Files
Ed Santiago 761eca9e75 e2e tests: fix racy flakes
...mostly just test code that wasn't doing the required waits.

My first approach in the kube-play test was to add "--wait".
Bit mistake! The --wait flag, counterintuitively and counter to
documentation, actually destroys all pods+containers+everything
on exit. (Or tries -- see #17803). Since this violates POLA
and is undocumented, I include here a fix to the man page.

Despite my best intentions, I can't reasonably check every single
test for missing waits, especially in kube-play where failing
containers will get retried forever so we can't wait. We'll
just have to fix flakes as we see them.

Fixes: #17958
Fixes: #18071

Signed-off-by: Ed Santiago <santiago@redhat.com>
2023-04-05 13:51:59 -06:00
..
2022-02-09 18:17:52 +01:00
2023-04-05 13:51:59 -06:00
2022-07-13 15:27:03 +01:00
2022-09-15 14:35:06 -06:00
2020-05-21 17:16:10 -04:00