Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Revise known-issues in the Egress document #6775

Merged
merged 1 commit into from
Oct 29, 2024
Merged

Revise known-issues in the Egress document #6775

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 14 additions & 12 deletions docs/egress.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -466,13 +466,15 @@ in a cluster using `kube-proxy` IPVS. The issue was fixed in Antrea v1.7.0.

## Known issues

To support `EgressSeparateSubnet` feature, VLAN sub-interfaces will be created by
Antrea Agents, the `rp_filter` of VLAN sub-interfaces should be 2, which enables loose
mode filtering. In a vanilla Kubernetes cluster, Antrea Agents will set the `rp_filter`
to 2 automatically without user intervention. However, it has been observed that
`rp_filter` update by Antrea has no effect on OpenShift clusters due to [a known issue](https://github.com/antrea-io/antrea/issues/6546).
A workaround is to leverage OpenShift Node Tuning Operator to update the `rp_filter`
for `all` interface on all Egress Nodes:
To support the `EgressSeparateSubnet` feature, VLAN sub-interfaces will be
created by Antrea Agent on a Node, and the `rp_filter` setting of the VLAN
sub-interfaces should be set to `2`, which configures loose reverse path
filtering. In a vanilla Kubernetes cluster, Antrea Agent will set `rp_filter` to
`2` automatically without user intervention. However, it has been observed that
the `rp_filter` update by Antrea takes no effect on an OpenShift cluster due to
[a known issue](https://github.com/antrea-io/antrea/issues/6546). A workaround
for this issue is to leverage OpenShift Node Tuning Operator to update
`rp_filter` for all interfaces on all Egress Nodes:

```yaml
apiVersion: tuned.openshift.io/v1
Expand All @@ -496,8 +498,8 @@ spec:
profile: openshift-antrea
```

After you apply above `Tuned` CR named `antrea` in a given OpenShift cluster, the Node
Tuning Operator will watch the CR and update `net.ipv4.conf.all.rp_filter` to 2 for all
matched Nodes (e.g. all Nodes with a label `network-role=egress-gateway`). Please refer
to the OpenShift official document about [Using the Node Tuning Operator](https://docs.openshift.com/container-platform/4.16/scalability_and_performance/using-node-tuning-operator.html)
for more details of `Tuned` CR.
After you apply the above `Tuned` CR named `antrea` in an OpenShift cluster, the
Node Tuning Operator will reconcile the CR and update
`net.ipv4.conf.all.rp_filter` to `2` for all the matched Nodes (e.g. all Nodes
with label `network-role=egress-gateway`). Please refer to the OpenShift
document about [Using the Node Tuning Operator](https://docs.openshift.com/container-platform/4.16/scalability_and_performance/using-node-tuning-operator.html).
Loading