Home Forums HDP on Linux – Installation Unable to install with Ambari

This topic contains 6 replies, has 4 voices, and was last updated by  tedr 1 year, 1 month ago.

  • Creator
    Topic
  • #25090

    Angel Cortes
    Member

    Good morning,
    Im trying to install HDP1.2.1 from ambari 1.2.2.5-1 on Suse11 SP2.

    I have all the repos in a mirror, Ambari-server is installed, ambari registration and ambari-services running in all the host. I configure and arrive to Step9 of the wizzard and then…..freeze.

    The service show 3-6%(depends of the try) and stay for 1hour and then report a fail with all the services in timeout in all the servers.
    And you can read in the logs this message:
    [HDP-UTILS-1.1.0.15|http://public-repo-1.hortonworks.com/HDP-UTILS-1.1.0.15/repos/suse11] Can’t provide file ‘./snappy/snappy-1.0.5-1.x86_64.rpm’ from repository ‘HDP-UTILS-1.1.0.15′
    History:
    – Download (curl) error for ‘http://public-repo-1.hortonworks.com/HDP-UTILS-1.1.0.15/repos/suse11/snappy/snappy-1.0.5-1.x86_64.rpm’:
    Error code: Connection failed
    Error message: Couldn’t resolve host ‘public-repo-1.hortonworks.com’

    I selected USE LOCAL REPOSITORIES, and i have the snappy in the HDP-UTILS repository, why is trying to get it outside?

    I tried to disable the HDP.repo and the HDP-UTILS.repo created by ambari, but then is created again in the next retry. I created my repositories with other names to not be overwriten, set priority to 1…Nothing works. It uses the HDP and the HDP-UTILS isntead of mine.

    Please any help?

Viewing 6 replies - 1 through 6 (of 6 total)

You must be logged in to reply to this topic.

  • Author
    Replies
  • #26936

    tedr
    Moderator

    Hi Pariksheet,

    Is your version of SLES registered? If it is not you’ll not have access to the zypper repositories necessary to download the sofware.

    Thanks,
    Ted.

    Collapse
    #26934

    Hi

    Can you tell me how to install postgre with below version in SLES
    libpq5 9.1.5-0.2.1
    postgresql 8.3.20-0.4.1
    postgresql-init 9.1-0.6.10.1
    postgresql-server 8.3.20-0.4.1

    Thanks
    Pariksheet

    Collapse
    #25688

    tedr
    Moderator

    Hi Angel,

    It’s good to hear that you finally got the POC cluster up and running.

    Thanks,
    Ted.

    Collapse
    #25679

    Angel Cortes
    Member

    Hi Ted,
    That’s the first try i did.

    I finally get the cluster for the PoC running.
    After setting all the repositories and a lot of different errors i decide to read the logs “carefully”(78.000 lines) and found that a lot of problems come from the paralel installation of components and the use of zypper.

    In a lot of nodes i found messages of “Zypper fail cause a lock by PID (xxxxx)” and that PID is from another zypper trying to install other component.

    After read this error, I tried the retry of the install and wait for timeout until the checks get all the component installed, and in the 3rd retry, all the components are installed and the cluster is running.

    Thx for your help.

    Collapse
    #25422

    tedr
    Member

    HI Angel,

    Can you try a re-install without doing the ambari-server reset? In this case you will need to give a different name for the cluster.

    Thanks,
    Ted.

    Collapse
    #25091

    Angel Cortes
    Member

    new problem…

    I passed the repo problem with a proxy, only to try this distribution as i need to compare with others to prepare a huge install for a bank account,
    now the ambari UI is halt in “installing ganglia” for 30 minutes with all the other services in “pending”, but in logs i can see:
    INFO 2013-05-10 09:17:12,861 LiveStatus.py:124 – The live status for component GANGLIA_SERVER of service GANGLIA is {‘msg’: ”, ‘componentName’: ‘GANGLIA_SERVER’, ‘serviceName’: u’GANGLIA’, ‘status’: ‘INSTALLED’, ‘clusterName’: u’HDP’}
    INFO 2013-05-10 09:17:12,862 ActionQueue.py:117 – Got live status for component GANGLIA_SERVER of service GANGLIA of cluster HDP
    {‘clusterName’: u’HDP’,
    ‘componentName’: ‘GANGLIA_SERVER’,
    ‘msg’: ”,
    ‘serviceName’: u’GANGLIA’,
    ‘status’: ‘INSTALLED’}

    And for other services…

    INFO 2013-05-10 09:17:12,845 ActionQueue.py:117 – Got live status for component JOBTRACKER of service MAPREDUCE of cluster HDP
    {‘clusterName’: u’HDP’,
    ‘componentName’: ‘JOBTRACKER’,
    ‘msg’: ”,
    ‘serviceName’: u’MAPREDUCE’,
    ‘status’: ‘INSTALLED’}

    Thats is absurd, yesterday i had this problem too and one hour later the ambari report a timeout and no errors, in logs.

    In every try to reinstal the cluster i do an “ambari-server reset” to avoid problems with the database.

    Why i cant install it?

    Collapse
Viewing 6 replies - 1 through 6 (of 6 total)