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
{{ message }}
This repository has been archived by the owner on Mar 17, 2023. It is now read-only.
Similar to #44, but when the gateway proxies may be running on only a subset of worker nodes.
# lives in the management cluster, in the same namespace as the Cluster it refsapiVersion: connectivity.tanzu.vmware.com/v1alpha1kind: GatewayDNSRecordmetadata:
name: strawberry-gatewaynamespace: foo-teamspec:
cluster: strawberrynamespace: projectcontourservice: envoyresolutionType: hostPort
The EndpointSlices would look like
# one of these on every managed workload cluster
kind: EndpointSlice
...
endpoints:
- addresses:
- "<IP of node A for strawberry cluster>"
- - "<IP of node B for strawberry cluster>"+ # no address for node B, if there isn't a gateway proxy pod running on that node
- "<IP of node C for strawberry cluster>"
Handle updates/deletes when node IPs change, or when pods get rescheduled to different nodes.
This may be more difficult than the DaemonSet one, since it requires joining Pod resources with Nodes to find node IPs hosting a pod.
But knowing this is coming might influence how we implement #44
The text was updated successfully, but these errors were encountered:
Similar to #44, but when the gateway proxies may be running on only a subset of worker nodes.
The EndpointSlices would look like
Handle updates/deletes when node IPs change, or when pods get rescheduled to different nodes.
This may be more difficult than the DaemonSet one, since it requires joining
Pod
resources withNodes
to find node IPs hosting a pod.But knowing this is coming might influence how we implement #44
The text was updated successfully, but these errors were encountered: