You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The job running the Containernet tests on an Ubuntu 20.04 host in Github Actions fails regularly. Cause is always that assertions like
self.assertTrue(self.net.pingAll() <=0.0)
fail, because the ping returns a packet drop of 100.0. However they don't always fail, just enough to have to rerun jobs multiple time to make the tests pass.
This does not occur on the Ubuntu 18.04 host, nor could I reproduce it locally so far.
The text was updated successfully, but these errors were encountered:
Seems like this issue is affecting all of my PRs (#238, #239, #240, #241) :(
What's interesting is that the set of failed tests for each GH action in Ubuntu 20.04 runner seems to vary every run. I think this flakiness is caused by the environment the GH runners are in.
Btw Ansible script containernet is using seems to be unable to detect docker preinstalled in GH runners, and attempts to install docker-ce alongside (which it succeeds). My theory (which is more like a random thought) is that the flakiness is coming from two different docker engines running simultaneously, but I haven't tested this theory (and it still doesn't explain why error is not occurring in 18.04 runner). (ignore this random thought - checking through APT showed that preinstalled Docker on GH action runner is automatically removed anyway)
While checking through upstream Mininet repository I've found this in their test action runner - maybe this has some connection with Containernet's issue?
The job running the Containernet tests on an Ubuntu 20.04 host in Github Actions fails regularly. Cause is always that assertions like
fail, because the ping returns a packet drop of
100.0
. However they don't always fail, just enough to have to rerun jobs multiple time to make the tests pass.This does not occur on the Ubuntu 18.04 host, nor could I reproduce it locally so far.
The text was updated successfully, but these errors were encountered: