c12s-kubespray/docs/calico.md
Cristian Calin ec0c0d4a28
Calico enable support for eBPF (#7618)
* Calico: align manifests with upstream

* allow enabling typha prometheus metrics

* Calico: enable eBPF support

* manage the kubernetes-services-endpoint configmap

* Calico: document the use of eBPF dataplane

* Calico: improve checks before deployment

* enforce disabling kube-proxy when using eBPF dataplane
* ensure calico_version is supported
2021-06-07 04:58:39 -07:00

11 KiB

Calico

Check if the calico-node container is running

docker ps | grep calico

The calicoctl.sh is wrap script with configured access credentials for command calicoctl allows to check the status of the network workloads.

  • Check the status of Calico nodes
calicoctl.sh node status
  • Show the configured network subnet for containers
calicoctl.sh get ippool -o wide
  • Show the workloads (ip addresses of containers and their location)
calicoctl.sh get workloadEndpoint -o wide

and

calicoctl.sh get hostEndpoint -o wide

Configuration

Optional : Define datastore type

The default datastore, Kubernetes API datastore is recommended for on-premises deployments, and supports only Kubernetes workloads; etcd is the best datastore for hybrid deployments.

Allowed values are kdd (default) and etcd.

Note: using kdd and more than 50 nodes, consider using the typha daemon to provide scaling.

To re-define you need to edit the inventory and add a group variable calico_datastore

calico_datastore: kdd

Optional : Define network backend

In some cases you may want to define Calico network backend. Allowed values are bird, vxlan or none. Bird is a default value.

To re-define you need to edit the inventory and add a group variable calico_network_backend

calico_network_backend: none

Optional : Define the default pool CIDRs

By default, kube_pods_subnet is used as the IP range CIDR for the default IP Pool, and kube_pods_subnet_ipv6 for IPv6. In some cases you may want to add several pools and not have them considered by Kubernetes as external (which means that they must be within or equal to the range defined in kube_pods_subnet and kube_pods_subnet_ipv6 ), it starts with the default IP Pools of which IP range CIDRs can by defined in group_vars (k8s_cluster/k8s-net-calico.yml):

calico_pool_cidr: 10.233.64.0/20
calico_pool_cidr_ipv6: fd85:ee78:d8a6:8607::1:0000/112

Optional : BGP Peering with border routers

In some cases you may want to route the pods subnet and so NAT is not needed on the nodes. For instance if you have a cluster spread on different locations and you want your pods to talk each other no matter where they are located. The following variables need to be set: peer_with_router to enable the peering with the datacenter's border router (default value: false). you'll need to edit the inventory and add a hostvar local_as by node.

node1 ansible_ssh_host=95.54.0.12 local_as=xxxxxx

Optional : Defining BGP peers

Peers can be defined using the peers variable (see docs/calico_peer_example examples). In order to define global peers, the peers variable can be defined in group_vars with the "scope" attribute of each global peer set to "global". In order to define peers on a per node basis, the peers variable must be defined in hostvars. NB: Ansible's hash_behaviour is by default set to "replace", thus defining both global and per node peers would end up with having only per node peers. If having both global and per node peers defined was meant to happen, global peers would have to be defined in hostvars for each host (as well as per node peers)

Since calico 3.4, Calico supports advertising Kubernetes service cluster IPs over BGP, just as it advertises pod IPs. This can be enabled by setting the following variable as follow in group_vars (k8s_cluster/k8s-net-calico.yml)

calico_advertise_cluster_ips: true

Since calico 3.10, Calico supports advertising Kubernetes service ExternalIPs over BGP in addition to cluster IPs advertising. This can be enabled by setting the following variable in group_vars (k8s_cluster/k8s-net-calico.yml)

calico_advertise_service_external_ips:
- x.x.x.x/24
- y.y.y.y/32

Optional : Define global AS number

Optional parameter global_as_num defines Calico global AS number (/calico/bgp/v1/global/as_num etcd key). It defaults to "64512".

Optional : BGP Peering with route reflectors

At large scale you may want to disable full node-to-node mesh in order to optimize your BGP topology and improve calico-node containers' start times.

To do so you can deploy BGP route reflectors and peer calico-node with them as recommended here:

You need to edit your inventory and add:

  • calico_rr group with nodes in it. calico_rr can be combined with kube_node and/or kube_control_plane. calico_rr group also must be a child group of k8s_cluster group.
  • cluster_id by route reflector node/group (see details here)

Here's an example of Kubespray inventory with standalone route reflectors:

[all]
rr0 ansible_ssh_host=10.210.1.10 ip=10.210.1.10
rr1 ansible_ssh_host=10.210.1.11 ip=10.210.1.11
node2 ansible_ssh_host=10.210.1.12 ip=10.210.1.12
node3 ansible_ssh_host=10.210.1.13 ip=10.210.1.13
node4 ansible_ssh_host=10.210.1.14 ip=10.210.1.14
node5 ansible_ssh_host=10.210.1.15 ip=10.210.1.15

