Compare commits

...
Author SHA1 Message Date
jschoubben e56416fa8c The mesh's one resolver: its seat, a provider's address, zones, and a node's hosts file (hq ADR 0194, 0196, 0199)
- mesh-dns-resolver: a mesh seat delivering wildcard-resolution, so every node's resolver
  configuration resolves to its one holder; node-dns-resolver kept until nothing claims it.
- ${bound:<provision>:address}: the providing machine's private address, for the one consumer
  that cannot use a name — a machine's resolver configuration.
- zone: a module declares the zone it answers and the listen that answers it; the controller
  settles it per node, refuses duplicates and shadowing, and hands the resolver .Zones to forward.
- node-hosts-file: a node seat whose holder owns /etc/hosts, with entries/add/remove.
The resolver tests follow the catalogue: no runtime dns (containers copy the machine's resolvers),
live-restore held by resolv-conf, resolv.conf naming the resolver by address then a public one.
2026-10-04 00:38:48 +02:00
12 changed files with 430 additions and 40 deletions
+8 -8
View File
@@ -254,11 +254,10 @@ func theResolver(t *testing.T) catalogue.Manifest {
return m
}
// The resolver is handed every machine on the private network as a wildcard, the same set and the
// same source as the hosts file, and is handed it again when a machine leaves — through the
// module's own manifest asking for the fact, with no module of the mesh's own in between (hal
// dnsmasq-app conversion, novox/hq 08-connectivity). The runtime on that machine is pointed at the
// machine's own address, where the resolver answers for its containers.
// The resolver is handed every machine on the private network as a wildcard, and is handed it again
// when a machine leaves — through the module's own manifest asking for the fact, with no module of the
// mesh's own in between (hal dnsmasq-app conversion, novox/hq 08-connectivity). It is the mesh's one
// resolver (ADR 0194), and the container runtime is given no resolver of its own (ADR 0196).
func TestTheResolverIsToldEveryMachineOnTheNetworkAndToldAgainWhenOneLeaves(t *testing.T) {
open := aMesh(t)
ctx := t.Context()
@@ -293,11 +292,12 @@ func TestTheResolverIsToldEveryMachineOnTheNetworkAndToldAgainWhenOneLeaves(t *t
t.Errorf("the resolver's machines lack %q:\n%s", want, first)
}
}
// The container runtime is given no resolver of its own (novox/hq ADR 0196): it copies its
// machine's, which name the mesh's resolver first. A `dns` key would be a second account of where a
// container asks, read only when the runtime starts.
for _, r := range composed(t, open, "anchor").Resources {
if r["id"] == "dnsmasq.runtime-dns" {
if !strings.Contains(r["content"].(string), `"10.77.0.1"`) || r["into"] != "json" {
t.Errorf("the runtime is not pointed at this machine's own address, written into its file: %v", r)
}
t.Errorf("the resolver still writes the runtime's own dns: %v", r)
}
}
+73 -1
View File
@@ -657,6 +657,13 @@ func renderingFor(ctx context.Context, open *stores, node string,
machines[name] = at
}
// And every zone a module in the mesh answers itself (novox/hq ADR 0199), for the mesh's resolver
// to forward.
zones, err := zonesInTheMesh(ctx, open)
if err != nil {
return catalogue.Rendering{}, inventory.Node{}, err
}
// **The bus is never public** (novox/hq ADR 0169). It was a foundation port — widened from the
// broker's own `from: mesh` to from-anywhere on the broker's host, so a machine could enrol
// before it had an address on the private network. A machine joins through the tunnel now, and
@@ -726,7 +733,7 @@ func renderingFor(ctx context.Context, open *stores, node string,
BusMembership: memberships[node],
Settings: settings, Generators: gens, Grants: grants, Needed: needed, Ports: ports,
Certificate: certificate, Authority: authority, Mesh: private, Names: names,
Machines: machines,
Machines: machines, Zones: zones,
Suffix: overlay.Suffix(), MeshRange: meshRange, TunnelInterface: overlay.Interface, Accounts: accounts, Foundation: foundation,
Kept: kept, Adopted: record.Adopted, OutwardLinks: outwardLinks,
Given: given, Taken: taken, Seats: seats, ArtifactStore: artifactStore, Built: built,
@@ -734,6 +741,71 @@ func renderingFor(ctx context.Context, open *stores, node string,
}, record, nil
}
// zonesInTheMesh is every zone a module in the mesh declares, where the mesh placed it (novox/hq ADR
// 0199): the zone settled from that node's settings, the node's private address, the port the
// answering listen is published on there.
