Hue Forum

oozie workflow fails on submit from hue

  • #49884
    D Blair Elzinga
    Participant

    I have a shared workflow that works fine when run as the owner as the workflow from hue. But if I launch it from another user account, then I get the following error:

    Failed to access deployment directory.

    AccessControlException: Permission denied: user=zing, access=READ_EXECUTE, inode=”/user/hue/oozie/workspaces/_mapred_-oozie-2614-1394040961.07″:mapred:hue:drwx–x–x (error 403)

    The python code /usr/lib/hue/desktop/libs/liboozie/src/liboozie/submittion.py doesn’t try to create the directory using a “doas” call, but rather just gets the owner (mapred in this case) to create the deployment directory, and then tries to access the directory (to copy files there for submit) using the current user.

    Looks like a bug to me, or perhaps I misunderstand the intent of hue an what it does with permissions. The python code has HARDCODED directory permissions for the directory created as 711 (!?why?) . If hue doesn’t want anyone but the owner to be able to submit a workflow, why doesn’t it just check that first before creating an empty directory that then just gets left lying there?

    I would expect that a shared workflow could be submitted by anyone that had permission to read it.

    Also note that even though it is a shared workflow, I’m unable to copy it from Hue as well.

to create new topics or reply. | New User Registration

  • Author
    Replies
  • #55261
    Patrick Alexander
    Participant

    I have same problem too, any solution yet ?

    #55406
    Dave
    Moderator

    Hi,

    Can you output here:

    hadoop version
    rpm -qa | grep hue

    Thanks

    Dave

You must be to reply to this topic. | Create Account

Support from the Experts

A HDP Support Subscription connects you experts with deep experience running Apache Hadoop in production, at-scale on the most demanding workloads.

Enterprise Support »

Become HDP Certified

Real world training designed by the core architects of Hadoop. Scenario-based training courses are available in-classroom or online from anywhere in the world

Training »

Hortonworks Data Platform
The Hortonworks Data Platform is a 100% open source distribution of Apache Hadoop that is truly enterprise grade having been built, tested and hardened with enterprise rigor.
Get started with Sandbox
Hortonworks Sandbox is a self-contained virtual machine with Apache Hadoop pre-configured alongside a set of hands-on, step-by-step Hadoop tutorials.
Modern Data Architecture
Tackle the challenges of big data. Hadoop integrates with existing EDW, RDBMS and MPP systems to deliver lower cost, higher capacity infrastructure.