HADOOP-17060. Clarify listStatus and getFileStatus behaviors inconsistent in the case of ViewFs implementation for isDirectory. Contributed by Uma Maheswara Rao G.

(cherry picked from commit 93b121a971)

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/viewfs/ViewFileSystem.java

@@ -488,6 +488,14 @@ private static FileStatus wrapLocalFileStatus(FileStatus orig,
         : new ViewFsFileStatus(orig, qualified);
   }
 
+  /**
+   * {@inheritDoc}
+   *
+   * If the given path is a symlink(mount link), the path will be resolved to a
+   * target path and it will get the resolved path's FileStatus object. It will
+   * not be represented as a symlink and isDirectory API returns true if the
+   * resolved path is a directory, false otherwise.
+   */
   @Override
   public FileStatus getFileStatus(final Path f) throws AccessControlException,
       FileNotFoundException, IOException {
@@ -505,6 +513,25 @@ public void access(Path path, FsAction mode) throws AccessControlException,
     res.targetFileSystem.access(res.remainingPath, mode);
   }
 
+  /**
+   * {@inheritDoc}
+   *
+   * Note: listStatus on root("/") considers listing from fallbackLink if
+   * available. If the same directory name is present in configured mount path
+   * as well as in fallback link, then only the configured mount path will be
+   * listed in the returned result.
+   *
+   * If any of the the immediate children of the given path f is a symlink(mount
+   * link), the returned FileStatus object of that children would be represented
+   * as a symlink. It will not be resolved to the target path and will not get
+   * the target path FileStatus object. The target path will be available via
+   * getSymlink on that children's FileStatus object. Since it represents as
+   * symlink, isDirectory on that children's FileStatus will return false.
+   *
+   * If you want to get the FileStatus of target path for that children, you may
+   * want to use GetFileStatus API with that children's symlink path. Please see
+   * {@link ViewFileSystem#getFileStatus(Path f)}
+   */
   @Override
   public FileStatus[] listStatus(final Path f) throws AccessControlException,
       FileNotFoundException, IOException {
@@ -1174,20 +1201,11 @@ public FileStatus getFileStatus(Path f) throws IOException {
       checkPathIsSlash(f);
       return new FileStatus(0, true, 0, 0, creationTime, creationTime,
           PERMISSION_555, ugi.getShortUserName(), ugi.getPrimaryGroupName(),
-
           new Path(theInternalDir.fullPath).makeQualified(
               myUri, ROOT_PATH));
     }
     
 
-    /**
-     * {@inheritDoc}
-     *
-     * Note: listStatus on root("/") considers listing from fallbackLink if
-     * available. If the same directory name is present in configured mount
-     * path as well as in fallback link, then only the configured mount path
-     * will be listed in the returned result.
-     */
     @Override
     public FileStatus[] listStatus(Path f) throws AccessControlException,
         FileNotFoundException, IOException {

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/viewfs/ViewFs.java

@@ -351,6 +351,14 @@ public FileChecksum getFileChecksum(final Path f)
     return res.targetFileSystem.getFileChecksum(res.remainingPath);
   }
 
+  /**
+   * {@inheritDoc}
+   *
+   * If the given path is a symlink(mount link), the path will be resolved to a
+   * target path and it will get the resolved path's FileStatus object. It will
+   * not be represented as a symlink and isDirectory API returns true if the
+   * resolved path is a directory, false otherwise.
+   */
   @Override
   public FileStatus getFileStatus(final Path f) throws AccessControlException,
       FileNotFoundException, UnresolvedLinkException, IOException {
@@ -436,6 +444,22 @@ public LocatedFileStatus getViewFsFileStatus(LocatedFileStatus stat,
     };
   }
   