//
// Read across every machine's resolution, as the roster once read routed names: a node whose set does
// not compose declares nothing and is passed over, so one broken machine does not cost the rest their
// zones; a store that cannot be read is raised, naming the machine, because returning the zones
// without it would withdraw them from the resolver as if the operator had (novox/hq 04-ISSUES/152).
// What the mesh refuses about the zones together — one declared twice, one shadowing the mesh's
// suffix or a node's public domain — is refused here, by name.
func zonesInTheMesh(ctx context.Context, open *stores) ([]catalogue.ZoneAt, error) {
inv := open.inventory
places, err := inv.Overlays(ctx)
if err != nil {
return nil, fmt.Errorf("where the machines are cannot be read: %w", err)
}
address := map[string]string{}
for _, p := range places {
if strings.TrimSpace(p.Address) != "" {
address[p.Name] = p.Address
}
}
nodes, err := inv.Nodes(ctx)
if err != nil {
return nil, fmt.Errorf("which machines the mesh has cannot be read: %w", err)
}
var zones []catalogue.ZoneAt
var public []string
for _, n := range nodes {
plan, _, err := planFor(ctx, open, n.Name)
switch {
case unresolvable(err):
continue
case err != nil:
return nil, fmt.Errorf("the zones %s answers cannot be read: %w", n.Name, err)
}
if plan.PublicDomain != "" {
public = append(public, plan.PublicDomain)
}
for _, m := range plan.Modules {
if m.Zone == nil {
continue
}
at := address[n.Name]
if at == "" {
// Not on the private network yet: nothing could reach its answerer.
continue
}
published, layers, err := portsGivenOn(ctx, inv, n.Name, m)
if err != nil {
return nil, fmt.Errorf("the zone %s declares on %s cannot be read: %w", m.Module, n.Name, err)
}
z, err := catalogue.ZoneOn(m, layers, published, n.Name, at)
if err != nil {
return nil, err
}
zones = append(zones, *z)
}
}
if problems := catalogue.ZonesProblems(zones, overlay.Suffix(), public); len(problems) > 0 {
return nil, fmt.Errorf("the mesh's zones cannot be forwarded:\n - %s", strings.Join(problems, "\n - "))
}
return zones, nil
}
// certificateFor is what the mesh certifies about one machine's internal name.
//
// It reaches across two contexts and reads neither one's store from the other: `inventory` knows
+3 -2
View File
@@ -21,8 +21,9 @@ import (
// formats and gains no fields.
//
// **It stays name-agnostic** ([ADR 0027]). The mesh does not learn what a `postgres-database` is:
// `at`, `as` and `from` are facts about any provision at all, and everything else comes from what
// the provider said it serves — whose keys are agreed by the requirement's name, not by this file.
// `at`, `as`, `from` and `address` (the providing machine's private address, novox/hq ADR 0194) are
// facts about any provision at all, and everything else comes from what the provider said it
// serves — whose keys are agreed by the requirement's name, not by this file.
// bound is where a module says a value from one of its bindings belongs:
// ${bound:<provision>.<key>}.
@@ -232,3 +232,34 @@ func TestTwoModulesOnOneNodeAreTwoIdentities(t *testing.T) {
t.Fatal("one module on two machines shares an identity")
}
}
// A machine's resolver configuration must name its resolver by address — it cannot resolve the name
// of the thing it resolves names with (novox/hq ADR 0194). So a binding offers the providing
// machine's private address beside its name, and only when the machine has one.
func TestABindingOffersTheProvidersAddress(t *testing.T) {
consumer := func() Resolution {
return Resolution{
Node: "workstation",
Modules: []Manifest{{
Module: "resolv-conf",
Requires: []string{"wildcard-resolution"},
Resources: []map[string]any{{
"id": "resolv", "type": "file", "path": "/etc/resolv.conf", "mode": "0644",
"content": "nameserver ${bound:wildcard-resolution:address}\n",
}},
}},
Needs: []Needed{{Name: "wildcard-resolution", From: "anchor", At: "anchor.internal", For: "resolv-conf"}},
}
}
out, err := consumer().Declaration(Rendering{Machines: map[string]string{"anchor.internal": "10.77.0.1"}})
if err != nil {
t.Fatal(err)
}
if got := fileNamed(out, "resolv-conf.resolv")["content"]; got != "nameserver 10.77.0.1\n" {
t.Fatalf("the resolver is not named by its address: %q", got)
}
// A machine with no address yet: refused, never written with a blank where the address belongs.
