tailscale/util/linuxfw/nftables_for_svcs_test.go
Irbe Krumina 096b090caf
cmd/containerboot,kube,util/linuxfw: configure kube egress proxies to route to 1+ tailnet targets (#13531)
* cmd/containerboot,kube,util/linuxfw: configure kube egress proxies to route to 1+ tailnet targets

This commit is first part of the work to allow running multiple
replicas of the Kubernetes operator egress proxies per tailnet service +
to allow exposing multiple tailnet services via each proxy replica.

This expands the existing iptables/nftables-based proxy configuration
mechanism.

A proxy can now be configured to route to one or more tailnet targets
via a (mounted) config file that, for each tailnet target, specifies:
- the target's tailnet IP or FQDN
- mappings of container ports to which cluster workloads will send traffic to
tailnet target ports where the traffic should be forwarded.

Example configfile contents:
{
  "some-svc": {"tailnetTarget":{"fqdn":"foo.tailnetxyz.ts.net","ports"{"tcp:4006:80":{"protocol":"tcp","matchPort":4006,"targetPort":80},"tcp:4007:443":{"protocol":"tcp","matchPort":4007,"targetPort":443}}}}
}

A proxy that is configured with this config file will configure firewall rules
to route cluster traffic to the tailnet targets. It will then watch the config file
for updates as well as monitor relevant netmap updates and reconfigure firewall
as needed.

This adds a bunch of new iptables/nftables functionality to make it easier to dynamically update
the firewall rules without needing to restart the proxy Pod as well as to make
it easier to debug/understand the rules:

- for iptables, each portmapping is a DNAT rule with a comment pointing
at the 'service',i.e:

-A PREROUTING ! -i tailscale0 -p tcp -m tcp --dport 4006 -m comment --comment "some-svc:tcp:4006 -> tcp:80" -j DNAT --to-destination 100.64.1.18:80
Additionally there is a SNAT rule for each tailnet target, to mask the source address.

- for nftables, a separate prerouting chain is created for each tailnet target
and all the portmapping rules are placed in that chain. This makes it easier
to look up rules and delete services when no longer needed.
(nftables allows hooking a custom chain to a prerouting hook, so no extra work
is needed to ensure that the rules in the service chains are evaluated).

The next steps will be to get the Kubernetes Operator to generate
the configfile and ensure it is mounted to the relevant proxy nodes.