+  /**
+   * {@inheritDoc}
+   *
+   * If any of the the immediate children of the given path f is a symlink(mount
+   * link), the returned FileStatus object of that children would be represented
+   * as a symlink. It will not be resolved to the target path and will not get
+   * the target path FileStatus object. The target path will be available via
+   * getSymlink on that children's FileStatus object. Since it represents as
+   * symlink, isDirectory on that children's FileStatus will return false.
+   *
+   * If you want to get the FileStatus of target path for that children, you may
+   * want to use GetFileStatus API with that children's symlink path. Please see
+   * {@link ViewFs#getFileStatus(Path f)}
+   *
+   * Note: In ViewFs, the mount links are represented as symlinks.
+   */
   @Override
   public FileStatus[] listStatus(final Path f) throws AccessControlException,
       FileNotFoundException, UnresolvedLinkException, IOException {

hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/fs/Hdfs.java

@@ -135,6 +135,14 @@ public FileChecksum getFileChecksum(Path f)
     return dfs.getFileChecksumWithCombineMode(getUriPath(f), Long.MAX_VALUE);
   }
 
+  /**
+   * {@inheritDoc}
+   *
+   * If the given path is a symlink, the path will be resolved to a target path
+   * and it will get the resolved path's FileStatus object. It will not be
+   * represented as a symlink and isDirectory API returns true if the resolved
+   * path is a directory, false otherwise.
+   */
   @Override
   public FileStatus getFileStatus(Path f) 
       throws IOException, UnresolvedLinkException {
@@ -269,6 +277,20 @@ public HdfsFileStatus getNext() throws IOException {
     }
   }
 
+  /**
+   * {@inheritDoc}
+   *
+   * If any of the the immediate children of the given path f is a symlink, the
+   * returned FileStatus object of that children would be represented as a
+   * symlink. It will not be resolved to the target path and will not get the
+   * target path FileStatus object. The target path will be available via
+   * getSymlink on that children's FileStatus object. Since it represents as
+   * symlink, isDirectory on that children's FileStatus will return false.
+   *
+   * If you want to get the FileStatus of target path for that children, you may
+   * want to use GetFileStatus API with that children's symlink path. Please see
+   * {@link Hdfs#getFileStatus(Path f)}
+   */
   @Override
   public FileStatus[] listStatus(Path f) 
       throws IOException, UnresolvedLinkException {

hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DistributedFileSystem.java

@@ -1143,10 +1143,21 @@ private FileStatus[] listStatusInternal(Path p) throws IOException {
   /**
    * List all the entries of a directory
    *
-   * Note that this operation is not atomic for a large directory.
+   * Note that this operation is not atomic for a large directory. The entries
-   * The entries of a directory may be fetched from NameNode multiple times.
+   * of a directory may be fetched from NameNode multiple times. It only
-   * It only guarantees that  each name occurs once if a directory
+   * guarantees that each name occurs once if a directory undergoes changes
-   * undergoes changes between the calls.
+   * between the calls.
+   *
+   * If any of the the immediate children of the given path f is a symlink, the
+   * returned FileStatus object of that children would be represented as a
+   * symlink. It will not be resolved to the target path and will not get the
+   * target path FileStatus object. The target path will be available via
+   * getSymlink on that children's FileStatus object. Since it represents as
+   * symlink, isDirectory on that children's FileStatus will return false.
+   *
+   * If you want to get the FileStatus of target path for that children, you may
+   * want to use GetFileStatus API with that children's symlink path. Please see
+   * {@link DistributedFileSystem#getFileStatus(Path f)}
    */
   @Override
   public FileStatus[] listStatus(Path p) throws IOException {
@@ -1712,6 +1723,12 @@ public FsServerDefaults getServerDefaults() throws IOException {
 
   /**
    * Returns the stat information about the file.
+   *
+   * If the given path is a symlink, the path will be resolved to a target path
+   * and it will get the resolved path's FileStatus object. It will not be
+   * represented as a symlink and isDirectory API returns true if the resolved
+   * path is a directory, false otherwise.
+   *
    * @throws FileNotFoundException if the file does not exist.
    */
   @Override