OCPBUGS-15544: Add adminpolicybasedexternalroutes rights for ovnkube-node.#1867
Conversation
|
Skipping CI for Draft Pull Request. |
|
WIP: waiting for test results to check ovnkube-node logs, every other job without this fix has mentioned errors, e.g. https://gcsweb-ci.apps.ci.l2s4.p1.openshiftapps.com/gcs/origin-ci-test/pr-logs/pull/openshift_cluster-network-operator/1866/pull-ci-openshift-cluster-network-operator-master-e2e-gcp-ovn/1676458790074978304/artifacts/e2e-gcp-ovn/gather-extra/artifacts/pods/openshift-ovn-kubernetes_ovnkube-node-jqwlf_ovnkube-node.log |
|
/retest-required |
|
looks like this PR solves memory leak problem: you can pick one job from this PR e.g. https://prow.ci.openshift.org/view/gs/origin-ci-test/pr-logs/pull/openshift_cluster-network-operator/1867/pull-ci-openshift-cluster-network-operator-master-e2e-azure-ovn/1676661712515764224 and same job for another pr e.g. https://prow.ci.openshift.org/view/gs/origin-ci-test/pr-logs/pull/openshift_cluster-network-operator/1866/pull-ci-openshift-cluster-network-operator-master-e2e-azure-ovn/1676458790066589696 |
|
@npinaeva: This pull request references Jira Issue OCPBUGS-15544, which is invalid:
Comment The bug has been updated to refer to the pull request using the external bug tracker. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
|
/lgtm |
|
/jira refresh |
|
@npinaeva: This pull request references Jira Issue OCPBUGS-15544, which is valid. The bug has been moved to the POST state. 3 validation(s) were run on this bug
Requesting review from QA contact: DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
|
perhaps add it to |
2beadee to
0a43705
Compare
|
nice catch! I didn't even know that file exists :) |
|
Would this be a good time to add |
I guess we can do this after. PTAL @jordigilh /lgtm |
Yeah, would make sense to have only access to the status in the controller rather than the whole object. I will raise another PR. Thanks for the suggestion 😄 |
I guess what he means is to change this line to point to the status |
Yeah, you need to take care also in the code that you are just updating the status (edit: using the But it's just the same with egressips, egresssvcs, ... so we can also do this as a general effort. edit: and we still need the read permission on the overall resource |
Yep, it is already using |
|
@npinaeva awesome that it fixes the memleak thing... can you put the details of why this happens into the commit message too though? |
It shouldn't change the object, therefore get, list, watch should be enough. The memory leak was hapenning because ovn-k doesn't fail when a controller can't watch resources, but it references other objects like pods and namespaces that have their own workqueues and can't be handled because the main watcher didn't start, thus the queues are overflowed. Add adminpolicybasedexternalroutes to cluster-reader permissions together with the other k8s.ovn.org resources Signed-off-by: Nadia Pinaeva <npinaeva@redhat.com>
0a43705 to
98c7509
Compare
|
added some memory leak details to the comment |
|
/lgtm |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: jcaamano, jordigilh, npinaeva The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/retest-required |
|
/override ci/prow/e2e-metal-ipi-ovn-ipv6 /override ci/prow/e2e-vsphere-ovn-windows /override ci/prow/e2e-gcp-ovn |
|
@dcbw: Overrode contexts on behalf of dcbw: ci/prow/e2e-gcp-ovn, ci/prow/e2e-metal-ipi-ovn-ipv6, ci/prow/e2e-vsphere-ovn-windows DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
|
@npinaeva: The following tests failed, say
Full PR test history. Your PR dashboard. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
@npinaeva: Jira Issue OCPBUGS-15544: All pull requests linked via external trackers have merged: Jira Issue OCPBUGS-15544 has been moved to the MODIFIED state. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |


It shouldn't change the object, therefore get, list, watch should be enough.
Initally added here #1765
Without this change non-OVN-IC deployments have errors like