For test_network_policy I recommend the following changes: 1) Add --request-timeout=1m to the `kubectl exec` calls. 2) Use ssh instead of unit.run to prevent it from acquiring the machine lock (which can prevent the debug action from running).
I have also requested changes to juju-crashdump that would help us diagnose future failures in more detail: 1) https://github.com/juju/juju-crashdump/issues/81 2) https://github.com/juju/juju-crashdump/issues/82
For test_network_policy I recommend the following changes: timeout= 1m to the `kubectl exec` calls.
1) Add --request-
2) Use ssh instead of unit.run to prevent it from acquiring the machine lock (which can prevent the debug action from running).
I have also requested changes to juju-crashdump that would help us diagnose future failures in more detail: /github. com/juju/ juju-crashdump/ issues/ 81 /github. com/juju/ juju-crashdump/ issues/ 82
1) https:/
2) https:/