if _, err := consumer().Declaration(Rendering{}); err == nil {
t.Fatal("a file naming an address the mesh does not have was composed")
}
}
+15 -1
View File
@@ -155,6 +155,10 @@ type Rendering struct {
// standing beside the machines and looking as real as they do.
Machines map[string]string
// Zones is every zone a module in the mesh answers itself, where it is answered (novox/hq ADR
// 0199): the mesh's resolver forwards each one there.
Zones []ZoneAt
Settings SettingsBy
Generators map[string]Generator
// Grants are the credentials this node must create, for the provisions it offers. Passed in
@@ -732,6 +736,16 @@ func (r Resolution) compose(with Rendering, owner map[string]string,
known[provision] = values
}
}
// And the providing machine's private address, beside its name (novox/hq ADR 0194). A name is
// what nearly every consumer wants; the one that cannot use it is a machine's resolver
// configuration, which must reach the resolver before it can resolve anything — the resolver's
// own name included. Absent when the machine has no address yet, so a file naming it is refused
// rather than written with a blank where an address belongs.
for _, values := range known {
if address := with.Machines[values["at"]]; address != "" {
values["address"] = address
}
}
// And what the module is called through each requirement it contributes to (novox/hq
// 04-ISSUES/122) — the same composition its binding file carries.
for provision, values := range known {
@@ -901,7 +915,7 @@ func (r Resolution) compose(with Rendering, owner map[string]string,
// plane's; making a name resolve is the module's software. Emitted as ordinary files under
// this module's name, so they are applied, reported and removed exactly as anything else
// it declares.
given, err := FactsInto(m, r, with.Names, with.Machines, with.Accounts, with.Suffix)
given, err := FactsWithZonesInto(m, r, with.Names, with.Machines, with.Accounts, with.Suffix, with.Zones)
if err != nil {
return nil, err
}
+5
View File
@@ -540,6 +540,10 @@ type Manifest struct {
// `restart-on` names to restart when the roster changes.
Facts map[string]RosterFile `json:"facts,omitempty"`
// Zone is the zone of names this module answers itself, and the listen that answers it (novox/hq
// ADR 0199). The mesh's resolver forwards the zone to it; nothing here names an address.
Zone *Zone `json:"zone,omitempty"`
// Certificate is where this module wants a certificate for its machine's name inside the
// mesh, and where the key that goes with it can be found.
//
@@ -1742,6 +1746,7 @@ func ParseManifest(raw []byte) (Manifest, error) {
}
}
problems = append(problems, zoneProblems(m)...)
if len(problems) > 0 {
sort.Strings(problems)
return Manifest{}, fmt.Errorf("this manifest cannot be used:\n - %s",
+44 -23
View File
@@ -48,9 +48,12 @@ func TestTheResolverForwardsToFixedUpstreamsAndNeverReadsResolvConf(t *testing.T
}
for _, want := range []string{
"\nno-resolv\n", "\nserver=1.1.1.1\n", "\nserver=8.8.8.8\n",
// Loopback is the mesh-wide setting's default; a machine answering its own LAN adds its
// address there (novox/hq issue 198).
"\nlisten-address=${setting:listen-addresses}\n", "\nlisten-address=${machine:address}\n", "\nbind-dynamic\n",
// The private address and loopback, never a LAN's (novox/hq ADR 0194): a device that is not a
// member cannot reach what the mesh's names point at.
"\nlisten-address=127.0.0.1\n", "\nlisten-address=${machine:address}\n", "\nbind-dynamic\n",
// No hosts file and no operator's files: the mesh's resolver answers every node (ADR 0199).
"\nno-hosts\n",
"\nconf-file=" + m.Facts["zones"].Path + "\n",
"\ndomain-needed\n", "\nbogus-priv\n",
"\nconf-file=" + m.Facts["node-zones"].Path + "\n",
} {
@@ -73,7 +76,15 @@ func TestTheResolverForwardsToFixedUpstreamsAndNeverReadsResolvConf(t *testing.T
t.Errorf("the resolver listens on %s", taken)
}
}
// And the file that decides what the machine asks names it there, alone.
