Fix flakiness of the TestAttachReconnectReadLogs test
What does this MR do?
Fix the flakiness of the TestAttachReconnectReadLogs
test
Why was this MR needed?
It brings reliability to the TestAttachReconnectReadLogs
test
What's the best way to test this MR?
How to test the MR
Run the following command :
go test -count 100 -run TestAttachReconnectReadLogs -v ./executors/kubernetes/
The tests must pass.
Note: The same test may passes on the master branch and it could be difficult to reproduce it.
Tests benchmarks
master
:
$ go test -count 100 -run TestAttachReconnectReadLogs -v ./executors/kubernetes/
ok gitlab.com/gitlab-org/gitlab-runner/executors/kubernetes 0.438s
flaky_test_attach_reconnect_read_logs
:
$ go test -count 100 -run TestAttachReconnectReadLogs -v ./executors/kubernetes/
ok gitlab.com/gitlab-org/gitlab-runner/executors/kubernetes 21.619s
That makes it clearer that the backoff period is now being respected.
What are the relevant issue numbers?
closes: #27924 (closed)
Edited by Romuald Atchadé