test-infra: Scheduler benchmark job missing some lines of stdout while printing results
I recently added a scheduler-benchmark job which basically runs make test-integration ... inside a dockerized environment. The job runs this script using the kubernetes_verify scenario. That script is quite similar to the test-dockerized.sh script which is used to run our normal integration tests - while this one is used to run benchmark integration tests.
When I locally run the make command, I get the desired output (with some lines of benchmark results in b/w):
$ make test-integration WHAT="./test/integration/scheduler_perf" KUBE_TEST_ARGS="-run='XXX' -bench=. -benchmem"
+++ [0123 20:46:51] Checking etcd is on PATH
/usr/local/google/home/shyamjvs/Downloads/etcd-v3.1.11-linux-amd64/etcd
+++ [0123 20:46:51] Starting etcd instance
etcd --advertise-client-urls http://127.0.0.1:2379 --data-dir /tmp/tmp.SzP8ZqrfTO --listen-client-urls http://127.0.0.1:2379 --debug > "/dev/null" 2>/dev/null
Waiting for etcd to come up.
+++ [0123 20:46:52] On try 2, etcd: : http://127.0.0.1:2379
{"action":"set","node":{"key":"/_test","value":"","modifiedIndex":4,"createdIndex":4}}
+++ [0123 20:46:52] Running integration test cases
Running tests for APIVersion: v1
+++ [0123 20:46:54] Running tests without code coverage
W0123 20:47:11.417751 90561 admission.go:61] AlwaysAdmit admission controller is deprecated. Please remove this controller from your configuration files and scripts.
I0123 20:47:11.418413 90561 services.go:35] Network range for service cluster IPs is unspecified. Defaulting to 10.0.0.0/24.
I0123 20:47:11.418442 90561 master.go:269] Node port range unspecified. Defaulting to 30000-32767.
...
...
BenchmarkScheduling/100Nodes/0Pods-12 500 2393175 ns/op 431003 B/op 4825 allocs/op
...
...
BenchmarkScheduling/100Nodes/1000Pods-12 500 2164436 ns/op 459811 B/op 5366 allocs/op
...
...
All those lines are also present in the CI job’s build-log.txt (e.g - https://storage.googleapis.com/kubernetes-jenkins/logs/ci-benchmark-scheduler-master/71/build-log.txt), except those lines starting with Benchmark which are basically results of the go benchmark.
cc @krzyzacy @kubernetes/test-infra-maintainers - Does someone know what’s happening with those lines?
About this issue
- Original URL
- State: closed
- Created 6 years ago
- Comments: 34 (34 by maintainers)
Commits related to this issue
- Merge pull request #58749 from shyamjvs/benchmark-script-fix Automatic merge from submit-queue (batch tested with PRs 58561, 58749). If you want to cherry-pick this change to another branch, please f... — committed to kubernetes/kubernetes by deleted user 6 years ago
- Merge pull request #58764 from shyamjvs/fix-benchmark-script Automatic merge from submit-queue (batch tested with PRs 58661, 58764, 58368, 58739, 58773). If you want to cherry-pick this change to ano... — committed to kubernetes/kubernetes by deleted user 6 years ago
- Merge pull request #59068 from shyamjvs/benchmarking-script-fix Automatic merge from submit-queue (batch tested with PRs 59106, 58985, 59068, 59120, 59126). If you want to cherry-pick this change to ... — committed to kubernetes/kubernetes by deleted user 6 years ago
Just checked that my fix works, and it indeed seems so - http://gcsweb.k8s.io/gcs/kubernetes-jenkins/logs/ci-benchmark-scheduler-master/297/ Thanks guys!
I’ve sent out the above PR to fix the non-existing file issue. Hope it helps.