accumulo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dylan Hutchison <dhutc...@cs.washington.edu>
Subject Re: Automating deploy+testing
Date Thu, 28 Jan 2016 12:01:14 GMT
One of my friends has had success using Ansible to install the Myria
database and its many roles like Ganglia on multiple nodes. You could look
at her control configs to get an idea of whether it would be helpful here:
https://github.com/parmitam/myria-ec2-ansible

General install and use info here:
http://docs.ansible.com/ansible/intro_installation.html
<http://docs.ansible.com/ansible/intro_installation.html#installing-%20%20%20%20%20the-control-machine,>

I imagine you have a decision whether to use an existing framework vs roll
your own.

~Dylan

On Tuesday, January 26, 2016, Josh Elser <josh.elser@gmail.com> wrote:

> Thanks Dylan. I think I played around with Mike's code too when the book
> was still being written.
>
> Multi-node is definitely a necessity, so I don't think Maven makes a good
> driver (Bash and some scripting language seems to fit nicely). I'm hoping
> to just consider Hadoop/ZooKeeper as pre-requisites. Other products (e.g.
> Ambari) likely will do this just fine for what I care.
>
> I'll take a look at the AccStack project too. Thanks for linking it.
>
> Dylan Hutchison wrote:
>
>> It might be useful to look at the Accumulo quickinstall maven project.
>>
>> https://github.com/accumulobook/quickinstall
>>
>>
>> Michael Wall wrote the project and I helped test and patch it.  It runs
>> from a release but could be modified to build from source instead.
>>
>> There's also a more messy script I used to build Hadoop and Accumulo from
>> source, including the Hadoop native stuff.
>>
>> https://github.com/Stevens-GraphGroup/AccStack/blob/master/setup.bash
>>
>>
>> Cheers, Dylan
>>
>> On Mon, Jan 25, 2016 at 9:38 PM, Josh Elser<josh.elser@gmail.com>  wrote:
>>
>> FYI, my evening hack plans this week are going to be centered around
>>> creating some automation around building Accumulo, installing it on some
>>> nodes and then automatically running Continuous Ingest/Randomwalk.
>>>
>>> Looking back now, I can't hardly believe I'e gone this long without
>>> writing something.
>>>
>>> If anyone else is thinking about this for the upcoming 1.6.5 and 1.7.1
>>> testing, I'd be happy to work together. The approach should be agnostic
>>> on
>>> platform and bootstraps Accumulo by hand (SSH, bash, and python) with
>>> arguments (selfishly defaulting to things useful to me). I presently have
>>> it pulling down some Git ref, building and installing the tarball. Need
>>> to
>>> get configuration, copying the install/configs to many nodes, running CI
>>> and reporting some results at the end.
>>>
>>> https://github.com/joshelser/accumulo-automated-testing
>>>
>>> - Josh
>>>
>>>
>>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message