SDN-3993: [DownstreamMerge] 06 jul 23 #1742
Conversation
Signed-off-by: Girish Moodalbail <[email protected]>
…ing/adding a pod from routeInfo and add unit tests Signed-off-by: jordigilh <[email protected]>
…() function Signed-off-by: jordigilh <[email protected]>
This commit adds the notion of zone to the EgressQoS controller so that it stops processing remote nodes and pods. Signed-off-by: Surya Seetharaman <[email protected]>
… a pod is deleted Signed-off-by: jordigilh <[email protected]>
rename incorrectly named test files - suit_test.go
This commit renames the subsystem from `master` to `ovnkube-controller` Signed-off-by: Surya Seetharaman <[email protected]>
…air_end Wait for APB controller's clean external gateway ECMP routes before resuming default net controller startup
EQoS: Consider only local nodes and pods
…_deleting_pod_in_egressgw Add check for overlapping external gateway IPs in egressgw when deleting a pod
Changes Metrics for IC
…tion Signed-off-by: jordigilh <[email protected]>
…n_master_controller APB: Change hybrid route and SNAT calls in unit tests to call the APB functions instead of the ones exposed for the legacy egressgw
When node from a remote zone is updated, we only perform the actual update when necessary. This commit improved the logic for doing the remote update in cases where the subnets of the remote node change. That is particularly needed when node changes from ipv4 to dual stack (ipv4 + ipv6) Reported-at: https://issues.redhat.com/browse/SDN-3993 Signed-off-by: Flavio Fernandes <[email protected]>
There are specific reasons for when we want to use the structured logs a.k.a InfoS versus when we want to use the formatted logs a.k.a Infof. Most of these logs are using the Infof format while using the InfoS function unfortunately. This causes logs like: I0629 20:41:28.798703 114685 master_controller.go:200] "Repairing Admin Policy Based External Route Services" I0629 20:41:28.798985 114685 repair.go:30] "Syncing exgw routes took %v" 265.415µs="(MISSING)" I0629 20:41:28.799084 114685 external_controller_pod.go:42] "Processing gateway pod %s/%s with matching policies %+v" openshift-dns="node-resolver-cmhvr" []="(MISSING)" I0629 20:41:28.799116 114685 external_controller_pod.go:42] "Processing gateway pod %s/%s with matching policies %+v" openshift-etcd="installer-7-ip-10-0-156-153.us-east-2.compute.internal" []="(MISSING)" I0629 20:41:28.799137 114685 external_controller_pod.go:42] "Processing gateway pod %s/%s with matching policies %+v" openshift-etcd="revision-pruner-7-ip-10-0-144-208.us-east-2.compute.internal" []="(MISSING)" I0629 20:41:28.799152 114685 external_controller_pod.go:42] "Processing gateway pod %s/%s with matching policies %+v" openshift-kube-apiserver="installer-8-ip-10-0-224-168.us-east-2.compute.internal" []="(MISSING)" I0629 20:41:28.799170 114685 external_controller_pod.go:42] "Processing gateway pod %s/%s with matching policies %+v" openshift-oauth-apiserver="apiserver-669859c5fb-7m7q6" []="(MISSING)" Let's change it back to Infof in places where the key=value format is not used. Signed-off-by: Surya Seetharaman <[email protected]>
Signed-off-by: jordigilh <[email protected]>
Disable UTs for APB Temporarily till the feature is stabilized. Reason: UTs are very flaky. Each PR is needed a minimum of 3 close/open combinations and luck for CI to pass to even run e2e's. At this stage where other features are trying to get in before the deadline this process is painful. Signed-off-by: Surya Seetharaman <[email protected]>
At ovnkube-controller start up, it may attempt to start adding resources for a node that does not yet have its annotation, but should be considered local. This is especially the case in OpenShift when going from phase1 -> phase 2 upgrade and ovnkube-node on the node will annotate its zone id. In this case ovnkube-controller will start too early and program NBDB resources as remote for the node. Then when the node is annotated, it will program NBDB resources as local. Since we cannot be sure that all remote feature configuration in NBDB is removed when going from remote -> local we need to avoid this case. It also makes no sense to program NBDB when there are no nodes in the zone. This commit adds a wait to make sure at least one node in the cluster is in the ovnkube-controller's managed zone. Thanks @numansiddique for the idea. Signed-off-by: Tim Rozet <[email protected]>
ovn-ic: node update missing static routes
|
@flavio-fernandes: This pull request references SDN-3993 which is a valid jira issue. In 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. |
openshift-ci-robot
added
the
jira/valid-reference
|
/assign @trozet @jordigilh @npinaeva @tssurya |
flavio-fernandes
force-pushed
the
merge-06-jul-23
branch
from
2f0be23
to
f3f14bc
Compare
flavio-fernandes
force-pushed
the
merge-06-jul-23
branch
from
f3f14bc
to
2f8e12e
Compare
|
/lgtm |
|
/retest |
|
/approve |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: dcbw, flavio-fernandes, npinaeva The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
openshift-ci
bot
added
the
approved
|
/override ci/prow/e2e-metal-ipi-ovn-dualstack |
|
/override ci/prow/e2e-gcp-ovn |
|
@dcbw: Overrode contexts on behalf of dcbw: ci/prow/e2e-metal-ipi-ovn-dualstack In 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. |
|
@dcbw: Overrode contexts on behalf of dcbw: ci/prow/e2e-gcp-ovn In 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. |
|
@flavio-fernandes: The following tests failed, say
Full PR test history. Your PR dashboard. 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. I understand the commands that are listed here. |
no conflicts.
based off of #1729