[ https://issues.apache.org/jira/browse/TEZ-3993?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16778599#comment-16778599 ]
Jonathan Eagles commented on TEZ-3993: -------------------------------------- [~daryn], you have mentioned there was some issues with this patch. Can you elaborated? > Tez fails to parse windows file paths in local mode > --------------------------------------------------- > > Key: TEZ-3993 > URL: https://issues.apache.org/jira/browse/TEZ-3993 > Project: Apache Tez > Issue Type: Bug > Reporter: Jonathan Eagles > Assignee: Jonathan Eagles > Priority: Major > Attachments: TEZ-3993.001.patch, TEZ-3993.002.patch > > > TezLocalCacheManager tries to generate symlinks to files that it puts in the > local cache, but the code that it uses to construct the path names is not > safe on Windows and causes bad file names to be constructed when run in a > Windows environment. On Windows, a path like file:/c:/path/to/my/file should > be legal and transform to c:\path\to\my\file, but with the invalid construct, > it turns into the illegal value /c:/path/to/my/file instead. > In TezLocalCacheManager, there is code that does > {code} > private boolean createSymlink(Path target, Path link) throws IOException { > LOG.info("Creating symlink: {} <- {}", target, link); > String targetPath = target.toUri().getPath(); > String linkPath = link.toUri().getPath(); > {code} > It looks like there are several other places in the Tez code that also use > the Path.toUri().getPath() construct that probably also need to be fixed in > order to work correctly on Windows. > The construct Path.toUri().getPath() doesn't handle windows directories > correctly. The Java File class understands how to do this correctly, so this > should really be replaced by > {code} > private boolean createSymlink(Path target, Path link) throws IOException { > LOG.info("Creating symlink: {} <- {}", target, link); > String targetPath = new File(target.toUri()).getCanonicalPath(); > String linkPath = new File(link.toUri()).getCanonicalPath(); > {code} > {code} > 2018-09-19T16:32:53,287 ERROR [LocalContainerLauncher-SubTaskRunner] > org.apache.tez.dag.app.launcher.LocalContainerLauncher - TezSubTaskRunner > failed due to exception > java.nio.file.InvalidPathException: Illegal char <:> at index 2: > /C:/Users/...fullpath > at sun.nio.fs.WindowsPathParser.normalize(WindowsPathParser.java:182) > at sun.nio.fs.WindowsPathParser.parse(WindowsPathParser.java:153) > at sun.nio.fs.WindowsPathParser.parse(WindowsPathParser.java:77) > at sun.nio.fs.WindowsPath.parse(WindowsPath.java:94) > at sun.nio.fs.WindowsFileSystem.getPath(WindowsFileSystem.java:255) > at java.nio.file.Paths.get(Paths.java:84) > at > org.apache.tez.dag.app.launcher.TezLocalCacheManager.createSymlink(TezLocalCacheManager.java:173) > at > org.apache.tez.dag.app.launcher.TezLocalCacheManager.localize(TezLocalCacheManager.java:126) > at > org.apache.tez.dag.app.launcher.LocalContainerLauncher.launch(LocalContainerLauncher.java:263) > at > org.apache.tez.dag.app.launcher.LocalContainerLauncher.access$300(LocalContainerLauncher.java:82) > at > org.apache.tez.dag.app.launcher.LocalContainerLauncher$TezSubTaskRunner.run(LocalContainerLauncher.java:207) > {code} -- This message was sent by Atlassian JIRA (v7.6.3#76005)