Jump to: navigation, search

Baremetal

Revision as of 15:28, 23 July 2012 by Mikyung (talk)

Summary

This blueprint proposes to support general bare-metal provisioning framework in OpenStack.

The target release for this is Folsom. USC/ISI and NTT docomo are working on integration of bare-metal provisioning implementation to support following stuff:

  • Support PXE and non-PXE bare-metal machines
  • Support several architecture types such as x86_64, tilepro64, and arm
  • Support fault-tolerance of bare-metal nova-compute node

The USC/ISI team has a branch here (general bare-metal provisioning framework and non-PXE support):

The NTT docomo team has a functional prototype here (additional bare-metal features and PXE support):

An etherpad for discussion of this blueprint is available at http://etherpad.openstack.org/FolsomBareMetalCloud

Final merged github branch: WORKING

Release Note

TO DO

Overview

1) A user requests a baremetal instance.

  • euca-run-instances -t baremetal.small --kernel aki-AAA --ramdisk ari-BBB ami-CCC

2) nova-scheduler selects a baremetal nova-compute.

  • nova.conf
    • --baremetal_driver = pxe/non-pxe(=tilera)
    • --physical_deploy_kernel = xxx
    • --physical_deploy_ramdisk = yyy
    • --instance_type_extra_specs=cpu_arch:x86_64/tilepro64/arm
  • vcpus = the maximum number of cpus/cores among bare-metal nodes' resources

3) The bare-metal nova-compute selects a bare-metal node from its pool based on hardware resources and the instance type (# of cpus, memory, HDDs).

4) The kernel and ramdisk are prepared and then TFTP/PXE is set up.

  • PXE: kernel and ramdisk for the deployment, and the user specified kernel and ramdisk are put to TFTP server. PXE is configured for the bare-metal nodes.
  • Non-PXE: NFS/TFTP is configured for the bare-metal nodes.

5) The bare-metal nova-compute powers on the bare-metal node.

  • PXE/IPMI: through IPMI
  • Non-IPMI: through PDU(Power Distribution Unit)

6) The images are deployed to bare-metal nodes.

  • PXE: the bare-metal node uses the deployment kernel and ramdisk, and the bare-metal nova-compute writes AMI to the bare-metal node's local disk via iSCSI.
  • Non-PXE: nova-compute mounts AMI into NFS directory or passes AMI to the bare-metal node's local disk.

7) The bare-metal node is rebooted.

  • The bare-metal node is booted up by the user specified kernel, ramdisk and its local disk.

Design

TO DO

Implementation

TO DO