Github user markhamstra commented on a diff in the pull request:
https://github.com/apache/spark/pull/16813#discussion_r99964246
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/SchedulableBuilder.scala ---
@@ -69,60 +72,81 @@ private[spark] class FairSchedulableBuilder(val
rootPool: Pool, conf: SparkConf)
val DEFAULT_WEIGHT = 1
override def buildPools() {
- var is: Option[InputStream] = None
+ var fileData: Option[FileData] = None
try {
- is = Option {
- schedulerAllocFile.map { f =>
- new FileInputStream(f)
- }.getOrElse {
-
Utils.getSparkClassLoader.getResourceAsStream(DEFAULT_SCHEDULER_FILE)
- }
- }
-
- is.foreach { i => buildFairSchedulerPool(i) }
+ fileData = getFileData()
+ fileData.foreach { data => buildFairSchedulerPool(data) }
+ } catch {
+ case NonFatal(t) =>
+ logError("Error while building the fair scheduler pools: ", t)
+ throw t
} finally {
- is.foreach(_.close())
+ fileData.foreach(_.inputStream.close())
}
// finally create "default" pool
buildDefaultPool()
}
+ private def getFileData(): Option[FileData] = {
+ schedulerAllocFile.map { f =>
+ val file = new File(f)
+ val fis = new FileInputStream(file)
--- End diff --
Yes, and if they do have multiple files with the same name on different
paths and are expecting one of them to be used when Spark is actually trying to
use one on a different path, then having the full path in the log message will
be crucial to short-circuiting the debugging confusion.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
