Mike,

Thanks for the reminder!  I've updated some of mine.

All > These are topics for review on Wednesday.  Please take time to review!

http://pad.opencrowbar.org/p/community-planning-meeting-2014-06-04

Here's a copy of the material:

# Stories for Review

## Ceph IPv6 (will be needed for OpenStack)
Victor...

## Online service for tools for CI/CD
Scott...

## Hadoop "Ready State" Prep:

Hadoop Readiness Barclamp
=========================

The hadoop readiness barclamp is responsible for transitioning nodes
from the base opencrowbar state to a state ready for an Apache
Hadoop installation.

Requirements
------------

General:
    * Hadoop distro agnostic

Supported Node Roles:
    * Hadoop Data node
    * Hadoop Infrastructure node (name node, resource manager, etc)
    * Hadoop edge or gateway node

Storage:
    * Format and create filesystems on data disks
    * mount disks using Hadoop naming convention (/data/<n>)
    * mount disks with appropriate options (noatime, etc)
    * create necessary directories with correct permissions

Operating System:
    * create core users and groups (hdfs, etc)
    * configure ulimits
    * set OS Kernel options (e.g. thp compaction, selinux, swapping)

Networking:
    * configure interface bonding and mode
    * configure network MTU
    * check / configure firewall

Anti-Requirements:
    * will not install JDK / JVM (licensing, version dependencies)
    * will not stage or install any Hadoop distro


## HW openWSMAN jig  +1
As an operator, I want to be able to use Crowbar to configure the RAID and BIOS 
of my servers using WSMAN interfaces (e.g.: Dell iDRAC, HP ILO).

I expect that
* my Crowbar admin server will have network access to the BMC network and know 
which IP/Credentials for servers to configure them.
* Crowbar will configure the server's out of band (no boot image is needed and 
the machine does not have to boot)
* default credentials will be assumed if none are give
* credentials will be set from the template or deployment attributes
* specific attributes will be set/determined using the attribute system 
attached to the deployment

Technical: In the opencrowbar/hardware repo, create a WSMAN jig using the 
openWSMAN gem (http://openwsman.github.io/).

need to review designs from Rajeev
Rob found Dell WSMAN simulator - it has an open license but its not clear that 
we can re-distribute

## HW split out for inband jig (OpenCompute)
As an operator, I want to be able to use Crowbar to configure RAID and BIOS on 
systems that do not have out-of-band BMC configuration.  These systems require 
that Crowbar boot an operating systems (assume Centos 6.5) and run local tools 
on those systems to complete the configuration.  A reboot will be required.

I expect that
* my Crowbar admin server will SSH into the server to complete these actions 
using the Admin network
* Crowbar will configure the server from the boot image (likely sledgehammer)
* no credentials are needed
* specific configuration attributes will be injected from the attribute system

Technical: I expect to use the SSH or Chef-Solo jig to complete these actions.

## OpenStack (Chef Server needed) - about getting Ready State for upstreams NOT 
advantaging CB ahead of upstream.  some work will be needed in upstreams

As an operator, I want to be able to use Crowbar to drive the OpenStack Chef 
Upsteams.  While I would like to have many cookbooks, I am ok to start with 
Nova + Nova-Net, Keystone, Glance and Cinder. My priority is to be able to use 
the upstreams first and then work in the community to help them mature.  I have 
no interest in getting them from a single vendor or being tied to a single OS.  
For that reason, I expect them to work on at least Ubuntu and Centos.

Overtime, I'd like to have (in order)
* a default HA configuration.
* Neutron as the networking default (GRE tunnels & VLAN)
* Heat
* Use the Ceph deploy as the storage target

Architecture:
As an operator, I expect that the controller nodes will be setup with RAID5 and 
standard BIOS.  Compute nodes will use RAID10 with virtualization enabled.  
Storage nodes will use JBOD with standard BIOS.  I also expect admin, public, 
private and storage networks to be configured.  The transport networks should 
use teamed NICs.

Technical: Needs to be able to use Chef server because upstreams assume 
encrypted databags.

_______________________________________________
Crowbar mailing list
[email protected]
https://lists.us.dell.com/mailman/listinfo/crowbar
For more information: http://crowbar.github.com/

Reply via email to