You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Vikram Dixit K (JIRA)" <ji...@apache.org> on 2015/11/19 21:01:10 UTC

[jira] [Commented] (HIVE-12459) Tez startup dislikes spaces in classpath on Windows

    [ https://issues.apache.org/jira/browse/HIVE-12459?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15014254#comment-15014254 ] 

Vikram Dixit K commented on HIVE-12459:
---------------------------------------

Hive derives the location of the hive-exec jar from its classpath much like MR. Not sure why this works in MR and not in Tez. Will investigate and get back to you on this.

> Tez startup dislikes spaces in classpath on Windows
> ---------------------------------------------------
>
>                 Key: HIVE-12459
>                 URL: https://issues.apache.org/jira/browse/HIVE-12459
>             Project: Hive
>          Issue Type: Bug
>          Components: Tez
>    Affects Versions: 1.2.1
>            Reporter: john lilley
>            Assignee: Vikram Dixit K
>
> We are seeing an issue that I summarize as “Tez doesn’t like spaces in the classpath”, but I wanted to check with the group before submitting a JIRA.  This is showing when we try to access Hive on HDP 2.3 from a Windows client, where we’ve put the client jars in a classpath that contains spaces.  
> The causing line in our code is:
> 	    state = SessionState.start(hiveConf);
> where SessionState is in org.apache.hadoop.hive.ql.session
> The exception stack is:
> net/redpoint/hiveclient/DMHCatClient.newInstance:java.lang.RuntimeException: java.io.FileNotFoundException: File file:/C:/Program%20Files/RedPointDM7/hadoop/clusters/hds-cent6/lib/hive-exec-1.2.1.2.3.0.0-2557.jar does not exist
> Additional message: 
>     org.apache.hadoop.hive.ql.session.SessionState.start(SessionState.java:535)
>     org.apache.hadoop.hive.ql.session.SessionState.start(SessionState.java:466)
>     net.redpoint.hiveclient.DMHCatClient.<init>(DMHCatClient.java:255)
>     net.redpoint.hiveclient.DMHCatClient.newInstance(DMHCatClient.java:59)
> Caused by: java.io.FileNotFoundException: File file:/C:/Program%20Files/RedPointDM7/hadoop/clusters/hds-cent6/lib/hive-exec-1.2.1.2.3.0.0-2557.jar does not exist
>     org.apache.hadoop.fs.RawLocalFileSystem.deprecatedGetFileStatus(RawLocalFileSystem.java:606)
>     org.apache.hadoop.fs.RawLocalFileSystem.getFileLinkStatusInternal(RawLocalFileSystem.java:819)
>     org.apache.hadoop.fs.RawLocalFileSystem.getFileStatus(RawLocalFileSystem.java:596)
>     org.apache.hadoop.fs.FilterFileSystem.getFileStatus(FilterFileSystem.java:421)
>     org.apache.hadoop.fs.ChecksumFileSystem$ChecksumFSInputChecker.<init>(ChecksumFileSystem.java:140)
>     org.apache.hadoop.fs.ChecksumFileSystem.open(ChecksumFileSystem.java:341)
>     org.apache.hadoop.fs.FileSystem.open(FileSystem.java:767)
>     org.apache.hadoop.hive.ql.exec.tez.TezSessionState.getSha(TezSessionState.java:356)
>     org.apache.hadoop.hive.ql.exec.tez.TezSessionState.createJarLocalResource(TezSessionState.java:332)
>     org.apache.hadoop.hive.ql.exec.tez.TezSessionState.open(TezSessionState.java:151)
>     org.apache.hadoop.hive.ql.exec.tez.TezSessionState.open(TezSessionState.java:116)
>     org.apache.hadoop.hive.ql.session.SessionState.start(SessionState.java:532)
>     org.apache.hadoop.hive.ql.session.SessionState.start(SessionState.java:466)
>     net.redpoint.hiveclient.DMHCatClient.<init>(DMHCatClient.java:255)
>     net.redpoint.hiveclient.DMHCatClient.newInstance(DMHCatClient.java:59)
> It sure looks like something in the client code is turning “C:/Program Files” into “C:/Program%20Files”.  I am certain that it is not our code, because we otherwise access all of the jars and Java classes just fine.
> Furthermore, disabling Tez for client-side Hive query in the configuration seems to fix or avoid the issue:
> theConfiguration.set("hive.execution.engine", "mr");
> The stack trace doesn’t make sense to me, because we use FileSystem all over the place and it doesn’t run into this problem when accessing HDFS, only when connecting to Hive.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)