Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/6743#discussion_r33702512
--- Diff: core/src/main/scala/org/apache/spark/api/r/RUtils.scala ---
@@ -0,0 +1,55 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.api.r
+
+import java.io.File
+
+import org.apache.spark.SparkException
+
+private[spark] object RUtils {
+ /**
+ * Get the SparkR package path in the local spark distribution.
+ */
+ def localSparkRPackagePath: Option[String] = {
+ val sparkHome = sys.env.get("SPARK_HOME")
+ sparkHome.map(
+ Seq(_, "R", "lib").mkString(File.separator)
+ )
+ }
+
+ /**
+ * Get the SparkR package path in various deployment modes.
+ */
+ def sparkRPackagePath(driver: Boolean): String = {
+ val yarnMode = sys.env.get("SPARK_YARN_MODE")
+ if (!yarnMode.isEmpty && yarnMode.get == "true" &&
+ !(driver && System.getProperty("spark.master") == "yarn-client")) {
+ // For workers in YARN modes and driver in yarn cluster mode,
+ // the SparkR package distributed as an archive resource should be
pointed to
+ // by a symbol link "sparkr" in the current directory.
+ new File("sparkr").getAbsolutePath
--- End diff --
I think we should rewrite this method in the following way. Right now it
reads the deploy mode from the master, which happens to work for YARN but isn't
really the right way to do it. Instead, we should pass the deploy mode from
`SparkSubmit.scala` as `spark.submit.deployMode` or something. Then this method
would look like
```
/**
* Get the SparkR package path in various deployment modes.
* This assumes that system properties `spark.master` and
`spark.submit.deployMode`
* and environment variable `SPARK_HOME` are set.
*/
def sparkRPackagePath(driver: Boolean): String = {
val master = sys.props("spark.master")
val deployMode = sys.props("spark.submit.deployMode")
val isYarnCluster = master.contains("yarn") && deployMode == "cluster"
val isYarnClient = master.contains("yarn") && deployMode == "client"
// In YARN mode, the SparkR package is distributed as an archive
symbolically
// linked to the "sparkr" file in the current directory. Note that this
does not apply
// to the driver in client mode because it is run outside of the cluster.
if (isYarnCluster || (isYarnClient && !driver)) {
new File("sparkr").getAbsolutePath
} else {
// Otherwise, assume the package is local
// TODO: support this for Mesos
localSparkRPackagePath.getOrElse {
throw new SparkException("SPARK_HOME not set. Can't locate SparkR
package.")
}
}
}
```
---
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]
