github/kubernetes
1
0
Fork 0
mirror of https://github.com/k3s-io/kubernetes.git synced 2025-07-31 07:20:13 +00:00
Code Issues Projects Releases Wiki Activity

Merge pull request #66827 from shyamjvs/fix-expected-pod-churn-in-load-test

Browse Source 
Automatic merge from submit-queue (batch tested with PRs 66827, 60550). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Fix expected pod creations/deletions estimate in load test

Fixing a wrongly computed expectation for first round of scaling in load test. Luckily as a side-effect, this should also speed up load test by:

- 25s on 100-node cluster
- 2m on 500-node cluster (so helps our presubmit - https://github.com/kubernetes/test-infra/issues/8348)
- 20m on 5000-node cluster

That said, I'm not 100% sure if this won't start causing failures in practice.

/cc @gmarek @mborsz 

```release-note
NONE
```
This commit is contained in:
Kubernetes Submit Queue 2018-07-31 20:42:04 -07:00 committed by GitHub
commit 77bac51db8
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 9 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
test/e2e/scalability/load.go
View File

@ -320,12 +320,19 @@ var _ = SIGDescribe("Load capacity", func() {
// We would like to spread scaling replication controllers over time
// to make it possible to create/schedule & delete them in the meantime.
// Currently we assume that <throughput> pods/second average throughput.
// The expected number of created/deleted pods is less than totalPods/3.
scalingTime := time.Duration(totalPods/(3*throughput)) * time.Second
// The expected number of created/deleted pods is totalPods/4 when scaling
// for the first time, with each RC changing its size from X to a uniform
// random value in the interval [X/2, 3X/2].
scalingTime := time.Duration(totalPods/(4*throughput)) * time.Second
framework.Logf("Starting to scale %v objects first time...", itArg.kind)
scaleAllResources(configs, scalingTime, testPhaseDurations.StartPhase(300, "scaling first time"))
By("============================================================================")
// The expected number of created/deleted pods is totalPods/3 when scaling for
// the second time, with each RC changing its size from a uniform random value
// in the interval [X/2, 3X/2] to another uniform random value in same interval.
scalingTime = time.Duration(totalPods/(3*throughput)) * time.Second
framework.Logf("Starting to scale %v objects second time...", itArg.kind)
scaleAllResources(configs, scalingTime, testPhaseDurations.StartPhase(400, "scaling second time"))
By("============================================================================")