// Never a directory or a file the operator keeps: a line written for one machine's programs would
// become an answer for every node (ADR 0199).
for _, never := range []string{"conf-dir=", "addn-hosts=", "listen-address=${setting:"} {
if strings.Contains(config, never) {
t.Errorf("the mesh's resolver still reads or listens on %q", never)
}
}
// And the file that decides what the machine asks names the mesh's resolver first, by address,
// and a public one second, asked only when the first is silent (ADR 0196).
var resolv string
for _, r := range catalogueManifest(t, "resolv-conf").Resources {
if r["path"] == "/etc/resolv.conf" {
@@ -86,13 +97,16 @@ func TestTheResolverForwardsToFixedUpstreamsAndNeverReadsResolvConf(t *testing.T
nameservers = append(nameservers, strings.TrimPrefix(line, "nameserver "))
}
}
if len(nameservers) != 1 || nameservers[0] != "127.0.0.1" {
t.Errorf("resolv.conf names %v; the predecessor's names the mesh's resolver alone at 127.0.0.1", nameservers)
if len(nameservers) != 2 || nameservers[0] != "${bound:wildcard-resolution:address}" || nameservers[1] != "1.1.1.1" {
t.Errorf("resolv.conf names %v; the mesh's resolver by address first, a public one second", nameservers)
}
// The split-DNS alternative points at the same address, or a machine that keeps
if !strings.Contains(resolv, "\noptions timeout:1 attempts:1") {
t.Errorf("the fallback is not reached after one short attempt:\n%s", resolv)
}
// The split-DNS alternative points at the same resolver, or a machine that keeps
// systemd-resolved in charge would route the mesh's suffix to nothing.
for _, r := range catalogueManifest(t, "resolved-split-dns").Resources {
if content, _ := r["content"].(string); content != "" && !strings.Contains(content, "DNS=127.0.0.1\n") {
if content, _ := r["content"].(string); content != "" && !strings.Contains(content, "DNS=${bound:wildcard-resolution:address}\n") {
t.Errorf("resolved-split-dns does not point at the resolver's address:\n%s", content)
}
}
@@ -100,7 +114,8 @@ func TestTheResolverForwardsToFixedUpstreamsAndNeverReadsResolvConf(t *testing.T
// The resolver and what points the machine at it compose on one machine, and what arrives is the
// mesh's account of every machine as a wildcard, the suffix kept local, the daemon restarting on
// that file, and the runtime pointed at this machine's own address.
// that file, the machine pointed at the resolver by address, and the runtime given no resolver of
// its own but kept running across a restart (ADR 0196).
func TestTheResolverAndWhatAsksItComposeOnOneMachine(t *testing.T) {
got, err := Resolve(resolverShelf(t), []string{"dnsmasq", "resolv-conf"},
Node{Name: "anchor", At: "anchor.internal"}, World{})
@@ -115,6 +130,7 @@ func TestTheResolverAndWhatAsksItComposeOnOneMachine(t *testing.T) {
// issue 111) — the resolver's zones read only the second, and in this scenario the two
// happen to be the same map, since nothing routed is part of it.
Names: twoMachines, Machines: twoMachines, Suffix: "internal",
Zones: []ZoneAt{{Zone: "incus", Address: "10.42.0.2", Port: 5353}},
Needed: map[string]map[string]string{"dnsmasq": {"broker": "sealed"}},
Settings: SettingsBy{"dnsmasq": {{From: "the mesh", Values: map[string]any{"listen-addresses": "127.0.0.1"}}}},
})
@@ -144,24 +160,29 @@ func TestTheResolverAndWhatAsksItComposeOnOneMachine(t *testing.T) {
for _, id := range service["restart-on"].([]any) {
reflects[id.(string)] = true
}
if !reflects["dnsmasq.config"] || !reflects["dnsmasq.fact-node-zones"] {
t.Errorf("the daemon does not restart on its configuration and the machines file both: %v", service["restart-on"])
if !reflects["dnsmasq.config"] || !reflects["dnsmasq.fact-node-zones"] || !reflects["dnsmasq.fact-zones"] {
t.Errorf("the daemon does not restart on its configuration, the machines and the zones: %v", service["restart-on"])
}
if z, _ := ids["dnsmasq.fact-zones"]["content"].(string); !strings.Contains(z, "server=/incus/10.42.0.2#5353\n") {
t.Errorf("the resolver was not told to forward the zone to its answerer:\n%s", z)
}
// The runtime's own file, written into (novox/hq ADR 0102) with the keys this module states:
// where containers resolve, and that a restart keeps them running — because the runtime reads
// `dns` only when it starts, and the one restart that needs is the operator's (issue 110).
