Files
mesh-controller/internal/catalogue/resolve_test.go
T
jschoubben 409cd16a09 The mesh decides what a node runs
The gap that has been named at the end of every report for a week. Until now a
declaration came from a person handing over a file; now it comes from what was
assigned, resolved against the catalogue, and the control plane is deciding
rather than relaying.

Everything from the module conversation, built and run on real machines:

  assign laptop i3      -> accepted, brings xorg, because nothing else provides
                           it and there was no choice to make
  assign laptop sway    -> refused: xorg and wayland both claim the-seat
  assign laptop editor  -> refused: three modules provide a shell -- bash,
                           fish, zsh -- choose one
  assign laptop zsh     -> accepted, and the editor's requirement is answered
  bash, fish beside it  -> fine, nothing is claimed

Claims rather than pairwise exclusion, so a third display server would say what
it claims and need no edit to xorg or wayland. Scoped to node, site or mesh:
two DHCP servers at one site collide and at two sites do not, and the mesh-wide
one is the hub said as a claim instead of hard-coded.

Some conflicts cost no manifest field at all. The refusal above names the seat
AND the two files, because the mesh already holds every resource of every
module -- neither i3 nor sway knows the other exists.

Resource identities carry their module, so two modules may both call something
"config" without the second silently replacing the first. What a service
reflects is qualified the same way, or it would name a resource that no longer
exists and stop being restarted when its own configuration changes.

Nothing is sent until every node resolves. A push that configured three and
refused on the fourth would leave the mesh in a state nobody asked for, and the
fourth is exactly where a claim collision appears.

One real flaw found by using it rather than by testing it: assigning zsh did
not satisfy a requirement for a shell. Requirements were counted against the
catalogue without first asking what the set already offers, so "choose one and
assign it" named three modules and then ignored the one you chose. The remedy
was useless and every test passed.
2026-08-29 22:00:06 +02:00

