Github user tgravescs commented on a diff in the pull request:
https://github.com/apache/spark/pull/4688#discussion_r25527152
--- Diff:
yarn/src/main/scala/org/apache/spark/deploy/yarn/YarnSparkHadoopUtil.scala ---
@@ -82,6 +93,102 @@ class YarnSparkHadoopUtil extends SparkHadoopUtil {
if (credentials != null) credentials.getSecretKey(new Text(key)) else
null
}
+ override def setPrincipalAndKeytabForLogin(principal: String, keytab:
String): Unit = {
+ loginPrincipal = Option(principal)
+ keytabFile = Option(keytab)
+ }
+
+ private[spark] override def scheduleLoginFromKeytab(
+ callback: (SerializableBuffer) => Unit): Unit = {
+
+ loginPrincipal match {
+ case Some(principal) =>
+ val keytab = keytabFile.get
+ val remoteFs = FileSystem.get(conf)
+ val remoteKeytabPath = new Path(
+ remoteFs.getHomeDirectory, System.getenv("SPARK_STAGING_DIR") +
Path.SEPARATOR + keytab)
+ val localFS = FileSystem.getLocal(conf)
+ // At this point, SparkEnv is likely no initialized, so create a
dir, put the keytab there.
+ val tempDir = Utils.createTempDir()
+ val localURI = new URI(tempDir.getAbsolutePath + Path.SEPARATOR +
keytab)
+ val qualifiedURI = new URI(localFS.makeQualified(new
Path(localURI)).toString)
+ FileUtil.copy(
--- End diff --
why not just use the distributed cache for this? I'm more hesitant about
us manually doing it and messing up permissions and such then allow YARN to do.
You can set it up so it only downloads to AM and not executors.
---
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]