runtime := ids["dnsmasq.runtime-dns"]
// The runtime's own file, written into (novox/hq ADR 0102) with one key, by what decides how the
// machine resolves: a restart keeps every container running. No `dns` — a container copies its
// machine's resolvers (ADR 0196), and the mesh's resolver is not written into the runtime twice.
if ids["dnsmasq.runtime-dns"] != nil {
t.Errorf("the resolver still writes the runtime's dns: %v", ids["dnsmasq.runtime-dns"])
}
runtime := ids["resolv-conf.runtime-config"]
if runtime == nil || runtime["path"] != "/etc/docker/daemon.json" || runtime["into"] != "json" {
t.Fatalf("the runtime's dns is not written into its file: %v", runtime)
t.Fatalf("live-restore is not written into the runtime's file: %v", runtime)
}
var keys map[string]any
if err := json.Unmarshal([]byte(runtime["content"].(string)), &keys); err != nil {
t.Fatalf("the runtime's keys are not JSON: %v", err)
}
dns, _ := keys["dns"].([]any)
if len(keys) != 2 || len(dns) != 1 || dns[0] != "10.42.0.1" || keys["live-restore"] != true {
t.Errorf("the runtime is given %v; containers resolve at this machine's own private-network address, a restart keeps them, and nothing else is written", keys)
if len(keys) != 1 || keys["live-restore"] != true {
t.Errorf("the runtime is given %v; live-restore and nothing else", keys)
}
// The runtime is reloaded when that file changes, and never restarted: a restart stops every
// container on the machine (ADR 0102), and a reload is what turns live-restore on.
@@ -171,10 +192,10 @@ func TestTheResolverAndWhatAsksItComposeOnOneMachine(t *testing.T) {
continue
}
if _, restarts := r["restart-on"]; restarts {
t.Errorf("the resolver orders the runtime restarted, which stops every container (ADR 0102): %v", r)
t.Errorf("the runtime is ordered restarted, which stops every container (ADR 0102): %v", r)
}
for _, on := range asStrings(r["reload-on"]) {
if on == "dnsmasq.runtime-dns" {
if on == "resolv-conf.runtime-config" {
reloaded = true
}
}
@@ -184,8 +205,8 @@ func TestTheResolverAndWhatAsksItComposeOnOneMachine(t *testing.T) {
}
resolv := ids["resolv-conf.resolv"]
if resolv == nil || !strings.Contains(resolv["content"].(string), "\nnameserver 127.0.0.1\n") {
t.Fatalf("the machine is not pointed at the resolver: %v", resolv)
if resolv == nil || !strings.Contains(resolv["content"].(string), "\nnameserver 10.42.0.1\nnameserver 1.1.1.1\n") {
t.Fatalf("the machine is not pointed at the resolver by address, with the public fallback: %v", resolv)
}
}
+27
View File
@@ -61,6 +61,16 @@ type rosterView struct {
Suffix string
Names []rosterEntry
Machines []rosterEntry
// Zones is every zone a module in the mesh answers itself, with where its answerer is (novox/hq
// ADR 0199) — what the mesh's resolver forwards. Ordered by zone.
Zones []rosterZone
}
// rosterZone is one zone as a template sees it: the zone, and the address and port answering it.
type rosterZone struct {
Zone string
Address string
Port int
}
// rosterEntry is one machine as a template sees it: its bare name, its full mesh name, its address,
@@ -80,6 +90,12 @@ type rosterEntry struct {
// `machines` is only the machines — the two must not be confused (novox/hq 04-ISSUES/111), so both
// are given and the template chooses.
func FactsInto(m Manifest, r Resolution, every, machines, accounts map[string]string, suffix string) ([]map[string]any, error) {
return FactsWithZonesInto(m, r, every, machines, accounts, suffix, nil)
}
// FactsWithZonesInto is FactsInto with the mesh's zones in the view, for a template that ranges them.
