Don't pick tried endpoint & count the latest in ewma balancer #6639

spacewander · 2020-12-17T01:48:39Z

What this PR does / why we need it:

Types of changes

Bug fix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (fix or feature that would cause existing functionality to change)

Which issue/s this PR fixes

fixes #6632

How Has This Been Tested?

Checklist:

My change requires a change to the documentation.
I have updated the documentation accordingly.
I've read the CONTRIBUTION guide
I have added tests to cover my changes.
All new and existing tests passed.

k8s-ci-robot · 2020-12-17T01:48:42Z

Thanks for your pull request. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

📝 Please follow instructions at https://git.k8s.io/community/CLA.md#the-contributor-license-agreement to sign the CLA.

It may take a couple minutes for the CLA signature to be fully registered; after that, please reply here with a new comment and we'll verify. Thanks.

If you've already signed a CLA, it's possible we don't have your GitHub username or you're using a different email address. Check your existing CLA data and verify that your email is set on your git commits.
If you signed the CLA as a corporation, please sign in with your organization's credentials at https://identity.linuxfoundation.org/projects/cncf to be authorized.
If you have done the above and are still having issues with the CLA being reported as unsigned, please log a ticket with the Linux Foundation Helpdesk: https://support.linuxfoundation.org/
Should you encounter any issues with the Linux Foundation Helpdesk, send a message to the backup e-mail support address at: [email protected]

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

k8s-ci-robot · 2020-12-17T01:48:47Z

Welcome @spacewander!

It looks like this is your first PR to kubernetes/ingress-nginx 🎉. Please refer to our pull request process documentation to help your PR have a smooth ride to approval.

You will be prompted by a bot to use commands during the review process. Do not be afraid to follow the prompts! It is okay to experiment. Here is the bot commands documentation.

You can also check if kubernetes/ingress-nginx has its own contribution guidelines.

You may want to refer to our testing guide if you run into trouble with your tests not passing.

If you are having difficulty getting your pull request seen, please follow the recommended escalation practices. Also, for tips and tricks in the contribution process you may want to read the Kubernetes contributor cheat sheet. We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to Kubernetes. 😃

k8s-ci-robot · 2020-12-17T01:48:48Z

Hi @spacewander. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

aledbf · 2020-12-17T03:03:23Z

/ok-to-test

aledbf · 2020-12-17T03:03:34Z

/assign @ElvinEfendi

ElvinEfendi · 2020-12-17T03:05:55Z

rootfs/etc/nginx/lua/util/split.lua

@@ -18,6 +18,12 @@ function _M.get_first_value(var)
  return t[1]
 end

+function _M.get_last_value(var)


can you write a unit test for this function?

ElvinEfendi · 2020-12-17T03:06:49Z

Close #6632.

this PR addresses only one part of that issue, the algorithm would still potentially try a failed endpoint

spacewander · 2020-12-17T08:16:49Z

@ElvinEfendi
Updated.

ElvinEfendi · 2020-12-17T14:04:45Z

rootfs/etc/nginx/lua/balancer/ewma.lua

+  local lowest_score
+  for i = 1, k do
+    local key = get_upstream_name(peers[i])
+    if chosen_endpoints[key] then


Let's assume peers has 10 elements, out of which 2 are already tried. With the approach you implemented there's still a chance that we pick already tried upstream. But alternatively we can introduce a new list i.e filtered_peers, which is a copy of peers without the 2 already tried peers. That way we guarantee that we don't pick an already tried peer. I'm curious why you did not take this approach instead?

@ElvinEfendi
Aha, I didn't notice there is a PICK_SET_SIZE limitation.
Just curiously, why we set a PICK_SET_SIZE limitation and its value is 2? The original implementation (https://github.com/twitter/finagle/blob/1bc837c4feafc0096e43c0e98516a8e1c50c4421/finagle-core/src/main/scala/com/twitter/finagle/loadbalancer/PeakEwma.scala) seems don't have the same limitation.

That thing seems to be specifically about calculating peak ewma. I don't see shuffling or picking nodes in that code.

I think value 2 is based on "Michael Mitzenmacher. 2001. The Power of Two Choices in Randomized Load Balancing. IEEE Trans. Parallel Distrib. Syst. 12, 10 (October 2001), 1094-1104."

ElvinEfendi · 2020-12-17T14:06:21Z

rootfs/etc/nginx/lua/balancer/ewma.lua

@@ -170,20 +190,30 @@ function _M.balance(self)
  if #peers > 1 then
    local k = (#peers < PICK_SET_SIZE) and #peers or PICK_SET_SIZE
    local peer_copy = util.deepcopy(peers)
-    endpoint, ewma_score = pick_and_score(peer_copy, k)
+
+    local chosen_endpoints


What about tried_endpoints?

aledbf · 2020-12-18T10:15:14Z

@spacewander please squash the commits and reword the commit message

fixes kubernetes#6632

spacewander · 2020-12-18T11:28:19Z

@aledbf
Done.

ElvinEfendi · 2020-12-23T17:04:37Z

rootfs/etc/nginx/lua/balancer/ewma.lua

@@ -170,20 +177,50 @@ function _M.balance(self)
  if #peers > 1 then
    local k = (#peers < PICK_SET_SIZE) and #peers or PICK_SET_SIZE
    local peer_copy = util.deepcopy(peers)


I think this is unnecessary now, no? Because below you effectively do the same by copying peers to filtered_peers?

