891409 - [socorro-bixie] Build Staging Environment

Status: RESOLVED INCOMPLETE

(Infrastructure & Operations Graveyard :: WebOps: Socorro, task, P1)

Description

[Brandon Burton [:solarce]]

From email

>> For testing Bixie we will need a home for the bixie-collector, bixie-processor, bixie-crontabber, bixie-crashmover, bixie-middleware, bixie-webapp(django). We also need access to RabbitMQ, Postgres, HBase, and ElasticSearch. Existing HBase and RabbitMQ can be reused. 

>> I haven't spoken to Adrian, but I'm reasonably sure that existing ES can be used. Selena is nervous about disk space and wants a clone of our staging DB setup before we put the bixie infra under any regular load.


STAGE:
    5 host cluster
      * admin node for crontabber, crashmover
      * collector, processor, mware, and webapp get one node each
      * netflows to socorro stage datastores

Data Stores:
    Clone stage's existing Database infra

Comment 1

[Brandon Burton [:solarce]]

The initial batch of manifests, templates, and files have been committed to Puppet svn with the completed of bug 904690

I've gone through everything and done a simple rename of 'stage' to 'bixie_stage' to start the customization

I haven't applied to the new Puppet classes to the node definitions yet, I'll work with :lonnen to get them into a better state and we'll begin testing them.