func FactsWithZonesInto(m Manifest, r Resolution, every, machines, accounts map[string]string, suffix string,
zones []ZoneAt) ([]map[string]any, error) {
if len(m.Facts) == 0 {
return nil, nil
}
@@ -94,6 +110,7 @@ func FactsInto(m Manifest, r Resolution, every, machines, accounts map[string]st
Suffix: strings.TrimPrefix(suffixOr(suffix), "."),
Names: entriesFrom(every, accounts, suffix),
Machines: entriesFrom(machines, accounts, suffix),
Zones: zonesFrom(zones),
}
out := make([]map[string]any, 0, len(names))
@@ -223,3 +240,13 @@ func sortedNames(addresses map[string]string) []string {
sort.Strings(out)
return out
}
// zonesFrom is the zones a template ranges, ordered by zone so two renderings of one mesh are one file.
func zonesFrom(zones []ZoneAt) []rosterZone {
out := make([]rosterZone, 0, len(zones))
for _, z := range zones {
out = append(out, rosterZone{Zone: z.Zone, Address: z.Address, Port: z.Port})
}
sort.Slice(out, func(i, j int) bool { return out[i].Zone < out[j].Zone })
return out
}
+24
View File
@@ -107,7 +107,31 @@ var defaultSeats = []Seat{
// that machine unresolvable in the meantime. Deleted once no registered manifest claims it.
{Name: "mesh-build-machine", Scope: ScopeMesh,
Accepts: []string{"build"}, Emits: []string{"started", "built", "log.*"}, Decision: "novox/hq ADR 0190"},
// **The mesh's one resolver** (novox/hq ADR 0194, 0196): every node's internal domain, held in one
// place, and every node and container asks it first. Delivers what a machine's resolver
// configuration requires, so that requirement resolves to the holder wherever it is placed.
{Name: "mesh-dns-resolver", Scope: ScopeMesh, Delivers: "wildcard-resolution", Decision: "novox/hq ADR 0194"},
// **Retired by ADR 0194, kept while a manifest still claims it** — the same reason as
// mesh-build-machine above: a machine still holds it until the mesh's resolver replaces it, and
// removing the row first would make that machine unresolvable. Deleted once nothing claims it.
{Name: "node-dns-resolver", Scope: ScopeNode, Decision: "novox/hq ADR 0121"},
// **A machine's /etc/hosts is one module's** (novox/hq ADR 0199): its holder writes the machine's
// own lines and keeps every other line as the operator's, changed through these three verbs on that
// machine alone. The controller holds none of it.
{Name: "node-hosts-file", Scope: ScopeNode, Decision: "novox/hq ADR 0199",
Serves: []Verb{
{Name: "entries", Description: "Every line of this machine's /etc/hosts, each marked whose it is: " +
"the operator's, or the block of the module or tool that writes it.",
Input: schema(map[string]string{}, nil)},
{Name: "add", Description: "Add one address and its names to the operator's lines of this machine's " +
"/etc/hosts — a name for this machine's own programs, not the mesh's.",
Input: schema(map[string]string{"address": "the IPv4 or IPv6 address",
"names": "the names for it, separated by spaces"}, []string{"address", "names"})},
{Name: "remove", Description: "Remove one name, or every line of one address, from the operator's " +
"lines of this machine's /etc/hosts. A line a module writes is refused, naming the module.",
Input: schema(map[string]string{"name": "a host name, or an address to remove every line of"},
[]string{"name"})},
}},
// The intrusion prevention's verbs (novox/hq ADR 0179): what a person asks a machine's ban list
// whatever keeps it — who is banned and why, ban one address, let one go. Every holder serves all
// four; the jails themselves are composed from the modules the machine runs (to-be 31).
