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
@pratyushprakash Sure! let's have a discussion on the design so that we are aligned.
Maybe a good start is a separate package that takes in a kubeconfig or uses active cluster context if not provided to run tests. this can live under /test-bdd/main_test.go and invoke the tests in individual files e.g. /test-bdd/podreaper.go
A pod which is stuck in Terminating state is reapable after N2 minutes (depending on --reap-after flag)
A pod which is stuck in Terminating state is not reapable after N1 minutes (depending on --reap-after flag)
If dry-run is enabled, reapable pods are not reaped
(--soft-reap=true) if containers are still running, pods are not reaped
How to reproduce a stuck pod for testing
Run some pod that is doing some infinite 'sleep' and issue a terminate command to it
It will then go into terminating state and stay in that state (with container running) until sleep is done. This means as long as --soft-reap is off, it will become reapable.
We can test that if --soft-reap is on, pod does not get deleted (testing the inverse of this will be difficult).
Once we have the code, we can add separate issue for making sure it runs nightly in travis-ci.
The above should be an easy start at this, WDYT @pratyushprakash ?
governor packages should have a end to end functional test.
we can have a travis cronjob that runs nightly and run this test
test should:
hostNetwork
and runip link set dev eth0 down
, node will become NotReady, let reaper kill itThe text was updated successfully, but these errors were encountered: