Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/7021#discussion_r33971391
--- Diff: core/src/main/scala/org/apache/spark/rdd/RDDCheckpointData.scala
---
@@ -44,6 +46,26 @@ private[spark] class RDDCheckpointData[T:
ClassTag](@transient rdd: RDD[T])
import CheckpointState._
+ // Because SparkContext is transient in RDD, so we can't get the id and
checkpointDir later.
+ // So keep a copy of the id and checkpointDir.
+ // The id of RDD
+ val rddId: Int = rdd.id
+
+ // The path the checkpoint data will write to.
+ val checkpointDir = rdd.context.checkpointDir
+ @transient var checkpointPath: Path = null
+ @transient var fs: FileSystem = null
+ if (checkpointDir.isDefined) {
+ checkpointPath = new Path(checkpointDir.get, "rdd-" + rddId)
+ fs = checkpointPath.getFileSystem(rdd.context.hadoopConfiguration)
+ if (!fs.mkdirs(checkpointPath)) {
+ throw new SparkException("Failed to create checkpoint path " +
checkpointPath)
+ }
+ }
--- End diff --
these don't need to be vars right? In fact, `fs`, `rddId` and
`checkpointDir` don't even need to exist. You can just do
```
@transient val checkpointPath: Path = {
val path = RDDCheckpointData.rddCheckpointDataPath(rdd.context, rdd.id)
val fs = path.getFileSystem(rdd.context.hadoopConfiguration)
if (!fs.mkdirs(path)) {
...
}
path
}
```
---
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