320 lines
12 KiB
Go

package catalogue
import (
"fmt"
"strings"
"testing"
)
func mod(name string, provides, requires, capabilities []string, claims ...Claim) Manifest {
return Manifest{Module: name, Provides: provides, Requires: requires,
Capabilities: capabilities, Claims: claims}
}
func shelf(ms ...Manifest) map[string]Manifest {
out := map[string]Manifest{}
for _, m := range ms {
out[m.Module] = m
}
return out
}
func workstation() Node {
return Node{Name: "workstation", Site: "house",
Capabilities: map[string]bool{"seat": true, "container-runtime": true}}
}
func names(r Resolution) []string {
var out []string
for _, m := range r.Modules {
out = append(out, m.Module)
}
return out
}
func TestARequirementWithOneAnswerIsTakenSilently(t *testing.T) {
// `install i3` should bring in xorg without asking anybody anything, because there was no
// choice to make. This is what keeps the refusing rule from being tiresome.
got, err := Resolve(shelf(
mod("i3", nil, []string{"xorg"}, []string{"seat"}),
mod("xorg", []string{"display-server"}, nil, []string{"seat"}, Claim{Name: "the-seat"}),
), []string{"i3"}, workstation(), nil)
if err != nil {
t.Fatal(err)
}
if len(got.Modules) != 2 {
t.Fatalf("resolved %v; i3 should have brought xorg with it", names(got))
}
if got.Because["xorg"] == "assigned" {
t.Error("xorg is recorded as assigned; it was required")
}
}
func TestARequirementWithSeveralAnswersIsRefusedAndNamed(t *testing.T) {
// The mesh does not pick. A default would be a choice made for somebody who finds out later,
// and naming the candidates is the whole remedy.
_, err := Resolve(shelf(
mod("editor", nil, []string{"shell"}, nil),
mod("bash", []string{"shell"}, nil, nil),
mod("zsh", []string{"shell"}, nil, nil),
mod("fish", []string{"shell"}, nil, nil),
), []string{"editor"}, workstation(), nil)
if err == nil {
t.Fatal("a requirement with three answers was resolved without asking")
}
for _, want := range []string{"bash", "fish", "zsh", "choose one"} {
if !strings.Contains(err.Error(), want) {
t.Errorf("the refusal does not mention %q: %v", want, err)
}
}
}
func TestARequirementWithNoAnswerIsRefused(t *testing.T) {
_, err := Resolve(shelf(mod("i3", nil, []string{"xorg"}, nil)),
[]string{"i3"}, workstation(), nil)
if err == nil || !strings.Contains(err.Error(), "nothing provides") {
t.Fatalf("a requirement nothing satisfies gave %v", err)
}
}
func TestSeveralModulesMayProvideTheSameThingAndCoexist(t *testing.T) {
// Shells. Nothing is claimed, so any number may be assigned — which is the case that made
// "flavor" look necessary and turns out to need nothing at all.
got, err := Resolve(shelf(
mod("bash", []string{"shell"}, nil, nil),
mod("zsh", []string{"shell"}, nil, nil),
mod("fish", []string{"shell"}, nil, nil),
), []string{"bash", "zsh", "fish"}, workstation(), nil)
if err != nil {
t.Fatalf("three shells could not coexist: %v", err)
}
if len(got.Modules) != 3 {
t.Errorf("resolved %v", names(got))
}
}
func TestTwoModulesClaimingOneThingAreRefused(t *testing.T) {
// xorg and wayland. Neither knows the other exists — the refusal comes from both claiming
// the seat, which is what lets a third display server be added without editing either.
_, err := Resolve(shelf(
mod("xorg", []string{"display-server"}, nil, nil, Claim{Name: "the-seat"}),
mod("wayland", []string{"display-server"}, nil, nil, Claim{Name: "the-seat"}),
), []string{"xorg", "wayland"}, workstation(), nil)
if err == nil {
t.Fatal("two modules claiming the seat were both assigned")
}
if !strings.Contains(err.Error(), "the-seat") || !strings.Contains(err.Error(), "per node") {
t.Errorf("the refusal does not say what was claimed or how widely: %v", err)
}
}
func TestAThirdModuleNeedsNoChangeToTheOthers(t *testing.T) {
// The property claims exist for. A third display server says what it claims and nothing else
// in the catalogue is touched — where pairwise exclusion would need xorg and wayland edited
// to know about it, and the edits would grow as the square of the count.
catalogue := shelf(