+4 -4
View File
@@ -44,10 +44,10 @@ func TestTheSeatsAreAClosedSetAndEachNamesItsDecision(t *testing.T) {
delivered[s.Delivers] = s.Name
}
}
// Seventeen since node-build-agent (novox/hq ADR 0190) — sixteen once the retired
// mesh-build-machine row goes, when no registered manifest claims it any more.
if len(Seats()) != 17 {
t.Errorf("the mesh defines %d seats rather than 17; the set is closed, so a change here is "+
// Nineteen since mesh-dns-resolver (novox/hq ADR 0194) and node-hosts-file (ADR 0199) — two fewer
// once the retired mesh-build-machine and node-dns-resolver rows go, when no manifest claims either.
if len(Seats()) != 19 {
t.Errorf("the mesh defines %d seats rather than 19; the set is closed, so a change here is "+
"a decision (novox/hq ADR 0110): %s", len(Seats()), seatNames())
}
}
+117
View File
@@ -0,0 +1,117 @@
package catalogue
import (
"fmt"
"sort"
"strings"
)
// Zones: names a module answers itself (novox/hq ADR 0199).
//
// The mesh's resolver holds each node's internal domain and nothing else (ADR 0191, 0194). A module
// whose names are its own — the lab's scenario machines, known only while a scenario runs — declares
// the zone it answers and the listen that answers it; the controller hands the resolver's holder every
// zone with the declaring node's private address and the port that listen is published on, and the
// holder forwards the zone there. **A definition names no address** (ADR 0112): the zone is a setting,
// the listen is the module's own, and where they are is the mesh's fact.
// Zone is the manifest's declaration that a module answers the names in one zone.
type Zone struct {
// Name is the zone: a label or a dotted name, normally `${setting:<key>}`, so the operator chooses
// it and the definition does not.
Name string `json:"name"`
// Listen names one of the module's listens: the DNS answerer for the zone.
Listen string `json:"listen"`
}
// ZoneAt is a declared zone where the mesh placed it: what the resolver's holder forwards, and where.
type ZoneAt struct {
Zone string
Node string
Module string
Address string
Port int
}
// zoneProblems is what is wrong with a module's zone declaration on its own, before any node.
func zoneProblems(m Manifest) []string {
if m.Zone == nil {
return nil
}
var problems []string
if strings.TrimSpace(m.Zone.Name) == "" {
problems = append(problems, fmt.Sprintf("%s declares a zone with no name", m.Module))
}
if !m.hasListen(m.Zone.Listen) {
problems = append(problems, fmt.Sprintf(
"%s declares zone %q answered by listen %q, and has no listen of that name",
m.Module, m.Zone.Name, m.Zone.Listen))
}
return problems
}
func (m Manifest) hasListen(name string) bool {
if name == "" {
return false
}
for _, l := range m.Listens {
if l.Name == name {
return true
}
}
return false
}
// ZoneOn is one module's zone as one node places it: the name settled from the node's settings, the
// port its answering listen is published on there. Nothing when the module declares no zone.
func ZoneOn(m Manifest, layers []Layer, published map[int]int, node, address string) (*ZoneAt, error) {
if m.Zone == nil {
return nil, nil
}
settled, err := Settle(map[string]any{"zone": m.Zone.Name}, layers)
if err != nil {
return nil, fmt.Errorf("%s's zone on %s: %w", m.Module, node, err)
}
zone := strings.Trim(strings.ToLower(fmt.Sprint(settled["zone"])), ".")
var port int
for _, l := range m.Listens {
if l.Name == m.Zone.Listen {
port = l.Port
if at, given := published[l.Port]; given {
port = at
}
}
}
return &ZoneAt{Zone: zone, Node: node, Module: m.Module, Address: address, Port: port}, nil
}
// ZonesProblems is what the mesh refuses about its zones together: one zone declared twice, a zone
// that is the mesh's suffix or under it, a zone that is a node's public domain or under one. A module
// may not shadow names the mesh's resolver or the public DNS answers.
func ZonesProblems(zones []ZoneAt, suffix string, publicDomains []string) []string {
var problems []string
under := func(zone, domain string) bool {
domain = strings.Trim(strings.ToLower(domain), ".")
return domain != "" && (zone == domain || strings.HasSuffix(zone, "."+domain))
}
seen := map[string]ZoneAt{}
for _, z := range zones {
if other, twice := seen[z.Zone]; twice && (other.Node != z.Node || other.Module != z.Module) {
problems = append(problems, fmt.Sprintf("zone %q is declared by %s on %s and by %s on %s; one module answers a zone",
z.Zone, other.Module, other.Node, z.Module, z.Node))
}
seen[z.Zone] = z
if under(z.Zone, suffix) {
problems = append(problems, fmt.Sprintf("%s on %s declares zone %q, which is the mesh's own suffix or under it",
z.Module, z.Node, z.Zone))
}
for _, d := range publicDomains {
if under(z.Zone, d) {
problems = append(problems, fmt.Sprintf("%s on %s declares zone %q, which is the public domain %q or under it",
z.Module, z.Node, z.Zone, d))
}
}
}
sort.Strings(problems)
return problems
}
+78
View File
@@ -0,0 +1,78 @@
package catalogue
import (
"strings"
"testing"
)
// A zone names the listen that answers it, or there is nothing to forward to (novox/hq ADR 0199).