[kube_control_plane]
node2
node3

[etcd]
node2
node3
node4

[kube_node]
node2
node3
node4
node5

[k8s_cluster:children]
kube_node
kube_control_plane
calico_rr

[calico_rr]
rr0
rr1

[rack0]
rr0
rr1
node2
node3
node4
node5

[rack0:vars]
cluster_id="1.0.0.1"

The inventory above will deploy the following topology assuming that calico's global_as_num is set to 65400:

Image

Optional : Define default endpoint to host action

By default Calico blocks traffic from endpoints to the host itself by using an iptables DROP action. When using it in kubernetes the action has to be changed to RETURN (default in kubespray) or ACCEPT (see https://github.com/projectcalico/felix/issues/660 and https://github.com/projectcalico/calicoctl/issues/1389). Otherwise all network packets from pods (with hostNetwork=False) to services endpoints (with hostNetwork=True) within the same node are dropped.

To re-define default action please set the following variable in your inventory:

calico_endpoint_to_host_action: "ACCEPT"

Optional : Define address on which Felix will respond to health requests

Since Calico 3.2.0, HealthCheck default behavior changed from listening on all interfaces to just listening on localhost.

To re-define health host please set the following variable in your inventory:

calico_healthhost: "0.0.0.0"

Config encapsulation for cross server traffic

Calico supports two types of encapsulation: VXLAN and IP in IP. VXLAN is supported in some environments where IP in IP is not (for example, Azure).

IP in IP and VXLAN is mutualy exclusive modes.

Configure Ip in Ip mode. Possible values is Always, CrossSubnet, Never.

calico_ipip_mode: 'Always'

Configure VXLAN mode. Possible values is Always, CrossSubnet, Never.

calico_vxlan_mode: 'Never'

If you use VXLAN mode, BGP networking is not required. You can disable BGP to reduce the moving parts in your cluster by calico_network_backend: vxlan

Cloud providers configuration

Please refer to the official documentation, for example GCE configuration requires a security rule for calico ip-ip tunnels. Note, calico is always configured with calico_ipip_mode: Always if the cloud provider was defined.

Optional : Ignore kernel's RPF check setting

By default the felix agent(calico-node) will abort if the Kernel RPF setting is not 'strict'. If you want Calico to ignore the Kernel setting:

calico_node_ignorelooserpf: true

Note that in OpenStack you must allow ipip traffic in your security groups, otherwise you will experience timeouts. To do this you must add a rule which allows it, for example:

Optional : Felix configuration via extraenvs of calico node

Possible environment variable parameters for configuring Felix

calico_node_extra_envs:
    FELIX_DEVICEROUTESOURCEADDRESS: 172.17.0.1
neutron  security-group-rule-create  --protocol 4  --direction egress  k8s-a0tp4t
neutron  security-group-rule-create  --protocol 4  --direction igress  k8s-a0tp4t

Optional : Use Calico CNI host-local IPAM plugin

Calico currently supports two types of CNI IPAM plugins, host-local and calico-ipam (default).

To allow Calico to determine the subnet to use from the Kubernetes API based on the Node.podCIDR field, enable the following setting.

calico_ipam_host_local: true

Refer to Project Calico section Using host-local IPAM for further information.

eBPF Support

Calico supports eBPF for its data plane see an introduction to the Calico eBPF Dataplane for further information.

Note that it is advisable to always use the latest version of Calico when using the eBPF dataplane.

Enabling eBPF support

To enable the eBPF dataplane support ensure you add the following to your inventory. Note that the kube-proxy is incompatible with running Calico in eBPF mode and the kube-proxy should be removed from the system.

calico_bpf_enabled: true
kube_proxy_remove: true

Cleaning up after kube-proxy

Calico node cannot clean up after kube-proxy has run in ipvs mode. If you are converting an existing cluster to eBPF you will need to ensure the kube-proxy DaemonSet is deleted and that ipvs rules are cleaned.

To check that kube-proxy was running in ipvs mode:

# ipvsadm -l

To clean up any ipvs leftovers:

# ipvsadm -C

Calico access to the kube-api

Calico node, typha and kube-controllers need to be able to talk to the kubernetes API. Please reference the Enabling eBPF Calico Docs for guidelines on how to do this.

Kubespray sets up the kubernetes-services-endpoint configmap based on the contents of the loadbalancer_apiserver inventory variable documented in HA Mode.

If no external loadbalancer is used, Calico eBPF can also use the localhost loadbalancer option. In this case Calico Automatic Host Endpoints need to be enabled to allow services like coredns and metrics-server to communicate with the kubernetes host endpoint. See this blog post on enabling automatic host endpoints.

loadbalancer_apiserver_localhost: true
use_localhost_as_kubeapi_loadbalancer: true

Tunneled versus Direct Server Return

By default Calico usese Tunneled service mode but it can use direct server return (DSR) in order to optimize the return path for a service.

To configure DSR:

calico_bpf_service_mode: "DSR"

eBPF Logging and Troubleshooting

In order to enable Calico eBPF mode logging:

calico_bpf_log_level: "Debug"

To view the logs you need to use the tc command to read the kernel trace buffer:

tc exec bpf debug

Please see Calico eBPF troubleshooting guide.