Gazelle has been experiencing a very high flake rate: 10 of the last 20 builds on master failed. Most of these are on Ubuntu, like this one.
(00:16:12) ERROR: /var/lib/buildkite-agent/.cache/bazel/_bazel_buildkite-agent/011561ce453ad491f744b6cf5241f06f/external/gazelle+/internal/go_repository.bzl:336:13: An error occurred during the fetch of repository 'gazelle++_go_deps+++go_tools+com_github_google_uuid':
--
| Traceback (most recent call last):
| File "/var/lib/buildkite-agent/.cache/bazel/_bazel_buildkite-agent/011561ce453ad491f744b6cf5241f06f/external/gazelle+/internal/go_repository.bzl", line 336, column 13, in _go_repository_impl
| fail("%s: %s" % (ctx.name, result.stderr))
| Error in fail: gazelle++_go_deps+++go_tools+com_github_google_uuid: fetch_repo: failed copying repo: lstat /var/lib/buildkite-agent/.cache/bazel/_bazel_buildkite-agent/011561ce453ad491f744b6cf5241f06f/external/gazelle++non_module_deps+bazel_gazelle_go_repository_cache/pkg/mod/github.com/google/uuid@v1.3.0/time.go: no such file or directory
| (00:16:12) ERROR: no such package '@@gazelle++_go_deps+++go_tools+com_github_google_uuid//': gazelle++_go_deps+++go_tools+com_github_google_uuid: fetch_repo: failed copying repo: lstat /var/lib/buildkite-agent/.cache/bazel/_bazel_buildkite-agent/011561ce453ad491f744b6cf5241f06f/external/gazelle++non_module_deps+bazel_gazelle_go_repository_cache/pkg/mod/github.com/google/uuid@v1.3.0/time.go: no such file or directory
|
I think this is a CI infrastructure problem: the repository rule directory is being deleted while the repository rule is running. This error comes from a part of the code using filepath.Walk, which reads each directory and lstat's each file. We'd get no such file or directory if something were racing to delete the directory while fetch_repo was running.
I've only seen this happen on Ubuntu jobs. Retrying the job usually doesn't fixing the problem, but pushing an empty commit often does.
Are any other projects seeing errors like this? Is there anything happening in the infrastructure that might cause this?
Gazelle has been experiencing a very high flake rate: 10 of the last 20 builds on
masterfailed. Most of these are on Ubuntu, like this one.I think this is a CI infrastructure problem: the repository rule directory is being deleted while the repository rule is running. This error comes from a part of the code using
filepath.Walk, which reads each directory andlstat's each file. We'd getno such file or directoryif something were racing to delete the directory whilefetch_repowas running.I've only seen this happen on Ubuntu jobs. Retrying the job usually doesn't fixing the problem, but pushing an empty commit often does.
Are any other projects seeing errors like this? Is there anything happening in the infrastructure that might cause this?