func TestAZoneMustNameOneOfTheModulesListens(t *testing.T) {
_, err := ParseManifest([]byte(`{"module":"lab","version":"1",
"listens":[{"name":"dns","port":5353,"protocol":"udp","from":"mesh","why":"the lab's names"}],
"zone":{"name":"${setting:zone}","listen":"web"}}`))
if err == nil || !strings.Contains(err.Error(), `has no listen of that name`) {
t.Fatalf("a zone answered by a listen the module does not have was accepted: %v", err)
}
if _, err := ParseManifest([]byte(`{"module":"lab","version":"1",
"listens":[{"name":"dns","port":5353,"protocol":"udp","from":"mesh","why":"the lab's names"}],
"zone":{"name":"${setting:zone}","listen":"dns"}}`)); err != nil {
t.Fatalf("a well-formed zone was refused: %v", err)
}
}
// The zone is the operator's (a setting) and the port is where this machine publishes the listen —
// neither is the definition's to state.
func TestAZoneIsPlacedFromTheNodesSettingAndPublishedPort(t *testing.T) {
m := Manifest{Module: "lab", Zone: &Zone{Name: "${setting:zone}", Listen: "dns"},
Listens: []Listening{{Name: "dns", Port: 5353, From: FromMesh}}}
z, err := ZoneOn(m, []Layer{{From: "node", Values: map[string]any{"zone": "Incus."}}},
map[int]int{5353: 15353}, "workstation", "10.77.0.3")
if err != nil {
t.Fatal(err)
}
if z.Zone != "incus" || z.Address != "10.77.0.3" || z.Port != 15353 || z.Node != "workstation" {
t.Fatalf("the zone was placed as %+v", *z)
}
if _, err := ZoneOn(m, nil, nil, "workstation", "10.77.0.3"); err == nil {
t.Fatal("a zone nobody named was placed")
}
}
// One module answers a zone, and none may shadow the mesh's names or a public domain.
func TestTheMeshRefusesAZoneTwiceOrOneThatShadows(t *testing.T) {
one := ZoneAt{Zone: "incus", Node: "workstation", Module: "lab", Address: "10.77.0.3", Port: 53}
if p := ZonesProblems([]ZoneAt{one}, "internal", []string{"example.tld"}); len(p) != 0 {
t.Fatalf("one ordinary zone was refused: %v", p)
}
twice := one
twice.Node, twice.Module = "laptop", "other"
cases := map[string][]ZoneAt{
"declared by": {one, twice},
"mesh's own suffix": {{Zone: "lab.internal", Node: "a", Module: "m"}},
"public domain": {{Zone: "dev.example.tld", Node: "a", Module: "m"}},
}
for want, zones := range cases {
p := strings.Join(ZonesProblems(zones, "internal", []string{"example.tld"}), "\n")
if !strings.Contains(p, want) {
t.Errorf("not refused for %q: %q", want, p)
}
}
}
// The resolver's template sees every zone with where it is answered, in zone order.
func TestTheResolversTemplateRangesTheZones(t *testing.T) {
m := Manifest{Module: "dnsmasq", Facts: map[string]RosterFile{"zones": {
Path: "/etc/mesh-resolver/zones.conf",
Template: "{{range .Zones}}server=/{{.Zone}}/{{.Address}}#{{.Port}}\n{{end}}",
}}}
out, err := FactsWithZonesInto(m, Resolution{Node: "anchor"}, nil, nil, nil, "", []ZoneAt{
{Zone: "zeta", Address: "10.77.0.2", Port: 53},
{Zone: "incus", Address: "10.77.0.3", Port: 15353},
})
if err != nil {
t.Fatal(err)
}
if got := out[0]["content"]; got != "server=/incus/10.77.0.3#15353\nserver=/zeta/10.77.0.2#53\n" {
t.Fatalf("the resolver was told %q", got)
}
}