kubevirt / hyperconverged-cluster-operator / 16877622514

76%

Build Type

push

github

Committed by web-flow

Commit Message

Always deploy Network Policies for operands (#3699)

* Bug fix: avoid nil object in slice

Fix a case where we create a slice with fixed size, but not not setting
all its values, causing nil dereference panic, when trying to use the
empty items.

Signed-off-by: Nahshon Unna-Tsameret <nunnatsa@redhat.com>

* Move the alert-manager network policy to be static

Remove the logic that creates the `hco-allow-egress-to-alert-manager`
network policy from the reconciler code, to the csv merger tool, to be
created as a static yaml file on CSV build time, instead of on runtime.

Signed-off-by: Nahshon Unna-Tsameret <nunnatsa@redhat.com>

* api-server-proxy network policy: add rules for api-server and DNS

remove the network-policy label from the api-server-proxy pods because
we should not use the operator network policies for opernds: pods that
are deployed by the operator. Removing the
`hco.kubevirt.io/allow-access-cluster-services`, and instead, add the
rules to allow egress to the DNS and and to the API-Server, to the
existing Network policy.

Also, produce the network policies for the console plugin and the
api-server pods anyway, even if OLM does not deploy network policies for
the operator.

Signed-off-by: Nahshon Unna-Tsameret <nunnatsa@redhat.com>

* Fix the console plugin e2e test

The test is trying to call the console plugin from within the hco pod.
This is now blocked by the new NetworkPolicy.

Changed the test to do the same from the OpenShift console pod.

Signed-off-by: Nahshon Unna-Tsameret <nunnatsa@redhat.com>

---------

Signed-off-by: Nahshon Unna-Tsameret <nunnatsa@redhat.com>

Run Details

59 of 59 new or added lines in 3 files covered. (100.0%)

3 existing lines in 1 file now uncovered.

7501 of 9770 relevant lines covered (76.78%)

1.86 hits per line