network: on premise Service load balancers #356
Conversation
openshift-ci-robot
added
the
approved
|
/assign @cybertron @yboaron |
|
Note: I will squash commits once review has completed. |
|
I believe I have addressed all open feedback. The enhancement does not clarify a detailed plan for OpenShift integration, but it does highlight the options that must be explored in the next phase of this work. The outcome I hope to reach here is consensus on whether MetalLB is a good direction to pursue. The next step would be two things in parallel:
As I noted earlier, I will squash this down to one commit before merging. |
|
Just wanted to note that I have a PoC operator for deploying metallb. I did a quick demo of it here: https://www.youtube.com/watch?v=WgOZno0D7nw |
OpenShift does not currently support Services of type=LoadBalancer on bare metal or other on premise infrastructure environments. This enhancement proposes a way forward, which is to adopt MetalLB. The document explores some high level requirements, discusses some alternatives considered, and maps out how we would get to work on this through technical due diligence, upstream community engagement, and careful planning of OpenShift integration with a new operator. There is not currently a target for when this would be fully supported, as this is a proposed enhancement on a direction to take. More extensive technical due diligence, development, and testing will help define the roadmap over time.
russellb
force-pushed
the
on-prem-load-balancers
branch
from
da78454 to
5999c9b
Compare
Thanks. I've added this to the doc, squashed all the commits, and rebased on master. |
Despite being called MetalLB, the project does not actually implement load balancing as it is typically thought of (something like haproxy). It is one piece of a larger setup that results in load balancing for Service Load Balancers. These additions to the doc attempt to help explain further how load balancing works with MetalLB.
Update this enhancement to list some approvers and reviewers. The reviewers reflect those who have commented so far.
|
@knobunc @danwinship @danehans The conversation on this one seems to have settled. Note that it's in a "proposed" state. I propose that we merge this and expect a follow-up later to address the remaining open questions around design details once we're ready to spend more time on this. |
|
/lgtm |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: danwinship, russellb The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
OpenShift does not currently support Services of type=LoadBalancer on
bare metal or other on premise infrastructure environments. This
enhancement proposes a way forward, which is to adopt MetalLB.
The document explores some high level requirements, discusses some
alternatives considered, and maps out how we would get to work on
this through technical due diligence, upstream community engagement,
and careful planning of OpenShift integration with a new operator.
There is not currently a target for when this would be fully
supported, as this is a proposed enhancement on a direction to take.
More extensive technical due diligence, development, and testing will
help define the roadmap over time.