Ambari 1.5.1 problem with upgrades

to create new topics or reply. | New User Registration

This topic contains 0 replies, has 1 voice, and was last updated by  Pramod Thangali 1 year, 3 months ago.

  • Creator
  • #52842

    Pramod Thangali

    After upgrading to Ambari 1.5.1, Ambari agents are configured with wrong repo URL for HDP and most management operations will fail.

    Root Cause

    Beginning in Ambari 1.5.1, the agents that report to the Ambari server have been changed to an OS Family instead of a specific Operating System. Centos6, RedhHat6, and Oracle Linux6 now all belong to the redhat6 family. If you have set a local repo for one of these specific operating systems, and you upgrade to Ambari 1.5.1, the local repo configuration in the Ambari database and what the agent reports will be out of sync. Refer to the following list for a complete mapping of operating systems to OS Family:

    centos6, oraclelinux6, redhat6: redhat6

    centos5, oraclelinux5, redhat: redhat5

    SLES, Suse: suse


    If you configured local repo for Centos6, after the upgrade agents will report host OS as ‘redhat6’ and will be configured with the wrong repo URL.

    Note: If all the hosts in your clusters are redhat6, you will not see this issue.

    This issue will be fixed in 1.6.0, currently planned for GA in May.

    Temporary Workaround

    Run the following command to set the local repo URL for the correct OS Family:

    ambari-server upgradestack <Stack Id> <local repo url> <OS Family>

    – Stack Id: Current stack id for the cluster, eg: HDP-2.0.6, HDP-2.1, etc.
    – Repo URL: Local repo URL
    – OS Family: redhat(5 /6), suse11

    Example for centos6 clusters with HDP 2.1 Stack:
    ambari-server upgradestack HDP-2.1 http://localrepo redhat6

    Thank you,

    Ambari Team

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

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.