SPARK-1557 Set permissions on event log files/directories

Author: tgravescs

core/src/main/scala/org/apache/spark/scheduler/EventLoggingListener.scala

@@ -51,10 +51,11 @@ private[spark] class EventLoggingListener(
   private val logBaseDir = conf.get("spark.eventLog.dir", "/tmp/spark-events").stripSuffix("/")
   private val name = appName.replaceAll("[ :/]", "-").toLowerCase + "-" + System.currentTimeMillis
   val logDir = logBaseDir + "/" + name
+  val LOG_FILE_PERMISSIONS: FsPermission = FsPermission.createImmutable(0770: Short)
 
   private val logger =
     new FileLogger(logDir, conf, hadoopConfiguration, outputBufferSize, shouldCompress,
-      shouldOverwrite)
+      shouldOverwrite, Some(LOG_FILE_PERMISSIONS))
 
   /**
    * Begin logging events.
@@ -64,10 +65,11 @@ private[spark] class EventLoggingListener(
     logInfo("Logging events to %s".format(logDir))
     if (shouldCompress) {
       val codec = conf.get("spark.io.compression.codec", CompressionCodec.DEFAULT_COMPRESSION_CODEC)
-      logger.newFile(COMPRESSION_CODEC_PREFIX + codec)
+      logger.newFile(COMPRESSION_CODEC_PREFIX + codec, Some(LOG_FILE_PERMISSIONS))
     }
-    logger.newFile(SPARK_VERSION_PREFIX + SparkContext.SPARK_VERSION)
-    logger.newFile(LOG_PREFIX + logger.fileIndex)
+    logger.newFile(SPARK_VERSION_PREFIX + SparkContext.SPARK_VERSION,
+      Some(LOG_FILE_PERMISSIONS))
+    logger.newFile(LOG_PREFIX + logger.fileIndex, Some(LOG_FILE_PERMISSIONS))
   }
 
   /** Log the event as JSON. */
@@ -114,7 +116,7 @@ private[spark] class EventLoggingListener(
    * In addition, create an empty special file to indicate application completion.
    */
   def stop() = {
-    logger.newFile(APPLICATION_COMPLETE)
+    logger.newFile(APPLICATION_COMPLETE, Some(LOG_FILE_PERMISSIONS))
     logger.stop()
   }
 }

core/src/main/scala/org/apache/spark/util/FileLogger.scala

@@ -27,6 +27,7 @@ import org.apache.hadoop.fs.{FileSystem, FSDataOutputStream, Path}
 
 import org.apache.spark.{Logging, SparkConf}
 import org.apache.spark.io.CompressionCodec
+import org.apache.hadoop.fs.permission.FsPermission
 
 /**
  * A generic class for logging information to file.
@@ -42,7 +43,8 @@ private[spark] class FileLogger(
     hadoopConfiguration: Configuration,
     outputBufferSize: Int = 8 * 1024, // 8 KB
     compress: Boolean = false,
-    overwrite: Boolean = true)
+    overwrite: Boolean = true,
+    dirPermissions: Option[FsPermission] = None)
   extends Logging {
 
   private val dateFormat = new ThreadLocal[SimpleDateFormat]() {
@@ -60,12 +62,12 @@ private[spark] class FileLogger(
 
   private var writer: Option[PrintWriter] = None
 
-  createLogDir()
+  createLogDir(dirPermissions)
 
   /**
    * Create a logging directory with the given path.
    */
-  private def createLogDir() {
+  private def createLogDir(dirPerms: Option[FsPermission]) {
     val path = new Path(logDir)
     if (fileSystem.exists(path)) {
       if (overwrite) {
@@ -79,16 +81,23 @@ private[spark] class FileLogger(
     if (!fileSystem.mkdirs(path)) {
       throw new IOException("Error in creating log directory: %s".format(logDir))
     }
+    if (dirPerms.isDefined) {
+      val fsStatus = fileSystem.getFileStatus(path)
+      if (fsStatus.getPermission().toShort() != dirPerms.get.toShort()) {
+        fileSystem.setPermission(path, dirPerms.get);
+      }
+    }
   }
 
