Return-Path: X-Original-To: apmail-hadoop-common-issues-archive@minotaur.apache.org Delivered-To: apmail-hadoop-common-issues-archive@minotaur.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 1AA7F9FBE for ; Mon, 21 May 2012 19:59:42 +0000 (UTC) Received: (qmail 48414 invoked by uid 500); 21 May 2012 19:59:41 -0000 Delivered-To: apmail-hadoop-common-issues-archive@hadoop.apache.org Received: (qmail 48372 invoked by uid 500); 21 May 2012 19:59:41 -0000 Mailing-List: contact common-issues-help@hadoop.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: common-issues@hadoop.apache.org Delivered-To: mailing list common-issues@hadoop.apache.org Received: (qmail 48260 invoked by uid 99); 21 May 2012 19:59:41 -0000 Received: from issues-vm.apache.org (HELO issues-vm) (140.211.11.160) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 21 May 2012 19:59:41 +0000 Received: from isssues-vm.apache.org (localhost [127.0.0.1]) by issues-vm (Postfix) with ESMTP id 829B9141887 for ; Mon, 21 May 2012 19:59:41 +0000 (UTC) Date: Mon, 21 May 2012 19:59:41 +0000 (UTC) From: "Daryn Sharp (JIRA)" To: common-issues@hadoop.apache.org Message-ID: <1611247932.4633.1337630381537.JavaMail.jiratomcat@issues-vm> In-Reply-To: <2015336447.12105.1337304247036.JavaMail.tomcat@hel.zones.apache.org> Subject: [jira] [Commented] (HADOOP-8408) MR doesn't work with a non-default ViewFS mount table and security enabled MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/HADOOP-8408?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13280425#comment-13280425 ] Daryn Sharp commented on HADOOP-8408: ------------------------------------- Yes, I noticed awhile back that HDFS-3062 was problematic, so it was superseded by my HDFS-3268. The handling of tokens for multi-token filesystem is rather funky right now. HADOOP-7967 and MAPREDUCE-3825 aim to fix the problem. The way it's intended to work is a filesystem with no tokens is supposed to return null for its service name. The {{TokenCache}} used to skip a fs with a null service. This was problematic for multi-token filesystems like {{ViewFileSystem}}, so someone changed the {{TokenCache}} to proceed to acquire tokens from a fs with a null service. A better approach is to request tokens from a filesystem: a "normal" fs returns its token if its service is non-null, and a multi-token filesystem gets the tokens for the leaf fs-es that don't have a null service. I hope that makes some sense... > MR doesn't work with a non-default ViewFS mount table and security enabled > -------------------------------------------------------------------------- > > Key: HADOOP-8408 > URL: https://issues.apache.org/jira/browse/HADOOP-8408 > Project: Hadoop Common > Issue Type: Bug > Components: fs > Affects Versions: 2.0.0 > Reporter: Aaron T. Myers > Assignee: Aaron T. Myers > Fix For: 2.0.1 > > Attachments: HDFS-8408.patch > > > With security enabled, if one sets up a ViewFS mount table using the default mount table name, everything works as expected. However, if you try to create a ViewFS mount table with a non-default name, you'll end up getting an error like the following (in this case "vfs-cluster" was the name of the mount table) when running an MR job: > {noformat} > java.lang.IllegalArgumentException: java.net.UnknownHostException: vfs-cluster > {noformat} -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira