Build tests are flakey

[bobcatfish]

Expected Behavior

Our tests should only fail when there is a real error! Especially the "build" tests.

Actual Behavior

Sometimes the build tests fail like this (e.g. #1203 (comment)):

I0816 13:41:12.000] ==== RUNNING BUILD TESTS ====
I0816 13:41:12.000] =============================
I0816 13:41:12.008] ----------------------------------------------
I0816 13:41:12.009] ---- Checking links in the markdown files ----
I0816 13:41:12.009] ----------------------------------------------
I0816 13:42:23.513] level=error msg="Deadline exceeded: try increase it by passing --deadline option"

Steps to Reproduce the Problem

Not sure! I saw this happen locally (i.e. when running the release pipeline) to @dibyom as well - it seems like if you run the build tests enough this happens sometimes.

[vdemeester]

This is the linters (golangci-lint) that time outs.. this might happen when they are scheduled on a slow node (or loaded node). We could bump the timeout for those, or split them to a job on their own 🤓

[bobcatfish]

Saw this when trying to run automated releases too (#860).

[lint : build-step-git-source-51fc6fe0-ce88-11e9-b641-0a580a340481-extr] {"level":"info","ts":1567542061.106349,"logger":"fallback-logger","caller":"git/git.go:105","msg":"Successfully cloned https://github.com/tektoncd/pipeline.git @ master in path /workspace/src/github.com/tektoncd/pipeline"}

[lint : build-step-lint] level=info msg="[config_reader] Config search paths: [./ /workspace/src/github.com/tektoncd/pipeline /workspace/src/github.com/tektoncd /workspace/src/github.com /workspace/src /workspace /]"
[lint : build-step-lint] level=info msg="[config_reader] Used config file .golangci.yml"
[lint : build-step-lint] level=info msg="[lintersdb] Active 12 linters: [deadcode errcheck gofmt goimports gosimple govet ineffassign staticcheck structcheck typecheck unused varcheck]"
[lint : build-step-lint] level=info msg="[lintersdb] Optimized sublinters [staticcheck gosimple unused] into metalinter megacheck"
[lint : build-step-lint] level=info msg="[loader] Go packages loading at mode load types and syntax took 20.52306475s"
[lint : build-step-lint] level=info msg="[loader] SSA repr building timing: packages building 234.47889ms, total 14.029861259s"
[lint : build-step-lint] level=info msg="[runner] worker.1 took 1m5.703529801s with stages: govet: 1m5.703488009s"
[lint : build-step-lint] level=info msg="[runner] worker.4 took 1m8.849588545s with stages: gofmt: 1m8.849572453s"
[lint : build-step-lint] level=info msg="Memory: 222 samples, avg is 720.2MB, max is 3694.5MB"
[lint : build-step-lint] level=info msg="Execution took 1m46.749045981s"
[lint : build-step-lint] level=info msg="[runner] worker.2 took 1m17.3554252s with stages: goimports: 1m17.355393415s"
[lint : build-step-lint] level=info msg="[runner] worker.3 took 1m34.527370732s with stages: megacheck: 1m34.527323136s"
[lint : build-step-lint] level=info msg="[runner] Workers idle times: #1: 28.819682124s, #2: 17.162955784s, #4: 25.601083978s"
[lint : build-step-lint] level=info msg="[runner] processing took 55.532µs with stages: max_same_issues: 29.937µs, nolint: 19.732µs, cgo: 1.419µs, skip_dirs: 1.337µs, max_from_linter: 465ns, filename_unadjuster: 311ns, skip_files: 276ns, path_prettifier: 257ns, path_shortener: 249ns, identifier_marker: 247ns, source_code: 241ns, diff: 232ns, autogenerated_exclude: 222ns, uniq_by_line: 186ns, exclude: 182ns, max_per_file_from_linter: 156ns, exclude-rules: 83ns"
[lint : build-step-lint] level=info msg="File cache stats: 0 entries of total size 0B"
[lint : build-step-lint] level=error msg="Deadline exceeded: try increase it by passing --deadline option"

Would be frustrating to have the nightly build fail b/c of this!

[vdemeester]

@bobcatfish yeah we might need to increase the deadline (can be done in the configuration file)

[bobcatfish]

I think this happened again #1475

[bobcatfish]

@vdemeester pointed out this is 5 min in the cli repo

[vdemeester]

Seems like #1613 got merge, so we can safely close this one
/close

[tekton-robot]

@vdemeester: Closing this issue.

In response to this:

Seems like #1613 got merge, so we can safely close this one
/close

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.