说明
操作
启用标志
yum -y install net-tools.x86_64
ifconfig | grep vxlan
cilium_vxlan: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
修改相关配置
ks get cm cilium-config -oyaml | grep -E 'tunnel|ipv4-native-routing-cidr|auto-direct-node-routes|bpf-lb-mode'
auto-direct-node-routes: "true"
bpf-lb-mode: dsr
ipv4-native-routing-cidr: 245.0.0.0/8
tunnel: disabled
重启所有cilium相关组件
ks delete $(ks get po -oname | grep cilium)
禁用标志
ifconfig | grep vxlan
端口修改
ks edit cm cilium-config
/usr/bin/cilium-agent --help | grep tun
-t, --tunnel string Tunnel mode {vxlan, geneve, disabled} (default "vxlan" for the "veth" datapath mode)
--tunnel-port int Tunnel port (default 8472 for "vxlan" and 6081 for "geneve")
/usr/bin/cilium-agent --help
Run the cilium agent
Usage:
cilium-agent [flags]
Flags:
--agent-health-port int TCP port for agent health status API (default 9879)
--agent-labels strings Additional labels to identify this agent
--agent-not-ready-taint-key string Key of the taint indicating that Cilium is not ready on the node (default "node.cilium.io/agent-not-ready")
--allocator-list-timeout duration Timeout for listing allocator state before exiting (default 3m0s)
--allow-icmp-frag-needed Allow ICMP Fragmentation Needed type packets for purposes like TCP Path MTU. (default true)
--allow-localhost string Policy when to allow local stack to reach local endpoints { auto | always | policy } (default "auto")
--annotate-k8s-node Annotate Kubernetes node
--api-rate-limit map API rate limiting configuration (example: --rate-limit endpoint-create=rate-limit:10/m,rate-burst:2)
--arping-refresh-period duration Period for remote node ARP entry refresh (set 0 to disable) (default 30s)
--auto-create-cilium-node-resource Automatically create CiliumNode resource for own node on startup (default true)
--auto-direct-node-routes Enable automatic L2 routing between nodes
--bgp-announce-lb-ip Announces service IPs of type LoadBalancer via BGP
--bgp-announce-pod-cidr Announces the node's pod CIDR via BGP
--bgp-config-path string Path to file containing the BGP configuration (default "/var/lib/cilium/bgp/config.yaml")
--bpf-ct-global-any-max int Maximum number of entries in non-TCP CT table (default 262144)
--bpf-ct-global-tcp-max int Maximum number of entries in TCP CT table (default 524288)
--bpf-ct-timeout-regular-any duration Timeout for entries in non-TCP CT table (default 1m0s)
--bpf-ct-timeout-regular-tcp duration Timeout for established entries in TCP CT table (default 6h0m0s)
--bpf-ct-timeout-regular-tcp-fin duration Teardown timeout for entries in TCP CT table (default 10s)
--bpf-ct-timeout-regular-tcp-syn duration Establishment timeout for entries in TCP CT table (default 1m0s)
--bpf-ct-timeout-service-any duration Timeout for service entries in non-TCP CT table (default 1m0s)
--bpf-ct-timeout-service-tcp duration Timeout for established service entries in TCP CT table (default 6h0m0s)
--bpf-ct-timeout-service-tcp-grace duration Timeout for graceful shutdown of service entries in TCP CT table (default 1m0s)
--bpf-fragments-map-max int Maximum number of entries in fragments tracking map (default 8192)
--bpf-lb-acceleration string BPF load balancing acceleration via XDP ("native", "disabled") (default "disabled")
--bpf-lb-algorithm string BPF load balancing algorithm ("random", "maglev") (default "random")
--bpf-lb-dev-ip-addr-inherit string Device name which IP addr is inherited by devices running LB BPF program (--devices)
--bpf-lb-dsr-dispatch string BPF load balancing DSR dispatch method ("opt", "ipip") (default "opt")
--bpf-lb-dsr-l4-xlate string BPF load balancing DSR L4 DNAT method for IPIP ("frontend", "backend") (default "frontend")
--bpf-lb-external-clusterip Enable external access to ClusterIP services (default false)
--bpf-lb-maglev-hash-seed string Maglev cluster-wide hash seed (base64 encoded) (default "JLfvgnHc2kaSUFaI")
--bpf-lb-maglev-table-size uint Maglev per service backend table size (parameter M) (default 16381)
--bpf-lb-map-max int Maximum number of entries in Cilium BPF lbmap (default 65536)
--bpf-lb-mode string BPF load balancing mode ("snat", "dsr", "hybrid") (default "snat")
--bpf-lb-rss-ipv4-src-cidr string BPF load balancing RSS outer source IPv4 CIDR prefix for IPIP
--bpf-lb-rss-ipv6-src-cidr string BPF load balancing RSS outer source IPv6 CIDR prefix for IPIP
--bpf-lb-sock Enable socket-based LB for E/W traffic
--bpf-lb-sock-hostns-only Skip socket LB for services when inside a pod namespace, in favor of service LB at the pod interface. Socket LB is still used when in the host namespace. Required by service mesh (e.g., Istio, Linkerd).
--bpf-map-dynamic-size-ratio float Ratio (0.0-1.0) of total system memory to use for dynamic sizing of CT, NAT