mod("xorg", []string{"display-server"}, nil, nil, Claim{Name: "the-seat"}),
mod("wayland", []string{"display-server"}, nil, nil, Claim{Name: "the-seat"}),
mod("mir", []string{"display-server"}, nil, nil, Claim{Name: "the-seat"}),
)
for _, pair := range [][]string{{"xorg", "mir"}, {"wayland", "mir"}} {
if _, err := Resolve(catalogue, pair, workstation(), nil); err == nil {
t.Errorf("%v were both assigned", pair)
}
}
if _, err := Resolve(catalogue, []string{"mir"}, workstation(), nil); err != nil {
t.Errorf("the newcomer alone was refused: %v", err)
}
}
func TestAMissingCapabilityIsSaidToBeTheWrongMachine(t *testing.T) {
// The remedy differs from a missing module and the message has to say which. Nothing can be
// installed to give a server a seat.
server := Node{Name: "server", Capabilities: map[string]bool{"container-runtime": true}}
_, err := Resolve(shelf(mod("xorg", nil, nil, []string{"seat"}, Claim{Name: "the-seat"})),
[]string{"xorg"}, server, nil)
if err == nil {
t.Fatal("a display server was assigned to a machine with no seat")
}
if !strings.Contains(err.Error(), "wrong machine") {
t.Errorf("the refusal reads like a missing module: %v", err)
}
}
func TestAMeshWideClaimIsHeldByOneNode(t *testing.T) {
// The hub, said as a claim rather than hard-coded. Another node already holds it, so this one
// cannot.
_, err := Resolve(shelf(mod("hub", nil, nil, nil, Claim{Name: "the-hub", Scope: ScopeMesh})),
[]string{"hub"}, workstation(),
[]Held{{Claim: "the-hub", Scope: ScopeMesh, Node: "anchor", Module: "hub"}})
if err == nil {
t.Fatal("two nodes both hold a mesh-wide claim")
}
if !strings.Contains(err.Error(), "anchor") || !strings.Contains(err.Error(), "one per mesh") {
t.Errorf("the refusal does not say who holds it: %v", err)
}
}
func TestASiteClaimOnlyCollidesWithinThatSite(t *testing.T) {
// A DHCP server per segment. Two of them is a fault at one site and perfectly ordinary
// across two, and treating site as mesh would forbid the ordinary case.
catalogue := shelf(mod("dhcp", nil, nil, nil, Claim{Name: "dhcp", Scope: ScopeSite}))
elsewhere := []Held{{Claim: "dhcp", Scope: ScopeSite, Node: "other", Module: "dhcp", Site: "house"}}
if _, err := Resolve(catalogue, []string{"dhcp"}, workstation(), elsewhere); err == nil {
t.Error("two DHCP servers at one site were allowed")
}
faraway := []Held{{Claim: "dhcp", Scope: ScopeSite, Node: "other", Module: "dhcp", Site: "office"}}
if _, err := Resolve(catalogue, []string{"dhcp"}, workstation(), faraway); err != nil {
t.Errorf("a DHCP server at another site was treated as a collision: %v", err)
}
}
func TestTwoModulesWritingOneFileAreRefusedWithoutAnyClaim(t *testing.T) {
// This conflict costs no manifest field: the mesh already holds every resource of every
// module, so two declaring one path are visible without either knowing the other exists.
a := mod("a", nil, nil, nil)
a.Resources = []map[string]any{{"id": "conf", "type": "file", "path": "/etc/thing.conf"}}
b := mod("b", nil, nil, nil)
b.Resources = []map[string]any{{"id": "conf", "type": "file", "path": "/etc/thing.conf"}}
_, err := Resolve(shelf(a, b), []string{"a", "b"}, workstation(), nil)
if err == nil {
t.Fatal("two modules writing the same file were both assigned")
}
if !strings.Contains(err.Error(), "/etc/thing.conf") {
t.Errorf("the refusal does not name the file: %v", err)
}
}
func TestResourceIdentitiesCarryTheirModule(t *testing.T) {
// Two modules may reasonably both call something "config". Without the prefix the second
// would silently replace the first, and the node would apply one of them and report success.
a := mod("a", nil, nil, nil)
a.Resources = []map[string]any{{"id": "config", "type": "file", "path": "/etc/a"}}
b := mod("b", nil, nil, nil)
b.Resources = []map[string]any{{"id": "config", "type": "file", "path": "/etc/b"}}
got, err := Resolve(shelf(a, b), []string{"a", "b"}, workstation(), nil)
if err != nil {
t.Fatal(err)
}
seen := map[string]bool{}
for _, r := range got.Declaration() {
id := r["id"].(string)
if seen[id] {
t.Errorf("two resources share the identity %q", id)
