Github user pwendell commented on a diff in the pull request:
https://github.com/apache/spark/pull/2087#discussion_r18939658
--- Diff: core/src/main/scala/org/apache/spark/deploy/SparkHadoopUtil.scala
---
@@ -121,6 +125,31 @@ class SparkHadoopUtil extends Logging {
UserGroupInformation.loginUserFromKeytab(principalName, keytabFilename)
}
+ /**
+ * Returns a function that can be called to find the number of Hadoop
FileSystem bytes read by
+ * this thread so far. Reflection is required because thread-level
FileSystem statistics are only
+ * available as of Hadoop 2.5 (see HADOOP-10688). Returns None if the
required method can't be
+ * found.
+ */
+ def getInputBytesReadCallback(path: Path, conf: Configuration):
Option[() => Long] = {
+ val qualifiedPath = path.getFileSystem(conf).makeQualified(path)
+ val scheme = qualifiedPath.toUri().getScheme()
+ val stats =
FileSystem.getAllStatistics().filter(_.getScheme().equals(scheme))
+ try {
+ val threadStats = stats.map(Utils.invoke(classOf[Statistics], _,
"getThreadStatistics"))
+ val statisticsDataClass =
+
Class.forName("org.apache.hadoop.fs.FileSystem$Statistics$StatisticsData")
+ val getBytesReadMethod =
statisticsDataClass.getDeclaredMethod("getBytesRead")
+ val f = () =>
threadStats.map(getBytesReadMethod.invoke(_).asInstanceOf[Long]).sum
+ val start = f()
+ Some(() => f() - start)
--- End diff --
ah I see now - so should we call this function `getThreadLocalBytesRead` or
something? It seems like it only semantically makes sense if the returned
function is called from the same thread as the function was generated in. That
might also be worth documenting in the javadoc. The use of the phrase "so far"
there also threw me off a bit - maybe there is a better phrase for that.
---
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
