hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "E. Sammer (JIRA)" <j...@apache.org>
Subject [jira] Commented: (MAPREDUCE-1378) Args in job details links on jobhistory.jsp are not URL encoded
Date Wed, 17 Feb 2010 16:21:27 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-1378?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12834820#action_12834820

E. Sammer commented on MAPREDUCE-1378:


Makes perfect sense, re: an improved web app. The testing of this is pretty straight forward.

Before the patch:

1. With a '+' or other non-encoded URL character in the real system path containing Hadoop,
go to http://jt:50030/jobhistory.jsp.
2. Click on one of the job ids to view job history details.
3.a. Get a 500 response due to a FileNotFound exception ex. java.io.FileNotFoundException:
File file:/Users/esammer/Applications/hadoop-0.20.1 152/logs/history/done/ (note the '+' in
the Hadoop dir is a space).

After the patch:

1.5. Observe the '+' is encoded as %2B in the job id link.
3.b. Instead of 3.a., observe that the page loads successfully.

This is the test case I followed. This patch is very specific in that it deals with this singular
link. The motivation behind doing a deep dive into the web app architecture is to address
other places where user input or OS file system paths are used to build links for similar
reasons. Let me know if this doesn't make sense.

> Args in job details links on jobhistory.jsp are not URL encoded
> ---------------------------------------------------------------
>                 Key: MAPREDUCE-1378
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1378
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: jobtracker
>    Affects Versions: 0.22.0
>            Reporter: E. Sammer
>            Priority: Trivial
>         Attachments: MAPREDUCE-1378.patch
> The logFile argument in the job links on the JT jobhistory.jsp page is not properly URL
encoded leading to links that result in 500 errors. I found the issue while working with the
Cloudera distro which contained a plus ('+') in the path which is interpreted as a space character
(%20) by Firefox. Here is the (trimmed) URL. Note the hadoop-0.20.1+152 directory which should
be hadoop-0.20.1%2B152. I have created a patch against current ASF svn trunk but it is untested
(although the jsp compiles to a class file ok).
> A job link from http://host:50030/jobhistory.jsp:
> http://host:50030/jobdetailshistory.jsp?jobid=job_201001141235_0001&logFile=file:/Users/esammer/hadoop-0.20.1+152/logs/history/done/...

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message