Robustness flake due to gofailpoint deactivation timeout

[serathius]

Which Github Action / Prow Jobs are flaking?

Test Robustness nightly

Which tests are flaking?

TestRobustnessExploratory/Kubernetes/HighTraffic/ClusterOfSize3

Github Action / Prow Job link

https://github.com/etcd-io/etcd/actions/runs/8314445428/job/22751625109

Reason for failure (if possible)

2024-03-17T11:27:29.9104803Z logger.go:130: 2024-03-17T11:27:29.907Z INFO goFailpoint deactivate failed {"failpoint": "raftBeforeSave=sleep(1s)", "error": "Delete "http://127.0.0.1:12381/raftBeforeSave\": context deadline exceeded (Client.Timeout exceeded while awaiting headers)"}

Anything else we need to know?

No response

[jamshidi799]

Hi, I am interested in this issue and in general robustness issues but I don't know exactly what should I do. In addition to that, most robustness tests fail in my local environment because the qps is lower than what expected

[serathius]

Expect this was caused by etcd-io/gofail#64

[serathius]

@jamshidi799 feel free to contact me on Slack (https://github.com/etcd-io/etcd#contact) if you need help with onboarding robustness test. I'm planning to create a dedicated community meeting dedicated to learning robustness test and triaging robustness test failures.

[henrybear327]

Expect this was caused by etcd-io/gofail#64

I am on it! :)

[serathius]

This was resolved. The root cause was fact that older etcd versions were using outdated gofail version.