Github user erenavsarogullari commented on a diff in the pull request:
https://github.com/apache/spark/pull/16813#discussion_r99965179
--- 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 --
Yep, it might be useful. I was thinking about the case: if user has two
`fairscheduler.xml` files (one of them is default file in classpath and second
one is in another path). I will also address this with the other comments,
thanks again ;)
---
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]
