Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/9428#discussion_r44320866
--- Diff:
core/src/main/scala/org/apache/spark/rdd/ReliableRDDCheckpointData.scala ---
@@ -63,11 +98,20 @@ private[spark] class ReliableRDDCheckpointData[T:
ClassTag](@transient private v
throw new SparkException(s"Failed to create checkpoint path $cpDir")
}
- // Save to file, and reload it as an RDD
- val broadcastedConf = rdd.context.broadcast(
- new SerializableConfiguration(rdd.context.hadoopConfiguration))
- // TODO: This is expensive because it computes the RDD again
unnecessarily (SPARK-8582)
- rdd.context.runJob(rdd,
ReliableCheckpointRDD.writeCheckpointFile[T](cpDir, broadcastedConf) _)
+ val checkpointedPartitionFiles =
fs.listStatus(path).map(_.getPath.getName).toSet
+ // Not all actions compute all partitions of the RDD (e.g. take). For
correctness, we
+ // must checkpoint any missing partitions. TODO: avoid running another
job here (SPARK-8582).
+ val missingPartitionIndices = rdd.partitions.map(_.index).filter { i =>
+
!checkpointedPartitionFiles(ReliableCheckpointRDD.checkpointFileName(i))
+ }
+ if (missingPartitionIndices.nonEmpty) {
+ // TODO: This is expensive because it computes the RDD again
unnecessarily (SPARK-8582)
--- End diff --
this TODO is outdated. This is only called when the iterator is not
drained, in which case it is necessary to run another job to compute the
missing partitions for correctness. I would just remove this TODO.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
