Files
weave/lb/admit/admit.go
T
levaliciousandClaude Opus 5.5 305ac4f3d3
images / images (push) Successful in 18m34s
lb: EgressService - a LoadBalancer Service's pods leave from its address
EgressService (reweave.levk.dev/v1alpha1, named after a LoadBalancer
Service of class reweave.levk.dev/node-address) makes every pod the
Service selects (by pod IP, never readiness) leave the cluster from the
Service's address, or not at all. spec.emit is explicit: Holder (the only
mode so far) tunnels members' packets to the node holding the address,
which emits them.

The holder (S1's: the one node with the address on an interface) is the
status' sole writer: host, address, and a 512-port block per node in
61000-65535 (above the default ephemeral range), stable across churn.
A member's node SNATs it to address:its-block before it leaves, so the
holder forwards it unchanged and routes replies back by port alone: ip
rules (to address, ipproto tcp/udp, dport block) ahead of the local
table, which moves from 0 to 100, since the address is local on the
holder. Members' packets travel both ways in reweave-eg, an ipip tunnel
between weave bridges (NPC drops new routed packets into weave), with
accept_local and loose rp_filter on that device only.

Members' new connections are classed at the tail of nat PREROUTING,
after kube-proxy's DNAT has ended the nat walk for Services: the pod
network, multicast and the nodes' addresses are not outside. Marked
packets follow a per-member rule to the holder's table (or to an
unreachable one), are SNATed into the tunnel, and a FORWARD guard drops
anything else a member sends outside (other protocols, other ways out).
The tunnel is accepted only from the nodes' bridges, a holder accepts
from it only the blocks' ports, and TCP MSS is clamped into it.

Fail closed throughout: until the holder has published the status,
without a block, with no single holder, or for a pod two EgressServices
select, a member's outside traffic is dropped. The CNI plugin asks
weave-npc to admit every pod (unix socket /run/weave-npc/egress.sock,
conflist "egressAdmission": true, now rewritten when it differs) after
IPAM and before the pod has an interface, so a member's first packet
already leaves from the address; no answer, no pod. weave-npc applies
nothing about egress from caches that are not synced (unless the API is
absent): a restart never strips members' rules.

reweave/test/egress/run.sh checks the design's assertions against a
live cluster; both end-to-end scripts now re-evaluate each check on
every retry (it was expanded once, before the loop).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UHXzGrsxfkGk8LmwrBbHNZ
2026-10-13 14:39:44 -04:00

89 lines
2.8 KiB
Go

// Package admit is the protocol between the weave CNI plugin and weave-npc's egress agent: at CNI ADD the
// plugin tells the agent which pod got which IP, and the agent answers once the pod's egress rules are in
// place, so an EgressService member's first packet already leaves from the Service's address (or not at all).
//
// It depends on the standard library only: the plugin binary must not grow a Kubernetes client.
package admit
import (
"bytes"
"context"
"encoding/json"
"fmt"
"io"
"net"
"net/http"
"strings"
"time"
)
// SocketPath is where the agent listens (a hostPath directory shared with the host, where the plugin runs).
const SocketPath = "/run/weave-npc/egress.sock"
// Request is one pod's admission: the IP the plugin just allocated to it.
type Request struct {
Namespace string `json:"namespace"`
Name string `json:"name"`
IP string `json:"ip"`
}
// Response says whether the pod is an EgressService member (its rules are installed either way when 200).
type Response struct {
Member bool `json:"member"`
}
// Timeout bounds one admission (the agent may have to fetch the pod from the API server).
const Timeout = 20 * time.Second
// PodFromArgs reads the pod's namespace and name from CNI_ARGS ("K8S_POD_NAMESPACE=..;K8S_POD_NAME=..").
func PodFromArgs(args string) (namespace, name string) {
for _, kv := range strings.Split(args, ";") {
k, v, ok := strings.Cut(kv, "=")
if !ok {
continue
}
switch k {
case "K8S_POD_NAMESPACE":
namespace = v
case "K8S_POD_NAME":
name = v
}
}
return namespace, name
}
// Admit asks the agent at socket to admit req, and returns only once the agent has installed the pod's rules.
func Admit(ctx context.Context, socket string, req Request) (Response, error) {
var resp Response
body, err := json.Marshal(req)
if err != nil {
return resp, err
}
client := &http.Client{Transport: &http.Transport{
DialContext: func(ctx context.Context, _, _ string) (net.Conn, error) {
var d net.Dialer
return d.DialContext(ctx, "unix", socket)
},
}}
ctx, cancel := context.WithTimeout(ctx, Timeout)
defer cancel()
hreq, err := http.NewRequestWithContext(ctx, http.MethodPost, "http://weave-npc/admit", bytes.NewReader(body))
if err != nil {
return resp, err
}
hreq.Header.Set("Content-Type", "application/json")
hresp, err := client.Do(hreq)
if err != nil {
return resp, fmt.Errorf("egress admission (weave-npc at %s): %v", socket, err)
}
defer hresp.Body.Close()
out, _ := io.ReadAll(io.LimitReader(hresp.Body, 1<<16))
if hresp.StatusCode != http.StatusOK {
return resp, fmt.Errorf("egress admission refused (%s): %s", hresp.Status, strings.TrimSpace(string(out)))
}
if err := json.Unmarshal(out, &resp); err != nil {
return resp, fmt.Errorf("egress admission: bad answer: %v", err)
}
return resp, nil
}