NetSys spark-monotasks issues

NetSys / spark-monotasks

Fast, predictable data analytics based on (and API-compatible with) Apache Spark

Apache License 2.0

25 stars 18 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

NoSuchFileException when HdfsDiskMonotask tries to access local disk

#34 hzhua opened 6 years ago
0
NetworkUtilization is not correctly decoded from JSON

#33 kayousterhout opened 8 years ago
0
Add off-heap memory use to the UI

#32 kayousterhout opened 8 years ago
0
DiskScheduler sometimes hangs on shutdown

#31 kayousterhout opened 9 years ago
0
Fix Java.util.concurrentModificationException in Hadoop code path

#30 kayousterhout opened 9 years ago
2
Clean up temporary data when a monotask fails.

#29 ccanel opened 9 years ago
0
Write unit test(s) for new code to read/write all shuffle data in a single file

#28 kayousterhout opened 9 years ago
0
Handle scenario where the entire input file isn't needed

#27 kayousterhout opened 9 years ago
0
Bug in BlockManager's concurrency control

#26 kayousterhout opened 9 years ago
2
Uncaught exceptions in DiskAccessorThreads lead to hang

#25 kayousterhout closed 9 years ago
3
Tell LocalDagScheduler when DiskMonotasks fail

#24 kayousterhout closed 9 years ago
1
Fix caching of ManagedBuffers in the BlockManager

#23 kayousterhout closed 9 years ago
1
Wait to serialize task metrics until all monotasks for a macrotask have completed

#22 kayousterhout closed 9 years ago
2
Remove all support for external shuffle and associated dynamic executor pool sizing

#21 kayousterhout opened 9 years ago
0
"Too many open files" error in CpuCounters

#20 ccanel closed 9 years ago
0
All types of ExecutorBackend should implement updateFreeCores

#19 kayousterhout opened 9 years ago
0
JsonProtcol.scala should properly decode BlockDeviceUtilizations

#18 kayousterhout opened 9 years ago
0
Throw exception in HadoopRDD.compute()

#17 kayousterhout opened 9 years ago
4
Scalastyle shouldn't fail when files only have the Berkeley header

#16 kayousterhout closed 9 years ago
2
Create a basic scheduler at the driver, that accounts for monotasks

#15 kayousterhout closed 9 years ago
0
Implement a control mechanism to prevent two macrotasks from loading or computing the same block at the same time.

#14 ccanel opened 9 years ago
0
Reimplement the BlockManager's ability to drop blocks from memory to disk if the MemoryStore becomes full.

#13 ccanel opened 9 years ago
0
Remove the BlockManager's ability to fetch remote blocks without using a NetworkMonotask.

#12 ccanel opened 9 years ago
0
NetworkMonotasks should schedule a DiskReadMonotask if a block that they are fetching is on disk.

#11 ccanel closed 9 years ago
0
Fail all dependent monotasks when a monotask fails

#10 kayousterhout opened 9 years ago
1
Support checkpoint RDDs

#9 kayousterhout opened 9 years ago
0
Delete intermediate data for monotasks that was stored in the memory store

#8 kayousterhout closed 9 years ago
0
Read Broadcast variables using a Network Monotask

#7 kayousterhout opened 9 years ago
0
Add back support to periodically update the master about metrics for running tasks

#6 kayousterhout opened 9 years ago
0
Change LocalDagScheduler to use an event loop

#5 kayousterhout closed 9 years ago
0
Support killing tasks

#4 kayousterhout opened 9 years ago
0
LocalDagScheduler.runningMonotasks Race Condition

#3 ccanel closed 9 years ago
2
Flushing to disk throws a ClassCastException when value is a single value (not an array or serialized buffer)

#2 kayousterhout opened 9 years ago
1
+ adding beginnings of a ReadTask. It includes:

#1 ryanra closed 7 years ago
11