-13 22:06:47 -08:00
Josh Rosen
cf52d9cade
Add try-finally to handle MapOutputTracker timeouts.
2012-12-13 21:53:30 -08:00
Matei Zaharia
05e225f988
Merge pull request #329 from woggling/executor-status-codes
...
Executor exit status codes
2012-12-13 20:14:10 -08:00
Charles Reiss
b054d3b222
ExecutorLostReason -> ExecutorLossReason
2012-12-13 18:44:07 -08:00
Charles Reiss
24d7aa2d15
Extra whitespace in ExecutorExitCode
2012-12-13 18:39:23 -08:00
Reynold Xin
dc7d7fc286
Merge branch 'master' of github.com:mesos/spark into spark-633
2012-12-13 16:48:34 -08:00
Reynold Xin
4f076e105e
SPARK-635: Pass a TaskContext object to compute() interface and use
...
that to close Hadoop input stream. Incorporated Matei's command.
2012-12-13 16:41:15 -08:00
Charles Reiss
829206f1a7
Explain slaveLost calls made by StandaloneSchedulerBackend
2012-12-13 16:23:36 -08:00
Charles Reiss
a4041dd87f
Log duplicate slaveLost() calls in ClusterScheduler.
2012-12-13 16:23:36 -08:00
Charles Reiss
fa9df4a45d
Normalize executor exit statuses and report them to the user.
2012-12-13 16:23:31 -08:00
Reynold Xin
eacb98e900
SPARK-635: Pass a TaskContext object to compute() interface and use that
...
to close Hadoop input stream.
2012-12-13 15:41:53 -08:00
Josh Rosen
7c9e3d1c21
Return success or failure in BlockStore.remove().
2012-12-13 15:22:27 -08:00
Reynold Xin
1b7a0451ed
Added the ability in block manager to remove blocks.
2012-12-13 00:04:42 -08:00
Charles Reiss
1d8e2e6cff
Call slaveLost on executor death for standalone clusters.
2012-12-12 21:15:34 -08:00
Tathagata Das
8e74fac215
Made checkpoint data in RDDs optional to further reduce serialized size.
2012-12-11 15:36:12 -08:00
Tathagata Das
fa28f25619
Fixed bug in UnionRDD and CoGroupedRDD
2012-12-11 13:59:43 -08:00
Tathagata Das
746afc2e65
Bunch of bug fixes related to checkpointing in RDDs. RDDCheckpointData object is used to lock all serialization and dependency changes for checkpointing. ResultTask converted to Externalizable and serialized RDD is cached like ShuffleMapTask.
2012-12-10 23:36:37 -08:00
Reynold Xin
21b271f5bd
Suppress shuffle block updates when a slave node comes back.
2012-12-10 20:36:03 -08:00
Matei Zaharia
a1a2daa7ef
Merge pull request #317 from woggling/block-manager-heartbeat
...
Implement block manager heartbeat
2012-12-10 11:03:55 -08:00
Charles Reiss
b6b62d774f
Decrease BlockManagerMaster logging verbosity
2012-12-10 00:31:55 -08:00
Charles Reiss
5d3e917d09
Use Akka scheduler for BlockManager heart beats.
...
Adds required ActorSystem argument to BlockManager constructors.
2012-12-10 00:31:50 -08:00
Charles Reiss
b53dd28c90
Changed default block manager heartbeat interval to 5 s
2012-12-09 23:03:34 -08:00
Matei Zaharia
e1d7cd2276
Search for a non-loopback address in Utils.getLocalIpAddress
2012-12-08 00:33:11 -08:00
Patrick Wendell
3e796bdd57
Changes in response to TD's review.
2012-12-07 19:34:05 -08:00
Patrick Wendell
c36ca10241
Adding locality aware parallelize
2012-12-07 16:42:36 -08:00
Tathagata Das
1f3a75ae9e
Modified checkpoint testsuite to more comprehensively test checkpointing of various RDDs. Fixed checkpoint bug (splits referring to parent RDDs or parent splits) in UnionRDD and CoalescedRDD. Fixed bug in testing ShuffledRDD. Removed unnecessary and useless map-side combining step for narrow dependencies in CoGroupedRDD. Removed unncessary WeakReference stuff from many other RDDs.
2012-12-07 13:45:52 -08:00
Charles Reiss
714c8d32d5
Don't divide by milliseconds by 1000 more.
2012-12-06 18:38:34 -08:00
Charles Reiss
8f0819520c
map -> foreach
2012-12-06 18:29:50 -08:00
Charles Reiss
7a033fd795
Make LocalSparkCluster use distinct IPs
2012-12-06 00:03:08 -08:00
Charles Reiss
d21ca010ac
Add block manager heart beats.
...
Renames old message called 'HeartBeat' to 'BlockUpdate'.
The BlockManager periodically sends a heart beat message to the master.
If the manager is currently not registered. The master responds to the
heart beat by indicating whether the BlockManager is currently registered
with the master. Additionally, the master now also responds to block
updates by indicating whether the BlockManager in question is registered.
When the BlockManager detects (by heart beat or failed block update)
that it stopped being registered, it reregisters and sends block
updates for all its blocks.
2012-12-05 23:35:20 -08:00
Charles Reiss
c9e54a6755
Track block managers by hostname; handle manager removal.
2012-12-05 23:35:20 -08:00
Charles Reiss
5afa2ee9e9
Actually put millis in _lastSeenMs
2012-12-05 23:35:20 -08:00
Charles Reiss
813ac71459
Don't use bogus port number in notifyADeadHost().