   /**
    * Create a new writer for the file identified by the given path.
    */
-  private def createWriter(fileName: String): PrintWriter = {
+  private def createWriter(fileName: String, perms: Option[FsPermission] = None): PrintWriter = {
     val logPath = logDir + "/" + fileName
     val uri = new URI(logPath)
     val defaultFs = FileSystem.getDefaultUri(hadoopConfiguration).getScheme
     val isDefaultLocal = (defaultFs == null || defaultFs == "file")
+    val path = new Path(logPath)
 
     /* The Hadoop LocalFileSystem (r1.0.4) has known issues with syncing (HADOOP-7844).
      * Therefore, for local files, use FileOutputStream instead. */
@@ -97,11 +106,11 @@ private[spark] class FileLogger(
         // Second parameter is whether to append
         new FileOutputStream(uri.getPath, !overwrite)
       } else {
-        val path = new Path(logPath)
         hadoopDataStream = Some(fileSystem.create(path, overwrite))
         hadoopDataStream.get
       }
 
+    if (perms.isDefined) fileSystem.setPermission(path, perms.get)
     val bstream = new BufferedOutputStream(dstream, outputBufferSize)
     val cstream = if (compress) compressionCodec.compressedOutputStream(bstream) else bstream
     new PrintWriter(cstream)
@@ -150,15 +159,16 @@ private[spark] class FileLogger(
   /**
    * Start a writer for a new file, closing the existing one if it exists.
    * @param fileName Name of the new file, defaulting to the file index if not provided.
+   * @param perms Permissions to put on the new file.
    */
-  def newFile(fileName: String = "") {
+  def newFile(fileName: String = "", perms: Option[FsPermission] = None) {
     fileIndex += 1
     writer.foreach(_.close())
     val name = fileName match {
       case "" => fileIndex.toString
       case _ => fileName
     }
-    writer = Some(createWriter(name))
+    writer = Some(createWriter(name, perms))
   }
 
   /**

docs/security.md

@@ -7,6 +7,8 @@ Spark currently supports authentication via a shared secret. Authentication can
 
 The Spark UI can also be secured by using javax servlet filters. A user may want to secure the UI if it has data that other users should not be allowed to see. The javax servlet filter specified by the user can authenticate the user and then once the user is logged in, Spark can compare that user versus the view acls to make sure they are authorized to view the UI. The configs 'spark.ui.acls.enable' and 'spark.ui.view.acls' control the behavior of the acls. Note that the person who started the application always has view access to the UI.
 
+If your applications are using event logging, the directory where the event logs go (`spark.eventLog.dir`) should be manually created and have the proper permissions set on it. If you want those log files secure, the permissions should be set to drwxrwxrwxt for that directory. The owner of the directory should be the super user who is running the history server and the group permissions should be restricted to super user group. This will allow all users to write to the directory but will prevent unprivileged users from removing or renaming a file unless they own the file or directory. The event log files will be created by Spark with permissions such that only the user and group have read and write access.
+
 For Spark on Yarn deployments, configuring `spark.authenticate` to true will automatically handle generating and distributing the shared secret. Each application will use a unique shared secret. The Spark UI uses the standard YARN web application proxy mechanism and will authenticate via any installed Hadoop filters. If an authentication filter is enabled, the acls controls can be used by control which users can via the Spark UI. 
 
 For other types of Spark deployments, the spark config `spark.authenticate.secret` should be configured on each of the nodes. This secret will be used by all the Master/Workers and applications. The UI can be secured using a javax servlet filter installed via `spark.ui.filters`. If an authentication filter is enabled, the acls controls can be used by control which users can via the Spark UI.