Work environment for Bootstrap/Get-started
Note: The below is a first draft cut at a work environment for Bootstrap/Get-started (BGS).
BGS is to deploy to both bare metal as well as to a virtual environment. This requires a physical server environment for BGS.
Possible hardware scenario's for OPNFV
Development Environment Layout
Assumption is that a total of 5 PODs is required:
BGS targets a deployment with HA/cluster support. As a result, the following POD configuration is assumed:
3 x Control node (for HA/clustered setup of OpenStack and OpenDaylight)
2 x Compute node (to bring up/run VNFs)
1 x Jump Server/Landing Server in which the installer runs in a VM (FUEL)
Total: A total of 30 servers (5 PODs with 6 Servers each) is required.
Server configuration
Typical server configuration (same server for all components of the POD assumed for reasons of simplicity)
Server:
CPU: Intel Xeon E5-2600 (IvyBridge at least, or similar)
Disk: 4 x 500G-2T + 1 x 300GB SSD (leave some room for experiments)
First 2 disks should be combined to form a 1 TB virtual store for the
OS/Software etc
Remaining should be combined to form a virtual disk for CEPH storage.
The 5'th disk (SSD) for distributed storage (CEPH) journal towards SSD technology.
Performance testing requires a mix of compute nodes that have CEPH(swift+Cinder) and without CEPH storage
Virtual ISO boot capabilities or a separate PXE boot server (DHCP/tftp or Cobbler)
Access to console ports/lights-out-management through management tool and/or serial console server
Lights-out-management/Out-of-band management for power on/off/reset
Memory: >= 32G RAM (Minimum)
Single Power supply active with spares in the Lab for power supply failure addressing.
I/O
Option 1: 4x1G Control, 2x40G Data, 48 Port Switch
Connectivity to each network is through a separate NIC that simplifies Switch Management. However, requires a more NICs on the server and also more switch ports.
1 x 1G for ILMI (Lights out Management )
1 x 1G for Admin/PXE boot
1 x 1G for control Plane connectivity
1 x 1G for storage
"NICs" can be internal in case a blade server is used
2 x 40G (or 10G) for data network (redundancy, NIC bonding, High bandwidth testing)
Option II: 1x1G Control, 2x 40G (or 10G) Data, 24 Port Switch
Option III: 2x1G Control, 2x10G Data, 2x40G Storage, 24 Port Switch
Power: Single power supply acceptable (redundant power not required/nice to have)
Switch:
TOR switch should support 1G/10G/40G links (either 4 of them or VLAN isolated to support 4 networks)
Uplink from the Jump server to Internet must be 1G or better.
Public IP address Pool per pod (8)
Private address pool per pod 3 x /24 subnets (either not shared or vlan isolated)
Additional links/port to support Ceph(Swift + Cinder Vols) on atleast 3-nodes to 5-nodes for opnfv controller and other Pocs.
Additional requirements (if servers are offered as MaaS):
Example Pod Configuration