ElvinEfendi · 2020-12-23T17:10:21Z

rootfs/etc/nginx/lua/test/balancer/ewma_test.lua

+
+      local peer = two_endpoints_instance:balance()
+      assert.equal("10.10.10.3:8080", peer)
+    end)


I don't understand the logic in these two tests, where are you marking the endpoints as tried? I'd have expected you modify ngx.ctx.balancer_ewma_tried_endpoints to control these tests.

--

Also an additional test is needed to ensure ngx.ctx.balancer_ewma_tried_endpoints is set in balance function.

ngx.ctx is shared between tests. I will change the test to use ngx.ctx directly so the test will be more clear.

@ElvinEfendi
Updated.

I see, yeah it is better to keep the tests independent of each other.

ElvinEfendi · 2020-12-24T02:49:39Z

/lgtm

k8s-ci-robot · 2020-12-24T02:49:55Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ElvinEfendi, spacewander

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:

~~OWNERS~~ [ElvinEfendi]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

k8s-ci-robot added do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. do-not-merge/invalid-commit-message Indicates that a PR should not merge because it has an invalid commit message. labels Dec 17, 2020

k8s-ci-robot added the cncf-cla: no Indicates the PR's author has not signed the CNCF CLA. label Dec 17, 2020

k8s-ci-robot added the needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. label Dec 17, 2020

k8s-ci-robot added the size/S Denotes a PR that changes 10-29 lines, ignoring generated files. label Dec 17, 2020

k8s-ci-robot requested review from cmluciano and ElvinEfendi December 17, 2020 01:48

spacewander marked this pull request as ready for review December 17, 2020 02:52

k8s-ci-robot added cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. and removed do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. cncf-cla: no Indicates the PR's author has not signed the CNCF CLA. labels Dec 17, 2020

k8s-ci-robot added ok-to-test Indicates a non-member PR verified by an org member that is safe to test. and removed needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. labels Dec 17, 2020

k8s-ci-robot assigned ElvinEfendi Dec 17, 2020

ElvinEfendi reviewed Dec 17, 2020

View reviewed changes

k8s-ci-robot added size/M Denotes a PR that changes 30-99 lines, ignoring generated files. and removed size/S Denotes a PR that changes 10-29 lines, ignoring generated files. labels Dec 17, 2020

k8s-ci-robot added size/L Denotes a PR that changes 100-499 lines, ignoring generated files. and removed size/M Denotes a PR that changes 30-99 lines, ignoring generated files. labels Dec 17, 2020

ElvinEfendi reviewed Dec 17, 2020

View reviewed changes

Don't pick tried endpoint & count the latest in ewma balancer

e118ebc

fixes kubernetes#6632

spacewander force-pushed the use_last_for_ewma branch from cab3c85 to e118ebc Compare December 18, 2020 11:27

k8s-ci-robot removed the do-not-merge/invalid-commit-message Indicates that a PR should not merge because it has an invalid commit message. label Dec 18, 2020

spacewander changed the title ~~Count every try in ewma balancer~~ Don't pick tried endpoint & count the latest in ewma balancer Dec 18, 2020

ElvinEfendi reviewed Dec 23, 2020

View reviewed changes

Update for review

06b200f

k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Dec 24, 2020

k8s-ci-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Dec 24, 2020

k8s-ci-robot merged commit b022ea8 into kubernetes:master Dec 24, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Don't pick tried endpoint & count the latest in ewma balancer #6639

Don't pick tried endpoint & count the latest in ewma balancer #6639

spacewander commented Dec 17, 2020 •

edited by aledbf

Loading

k8s-ci-robot commented Dec 17, 2020

k8s-ci-robot commented Dec 17, 2020

k8s-ci-robot commented Dec 17, 2020

aledbf commented Dec 17, 2020

aledbf commented Dec 17, 2020

ElvinEfendi Dec 17, 2020

ElvinEfendi commented Dec 17, 2020

spacewander commented Dec 17, 2020

ElvinEfendi Dec 17, 2020

spacewander Dec 18, 2020 •

edited

Loading

ElvinEfendi Dec 18, 2020

ElvinEfendi Dec 17, 2020

spacewander Dec 18, 2020

aledbf commented Dec 18, 2020

spacewander commented Dec 18, 2020

ElvinEfendi Dec 23, 2020

ElvinEfendi Dec 23, 2020

spacewander Dec 24, 2020

spacewander Dec 24, 2020

ElvinEfendi Dec 24, 2020

ElvinEfendi commented Dec 24, 2020

k8s-ci-robot commented Dec 24, 2020

Don't pick tried endpoint & count the latest in ewma balancer #6639

Don't pick tried endpoint & count the latest in ewma balancer #6639

Conversation

spacewander commented Dec 17, 2020 • edited by aledbf Loading

What this PR does / why we need it:

Types of changes

Which issue/s this PR fixes

How Has This Been Tested?

Checklist:

k8s-ci-robot commented Dec 17, 2020

k8s-ci-robot commented Dec 17, 2020

k8s-ci-robot commented Dec 17, 2020

aledbf commented Dec 17, 2020

aledbf commented Dec 17, 2020

Choose a reason for hiding this comment

ElvinEfendi commented Dec 17, 2020

spacewander commented Dec 17, 2020

Choose a reason for hiding this comment

spacewander Dec 18, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

aledbf commented Dec 18, 2020

spacewander commented Dec 18, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ElvinEfendi commented Dec 24, 2020

k8s-ci-robot commented Dec 24, 2020

spacewander commented Dec 17, 2020 •

edited by aledbf

Loading

spacewander Dec 18, 2020 •

edited

Loading