You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ambari.apache.org by Jabir Ahmed <ja...@gmail.com> on 2018/02/13 18:44:33 UTC

Re: Ansible module for Ambari

From my experience over using puppet , ambari and ansible for managing
hadoop, its better to stop using ambari and just write your own ansible
scripts to deploy the cluster & configs.

It would surely give you more control over the deployment also it would
give you a deeper understanding of how an incremental update happens
internally.

Thanks

Jabir

--
Linkedin <https://www.linkedin.com/in/jabirahmed/>


On Tue, Jan 30, 2018 at 8:44 AM, Ali Nazemian <al...@gmail.com> wrote:

> How can we update cluster incrementally if there is no Ambari module for
> Ansible? We will end up implementing Ambari APIs in Ansible playbook
> anyway, or we may ignore idempotency completely to simplify it or even we
> may say update cluster manually!
>
> Wouldn't be better to rely on an approach that can support incremental
> changes from the first place? Since we need to follow two different paths
> (one for initial build and another for incremental updates), how can we
> make sure the output would be the same?
>
> Regards,
> Ali
>
> On Thu, Jan 25, 2018 at 6:54 PM, Doroszlai, Attila <ad...@apache.org>
> wrote:
>
>> Hi Ali,
>>
>> I think this should work:
>>
>>  * install via blueprint
>>  * update cluster incrementally
>>  * export cluster in blueprint format (GET /clusters/:name?format=bluepri
>> nt)
>>  * (might need to replace a few properties depending on your services)
>>  * use exported blueprint for next install
>>
>> -Attila
>>
>> On Thu, Jan 25, 2018 at 8:41 AM, Ali Nazemian <al...@gmail.com>
>> wrote:
>> > Since the incremental update is not supported by blueprint, I was
>> hoping to
>> > use an Ansible module for Ambari to first install Ambari and then use
>> Ambari
>> > API calls to install different services one-by-one. What is the
>> recommended
>> > solution for build and operation automation on Ambari managed-cluster,
>> then?
>> > How can we make sure after a cluster is built we can update it and
>> create
>> > another completely identical platform with minimum effort? The
>> blueprint can
>> > only help if you want to create it for the initial time, but making sure
>> > that you can manage all the changes automatically and create another
>> > identical platform easily is almost impossible with Blueprint.
>> >
>> > Regards,
>> > Ali
>> >
>> > On Thu, Jan 25, 2018 at 6:32 PM, aman poonia <am...@gmail.com>
>> > wrote:
>> >>
>> >> This might help
>> >> https://github.com/hortonworks/ansible-hortonworks
>> >>
>> >> Also there is a puppet module for ambari in bigtop. That might help you
>> >> and give you more flexibility but that depends on what you want to
>> achieve.
>> >>
>> >> https://github.com/apache/bigtop/tree/master/bigtop-deploy/
>> puppet/modules/ambari
>> >>
>> >>
>> >> --
>> >> With Regards,
>> >> Aman Poonia
>> >>
>> >> On Thu, Jan 25, 2018 at 12:50 PM, Luciano Resende <
>> luckbr1975@gmail.com>
>> >> wrote:
>> >>>
>> >>> Take a look at the following repository:
>> >>>
>> >>> https://github.com/lresende/spark-cluster-install
>> >>>
>> >>> On Wed, Jan 24, 2018 at 10:59 PM Ali Nazemian <al...@gmail.com>
>> >>> wrote:
>> >>>>
>> >>>> Hi All,
>> >>>>
>> >>>> I was wondering whether there is any Ansible module available for
>> >>>> Ambari. What about Ambari Python client? We want to automate our
>> cluster
>> >>>> deployment and operations with Ansible, but we couldn't find an
>> Ansible
>> >>>> module for it yet. Problem is if we use blueprint for this purpose
>> we are
>> >>>> not able to address incremental updates, so we need to rely on API
>> calls for
>> >>>> any changes.
>> >>>>
>> >>>> Regards,
>> >>>> Ali
>> >>>
>> >>> --
>> >>> Sent from my Mobile device
>> >>
>> >>
>> >
>> >
>> >
>> > --
>> > A.Nazemian
>>
>
>
>
> --
> A.Nazemian
>