[Starlingx-discuss] [DOCS] Change compute node to worker node personality

Tullis, Michael L michael.l.tullis at intel.com
Thu May 9 20:38:16 UTC 2019


Thanks for this submission. We'll add this to the agenda for the upcoming doc team meeting.

-- Mike

________________________________
From: Sullivan, David [David.Sullivan at windriver.com]
Sent: Thursday, May 09, 2019 1:06 PM
To: starlingx-discuss at lists.starlingx.io
Subject: [Starlingx-discuss] [DOCS] Change compute node to worker node personality

As part of this story there are number of changes that can impact documentation. I’ve described the changes below. Let me know if you require any further details.

-David

---

This story encompasses the move from openstack compute nodes to docker/kubernetes worker nodes. The changes involved renaming compute nodes to worker nodes, renaming VM functions to Application and making vswitch and nova-local optional.


Changed functionality:

Compute personality/subfunction renamed to 'worker'
CLI changes: host-add, host-bulk-add, host-update, host-show, host-list, dcmanager patch
Horizon Changes: Host Inventory, Host personality, edit host, host overview, patch orchestration, dcmanager patch
The pxeboot console has also been updated to use worker

CPU VM function renamed to 'Application'
CLI changes: host-cpu-*
Horizon changes: host processor

VM huge pages renamed to 'Application huge pages'
CLI changes: host-memory-*
Horizon changes: host memory, host update memory, force-host-lock


vswitch changes
* If vswitch_type is 'None'
  - newly installed worker hosts will start with 0 vswitch cpus
  - vswitch cpus can only be set to 0 through 'system host-cpu-modify' or platform horizon
  - vswitch cpus must be 0 when a worker node is unlocked

* If vswitch_type is 'ovs-dpdk'
  - newly installed worker hosts will start with 2 vswitch cpus (this behavior is unchanged)
  - vswitch cpus must be at least 1 when a worker node is unlocked


nova-local changes
* The nova-local volume group is now optional. It is only required if the worker node has the openstack-compute-node=enabled label present.


huge page changes
* vswitch huge pages will only be allocated if vswitch_type is not 'None'


Stories:
https://storyboard.openstack.org/#!/story/2004022
https://storyboard.openstack.org/#!/story/2004763
https://storyboard.openstack.org/#!/story/2004649


Reviews:
https://review.opendev.org/#/c/656586/
https://review.opendev.org/#/c/656583/
https://review.opendev.org/#/c/651889/
https://review.opendev.org/#/c/631541/
https://review.opendev.org/#/c/646038/
https://review.opendev.org/#/c/633924/

https://review.opendev.org/#/c/624452/
https://review.opendev.org/#/c/624564/
https://review.opendev.org/#/c/624715/
https://review.opendev.org/#/c/624718/
https://review.opendev.org/#/c/624512/
https://review.opendev.org/#/c/624513/
https://review.opendev.org/#/c/624727/
https://review.opendev.org/#/c/624517/
https://review.opendev.org/#/c/626366/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.starlingx.io/pipermail/starlingx-discuss/attachments/20190509/ed5ff8b5/attachment-0001.html>


More information about the Starlingx-discuss mailing list