2012-12-05 23:35:20 -08:00
Tathagata Das
21a0852976
Refactored RDD checkpointing to minimize extra fields in RDD class.
2012-12-04 22:10:25 -08:00
Tathagata Das
a69a82be26
Added metadata cleaner to HttpBroadcast to clean up old broacast files.
2012-12-03 22:37:31 -08:00
Josh Rosen
cdaa0fad51
Use external addresses in standalone WebUI on EC2.
2012-12-01 18:19:13 -08:00
Tathagata Das
b4dba55f78
Made RDD checkpoint not create a new thread. Fixed bug in detecting when spark.cleaner.delay is insufficient.
2012-12-02 02:03:05 +00:00
Tathagata Das
477de94894
Minor modifications.
2012-12-01 13:15:06 -08:00
Tathagata Das
6fcd09f499
Added TimeStampedHashSet and used that to cleanup the list of registered RDD IDs in CacheTracker.
2012-11-29 02:06:33 -08:00
Tathagata Das
c9789751bf
Added metadata cleaner to BlockManager to remove old blocks completely.
2012-11-28 23:18:24 -08:00
Tathagata Das
9e9e9e1d89
Renamed CleanupTask to MetadataCleaner.
2012-11-28 18:48:14 -08:00
Tathagata Das
e463ae4920
Modified StorageLevel and BlockManagerId to cache common objects and use cached object while deserializing.
2012-11-28 14:05:01 -08:00
Tathagata Das
d5e7aad039
Bug fixes
2012-11-28 08:36:55 +00:00
Matei Zaharia
f86960cba9
Merge pull request #313 from rxin/pde_size_compress
...
Added a partition preserving flag to MapPartitionsWithSplitRDD.
2012-11-27 22:39:25 -08:00
Matei Zaharia
3ebd8e1885
Added zip to Java API
2012-11-27 22:38:09 -08:00
Matei Zaharia
27e43abd19
Added a zip() operation for RDDs with the same shape (number of
...
partitions and number of elements in each partition)
2012-11-27 22:27:47 -08:00
Matei Zaharia
f410a111ad
Merge branch 'master' of github.com:mesos/spark
2012-11-27 20:51:58 -08:00
Josh Rosen
7d71b9a56a
Fix NullPointerException caused by unregistered map outputs.
2012-11-27 20:51:51 -08:00
Matei Zaharia
935c468b71
Merge pull request #311 from woggling/map-output-npe
...
Fix NullPointerException when map output unregistered from MapOutputTracker twice
2012-11-27 20:50:48 -08:00
Reynold Xin
bd6dd1a3a6
Added a partition preserving flag to MapPartitionsWithSplitRDD.
2012-11-27 19:43:30 -08:00
Reynold Xin
f24bfd2dd1
For size compression, compress non zero values into non zero values.
2012-11-27 19:20:45 -08:00
Charles Reiss
cf79de425d
Fix NullPointerException when unregistering a map output twice.
2012-11-27 16:12:05 -08:00
Tathagata Das
b18d70870a
Modified bunch HashMaps in Spark to use TimeStampedHashMap and made various modules use CleanupTask to periodically clean up metadata.
2012-11-27 15:08:49 -08:00
Tathagata Das
0fe2fc4d5e
Merged branch mesos/master to branch dev.
2012-11-26 13:16:59 -08:00
Tathagata Das
c97ebf6437
Fixed bug in the number of splits in RDD after checkpointing. Modified reduceByKeyAndWindow (naive) computation from window+reduceByKey to reduceByKey+window+reduceByKey.
2012-11-19 23:22:07 +00:00
Matei Zaharia
3ff6f4bdee
Merge pull request #304 from mbautin/configurable_local_ip
...
SPARK-624: make the default local IP customizable
2012-11-19 13:23:39 -08:00
mbautin
00f4e3ff9c
Addressing Matei's comment: SPARK_LOCAL_IP environment variable
2012-11-19 11:52:10 -08:00
Tathagata Das
10c1abcb6a
Fixed checkpointing bug in CoGroupedRDD. CoGroupSplits kept around the RDD splits of its parent RDDs, thus checkpointing its parents did not release the references to the parent splits.
2012-11-17 17:27:00 -08:00
Charles Reiss
12c24e786c
Set default uncaught exception handler to exit.
...
Among other things, should prevent OutOfMemoryErrors in some daemon threads
(such as the network manager) from causing a spark executor to enter a state
where it cannot make progress but does not report an error.
2012-11-16 20:12:31 -08:00
mbautin
1f5a7e0e64
SPARK-624: make the default local IP customizable
2012-11-15 13:57:47 -08:00
Matei Zaharia
c23a74df0a
Use DNS names instead of IP addresses in standalone mode, to allow
...
matching with data locality hints from storage systems.
2012-11-15 00:10:52 -08:00
Tathagata Das
8a25d530ed
Optimized checkpoint writing by reusing FileSystem object. Fixed bug in updating of checkpoint data in DStream where the checkpointed RDDs, upon recovery, were not recognized as checkpointed RDDs and therefore deleted from HDFS. Made InputStreamsSuite more robust to timing delays.
2012-11-13 02:16:28 -08:00
Matei Zaharia
173e0354c0
Detect correctly when one has disconnected from a standalone cluster.
...
SPARK-617 #resolve
2012-11-11 21:06:57 -08:00