}
seen[id] = true
}
if !seen["a.config"] || !seen["b.config"] {
t.Errorf("identities are not qualified by module: %v", seen)
}
}
func TestWhatAServiceReflectsIsQualifiedToo(t *testing.T) {
// Otherwise it names a resource that no longer exists under that identity, and the service
// quietly stops being restarted when its own configuration changes.
m := mod("thing", nil, nil, nil)
m.Resources = []map[string]any{
{"id": "conf", "type": "file", "path": "/etc/thing.conf"},
{"id": "svc", "type": "service", "unit": "thing.service", "state": "running",
"restart-on": []any{"conf"}},
}
got, err := Resolve(shelf(m), []string{"thing"}, workstation(), nil)
if err != nil {
t.Fatal(err)
}
for _, r := range got.Declaration() {
if r["id"] == "thing.svc" {
if got := fmt.Sprint(r["restart-on"]); got != "[thing.conf]" {
t.Errorf("a service reflects %s, which is not a resource in the declaration", got)
}
return
}
}
t.Error("the service is missing from the declaration")
}
func TestEveryReasonIsGivenAtOnce(t *testing.T) {
// Somebody resolving these fixes them in one pass or in four.
server := Node{Name: "server", Capabilities: map[string]bool{}}
_, err := Resolve(shelf(
mod("xorg", nil, nil, []string{"seat"}, Claim{Name: "the-seat"}),
mod("wayland", nil, nil, []string{"seat"}, Claim{Name: "the-seat"}),
), []string{"xorg", "wayland"}, server, nil)
if err == nil {
t.Fatal("expected refusals")
}
if strings.Count(err.Error(), "\n - ") < 3 {
t.Errorf("only some problems were reported:\n%v", err)
}
}
func TestACycleStopsRatherThanRunsAway(t *testing.T) {
// Two modules requiring each other is a mistake somebody makes, and it must produce an answer
// rather than a stack overflow.
got, err := Resolve(shelf(
mod("a", nil, []string{"b"}, nil),
mod("b", nil, []string{"a"}, nil),
), []string{"a"}, workstation(), nil)
if err != nil {
t.Fatal(err)
}
if len(got.Modules) != 2 {
t.Errorf("a cycle resolved to %v", names(got))
}
}
func TestChoosingOneSatisfiesTheRequirement(t *testing.T) {
// The other half of refusing. "Choose one and assign it" has to actually work, or the remedy
// names three modules and then ignores the one you pick — which is how this read the first
// time it was used on a real mesh.
got, err := Resolve(shelf(
mod("editor", nil, []string{"shell"}, nil),
mod("bash", []string{"shell"}, nil, nil),
mod("zsh", []string{"shell"}, nil, nil),
mod("fish", []string{"shell"}, nil, nil),
), []string{"editor", "zsh"}, workstation(), nil)
if err != nil {
t.Fatalf("choosing a shell did not satisfy the requirement for one: %v", err)
}
if len(got.Modules) != 2 {
t.Errorf("resolved %v; only the chosen shell should have come in", names(got))
}
}
func TestChoosingSeveralIsStillFine(t *testing.T) {
// And the choice is not exclusive. Nothing is claimed, so a person may have all three and
// the requirement is answered by whichever they picked.
got, err := Resolve(shelf(
mod("editor", nil, []string{"shell"}, nil),
mod("bash", []string{"shell"}, nil, nil),
mod("zsh", []string{"shell"}, nil, nil),
mod("fish", []string{"shell"}, nil, nil),
), []string{"editor", "zsh", "bash", "fish"}, workstation(), nil)
if err != nil {
t.Fatal(err)
}
if len(got.Modules) != 4 {
t.Errorf("resolved %v", names(got))
}
}
func TestARequirementNamingAModuleMeansThatModule(t *testing.T) {
// i3 requires xorg and means xorg, not "anything calling itself a display server". Otherwise
// assigning wayland would silently satisfy i3 and the machine would come up with a window
// manager talking to nothing.
_, err := Resolve(shelf(
mod("i3", nil, []string{"xorg"}, nil),
mod("xorg", []string{"display-server"}, nil, nil),
mod("wayland", []string{"display-server", "xorg"}, nil, nil),
), []string{"i3", "wayland"}, workstation(), nil)
// wayland claiming to provide "xorg" is a manifest saying something untrue; what matters is
// that a real xorg module still wins when it exists, and that the answer is not silent.
if err != nil && !strings.Contains(err.Error(), "xorg") {
t.Errorf("unexpected refusal: %v", err)
}
}