Updates tailscale/tailscale#13406

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
2024-09-29 16:30:53 +01:00

157 lines
5.4 KiB
Go

// Copyright (c) Tailscale Inc & AUTHORS
// SPDX-License-Identifier: BSD-3-Clause
//go:build linux
package linuxfw
import (
"net/netip"
"testing"
"github.com/google/nftables"
)
// This test creates a temporary network namespace for the nftables rules being
// set up, so it needs to run in a privileged mode. Locally it needs to be run
// by root, else it will be silently skipped. In CI it runs in a privileged
// container.
func Test_nftablesRunner_EnsurePortMapRuleForSvc(t *testing.T) {
conn := newSysConn(t)
runner := newFakeNftablesRunnerWithConn(t, conn, true)
ipv4, ipv6 := netip.MustParseAddr("100.99.99.99"), netip.MustParseAddr("fd7a:115c:a1e0::701:b62a")
pmTCP := PortMap{MatchPort: 4003, TargetPort: 80, Protocol: "TCP"}
pmTCP1 := PortMap{MatchPort: 4004, TargetPort: 443, Protocol: "TCP"}
// Create a rule for service 'foo' to forward TCP traffic to IPv4 endpoint
runner.EnsurePortMapRuleForSvc("foo", "tailscale0", ipv4, pmTCP)
svcChains(t, 1, conn)
chainRuleCount(t, "foo", 1, conn, nftables.TableFamilyIPv4)
chainRule(t, "foo", ipv4, pmTCP, runner, nftables.TableFamilyIPv4)
// Create another rule for service 'foo' to forward TCP traffic to the
// same IPv4 endpoint, but to a different port.
runner.EnsurePortMapRuleForSvc("foo", "tailscale0", ipv4, pmTCP1)
svcChains(t, 1, conn)
chainRuleCount(t, "foo", 2, conn, nftables.TableFamilyIPv4)
chainRule(t, "foo", ipv4, pmTCP1, runner, nftables.TableFamilyIPv4)
// Create a rule for service 'foo' to forward TCP traffic to an IPv6 endpoint
runner.EnsurePortMapRuleForSvc("foo", "tailscale0", ipv6, pmTCP)
svcChains(t, 2, conn)
chainRuleCount(t, "foo", 1, conn, nftables.TableFamilyIPv6)
chainRule(t, "foo", ipv6, pmTCP, runner, nftables.TableFamilyIPv6)
// Create a rule for service 'bar' to forward TCP traffic to IPv4 endpoint
runner.EnsurePortMapRuleForSvc("bar", "tailscale0", ipv4, pmTCP)
svcChains(t, 3, conn)
chainRuleCount(t, "bar", 1, conn, nftables.TableFamilyIPv4)
chainRule(t, "bar", ipv4, pmTCP, runner, nftables.TableFamilyIPv4)
// Create a rule for service 'bar' to forward TCP traffic to an IPv6 endpoint
runner.EnsurePortMapRuleForSvc("bar", "tailscale0", ipv6, pmTCP)
svcChains(t, 4, conn)
chainRuleCount(t, "bar", 1, conn, nftables.TableFamilyIPv6)
chainRule(t, "bar", ipv6, pmTCP, runner, nftables.TableFamilyIPv6)
// Delete service bar
runner.DeleteSvc("bar", "tailscale0", []netip.Addr{ipv4, ipv6}, []PortMap{pmTCP})
svcChains(t, 2, conn)
// Delete a rule from service foo
runner.DeletePortMapRuleForSvc("foo", "tailscale0", ipv4, pmTCP)
svcChains(t, 2, conn)
chainRuleCount(t, "foo", 1, conn, nftables.TableFamilyIPv4)
// Delete service foo
runner.DeleteSvc("foo", "tailscale0", []netip.Addr{ipv4, ipv6}, []PortMap{pmTCP, pmTCP1})
svcChains(t, 0, conn)
}
// svcChains verifies that the expected number of chains exist (for either IP
// family) and that each of them is configured as NAT prerouting chain.
func svcChains(t *testing.T, wantCount int, conn *nftables.Conn) {
t.Helper()
chains, err := conn.ListChains()
if err != nil {
t.Fatalf("error listing chains: %v", err)
}
if len(chains) != wantCount {
t.Fatalf("wants %d chains, got %d", wantCount, len(chains))
}
for _, ch := range chains {
if *ch.Policy != nftables.ChainPolicyAccept {
t.Fatalf("chain %s has unexpected policy %v", ch.Name, *ch.Policy)
}
if ch.Type != nftables.ChainTypeNAT {
t.Fatalf("chain %s has unexpected type %v", ch.Name, ch.Type)
}
if *ch.Hooknum != *nftables.ChainHookPrerouting {
t.Fatalf("chain %s is attached to unexpected hook %v", ch.Name, ch.Hooknum)
}
if *ch.Priority != *nftables.ChainPriorityNATDest {
t.Fatalf("chain %s has unexpected priority %v", ch.Name, ch.Priority)
}
}
}
// chainRuleCount returns number of rules in a chain identified by service name and IP family.
func chainRuleCount(t *testing.T, svc string, count int, conn *nftables.Conn, fam nftables.TableFamily) {
t.Helper()
chains, err := conn.ListChainsOfTableFamily(fam)
if err != nil {
t.Fatalf("error listing chains: %v", err)
}
found := false
for _, ch := range chains {
if ch.Name == svc {
found = true
rules, err := conn.GetRules(ch.Table, ch)
if err != nil {
t.Fatalf("error getting rules: %v", err)
}
if len(rules) != count {
t.Fatalf("unexpected number of rules, wants %d got %d", count, len(rules))
}
break
}
}
if !found {
t.Fatalf("chain for service %s does not exist", svc)
}
}
// chainRule verifies that rule for the provided target IP and PortMap exists in
// a chain identified by service name and IP family.
func chainRule(t *testing.T, svc string, targetIP netip.Addr, pm PortMap, runner *nftablesRunner, fam nftables.TableFamily) {
t.Helper()
chains, err := runner.conn.ListChainsOfTableFamily(fam)
if err != nil {
t.Fatalf("error listing chains: %v", err)
}
var chain *nftables.Chain
for _, ch := range chains {
if ch.Name == svc {
chain = ch
break
}
}
if chain == nil {
t.Fatalf("chain for service %s does not exist", svc)
}
meta := svcPortMapRuleMeta(svc, targetIP, pm)
p, err := protoFromString(pm.Protocol)
if err != nil {
t.Fatalf("error converting protocol: %v", err)
}
wantsRule := portMapRule(chain.Table, chain, "tailscale0", targetIP, pm.MatchPort, pm.TargetPort, p, meta)
gotRule, err := findRule(runner.conn, wantsRule)
if err != nil {
t.Fatalf("error looking up rule: %v", err)
}
if gotRule == nil {
t.Fatalf("rule not found")
}
}