Hmaster HA via an API call

to create new topics or reply. | New User Registration

This topic contains 1 reply, has 2 voices, and was last updated by  Sasha J 2 years, 5 months ago.

  • Creator
  • #15987

    Francois BORIE

    Hi !

    I know for the moment it’s not possible to deploy other Hmaster instances with Ambari Web UI.

    I was wondering if there were any possibilities to do that via an API call. For example, if I want to deploy another HBASE_MASTER on a node which is just running a zookeeper instance, I try something like below but it fails with a 404 error :

    [root@oben**** ambari-server]# curl –user admin:admin -d ‘{“HostRoles”:{“cluster_name” : “hadoop_poc”,”component_name”: “HBASE_MASTER”,”host_name”:”mytestnode.priv”}}’ -X PUT http://oben****:8001/api/v1/clusters/hadoop_poc/hosts/mytestnode.priv/host_components/HBASE_MASTER{
    “status” : 404,
    “message” : “org.apache.ambari.server.controller.spi.NoSuchResourceException: The requested resource doesn’t exist: ServiceComponentHost not found, clusterName=hadoop_poc, serviceName=HBASE, serviceComponentName=HBASE_MASTER, hostName=mytestnode.priv”

    So I took an alternative way by replaying the puppet manifest of the main Hbase master visible in Ambari, on my testnode where I want to deploy my second hmaster.

    (I do something like :
    export RUBYLIB=’/usr/lib/ambari-agent/lib/puppet-2.7.9/lib/:/usr/lib/ambari-agent/lib/facter-1.6.10/lib’
    /usr/lib/ambari-agent/lib/ruby-1.8.7-p370/bin/ruby /usr/lib/ambari-agent/lib/puppet-2.7.9/bin/puppet apply –confdir=/var/lib/ambari-agent/puppet –detailed-exitcodes /var/lib/ambari-agent/data/mymanifestwhichdeployshmaster.pp

    It’s working and I get a new Hmaster automatically configured on another node, working in a cluster with the other Hmaster, but it looks like a quick & dirty way to achieve Hmaster HA for the moment.

    Many thanks for your answer,



Viewing 1 replies (of 1 total)

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

  • Author
  • #16008

    Sasha J

    Ambari in it’s current release not supporting services addition or moving to different node.
    Your manual trick may work, but it is not supported.

    Thank you!

Viewing 1 replies (of 1 total)
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.