|
90 | 90 | | metricAgent.pod.labels | object | `{}` | Labels for the metricAgent pod. | |
91 | 91 | | metricAgent.pod.resources | object | `{"limits":{},"requests":{}}` | Resource requests and limits (https://kubernetes.io/docs/user-guide/compute-resources/) for the metricAgent pod. | |
92 | 92 | | nameOverride | string | `""` | Override the standard name used by Helm and associated to Kubernetes/Liqo resources. | |
93 | | -| networkConfig.mtu | int | `1340` | Set the MTU for the interfaces managed by liqo: vxlan, tunnel and veth interfaces. The value is used by the gateway and route operators. The default value is configured to ensure correct behavior regardless of the combination of the underlying environments (e.g., cloud providers). This guarantees improved compatibility at the cost of possible limited performance drops. | |
94 | 93 | | networkManager.config.additionalPools | list | `[]` | Set of additional network pools to perform the automatic address mapping in Liqo. Network pools are used to map a cluster network into another one in order to prevent conflicts. Default set of network pools is: [10.0.0.0/8, 192.168.0.0/16, 172.16.0.0/12] | |
95 | 94 | | networkManager.config.podCIDR | string | `""` | The subnet used by the pods in your cluster, in CIDR notation (e.g., 10.0.0.0/16). | |
96 | 95 | | networkManager.config.reservedSubnets | list | `[]` | List of IP subnets that do not have to be used by Liqo. Liqo can perform automatic IP address remapping when a remote cluster is peering with you, e.g., in case IP address spaces (e.g., PodCIDR) overlaps. In order to prevent IP conflicting between locally used private subnets in your infrastructure and private subnets belonging to remote clusters you need tell liqo the subnets used in your cluster. E.g if your cluster nodes belong to the 192.168.2.0/24 subnet, then you should add that subnet to the reservedSubnets. PodCIDR and serviceCIDR used in the local cluster are automatically added to the reserved list. | |
|
103 | 102 | | networkManager.pod.labels | object | `{}` | Labels for the networkManager pod. | |
104 | 103 | | networkManager.pod.resources | object | `{"limits":{},"requests":{}}` | Resource requests and limits (https://kubernetes.io/docs/user-guide/compute-resources/) for the networkManager pod. | |
105 | 104 | | networking.internal | bool | `true` | Use the default Liqo network manager. | |
| 105 | +| networking.iptables | object | `{"mode":"nf_tables"}` | Iptables configuration tuning. | |
| 106 | +| networking.iptables.mode | string | `"nf_tables"` | Select the iptables mode to use. Possible values are "legacy" and "nf_tables". | |
| 107 | +| networking.mtu | int | `1340` | Set the MTU for the interfaces managed by liqo: vxlan, tunnel and veth interfaces. The value is used by the gateway and route operators. The default value is configured to ensure correct behavior regardless of the combination of the underlying environments (e.g., cloud providers). This guarantees improved compatibility at the cost of possible limited performance drops. | |
106 | 108 | | networking.reflectIPs | bool | `true` | Reflect pod IPs and EnpointSlices to the remote clusters. | |
107 | 109 | | openshiftConfig.enable | bool | `false` | Enable/Disable the OpenShift support, enabling Openshift-specific resources, and setting the pod security contexts in a way that is compatible with Openshift. | |
108 | 110 | | openshiftConfig.virtualKubeletSCCs | list | `["anyuid"]` | Security context configurations granted to the virtual kubelet in the local cluster. The configuration of one or more SCCs for the virtual kubelet is not strictly required, and privileges can be reduced in production environments. Still, the default configuration (i.e., anyuid) is suggested to prevent problems (i.e., the virtual kubelet fails to add the appropriate labels) when attempting to offload pods not managed by higher-level abstractions (e.g., Deployments), and not associated with a properly privileged service account. Indeed, "anyuid" is the SCC automatically associated with pods created by cluster administrators. Any pod granted a more privileged SCC and not linked to an adequately privileged service account will fail to be offloaded. | |
|
0 commit comments