banger/internal/daemon/dispatch_test.go
Thales Maciel e47b8146dc
daemon: thread per-RPC op_id end-to-end
Today there's no way to correlate a CLI failure with a daemon log
line. operationLog records relative timing but no id, two concurrent
vm.start calls log indistinguishably, and the async
vmCreateOperationState.ID is user-facing yet never reaches the
journal. The root helper logs plain text to stderr while bangerd
logs JSON, so a merged journalctl is hard to grep across the
trust-boundary split.

Mint a per-RPC op id at dispatch entry, store it on context, and
include it as an "op_id" attr on every operationLog record. The
id is stamped onto every error response (including the early
short-circuit paths bad_version and unknown_method). rpc.Call
forwards the context op id on requests so a daemon RPC and the
helper RPCs it triggers all share one id. The helper now logs
JSON to match bangerd, adopts the inbound id, and emits a single
"helper rpc completed" / "helper rpc failed" line per call so
operators can see at a glance how long each privileged op took.

vmCreateOperationState.ID is now the same id dispatch generated
for vm.create.begin — one identifier between client status polls,
daemon logs, and helper logs.

The wire format gains two optional fields: rpc.Request.OpID and
rpc.ErrorResponse.OpID, both omitempty so older peers (and the
opposite direction) ignore them. ErrorResponse.Error() now appends
"(op-XXXXXX)" to its string form when set; existing callers that
just print err.Error() get the id for free.

Tests cover: dispatch stamps op_id on unknown_method, bad_version,
and handler-returned errors; rpc.Call exposes the typed
*ErrorResponse via errors.As so the CLI can read code/op_id; ctx
op_id is forwarded to the server in the request envelope.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 22:13:44 -03:00

140 lines
3.9 KiB
Go

package daemon
import (
"context"
"sort"
"strings"
"testing"
"banger/internal/rpc"
)
// TestRPCHandlersMatchDocumentedMethods pins the surface of the RPC
// table: adding or removing a method should be an explicit, reviewable
// change. If the keyset drifts and this test isn't updated alongside,
// that's a red flag — either the documented list is stale, or a
// method sneaked in without being discussed.
//
// The expected list is the single source of truth for "methods
// banger speaks." Any production code consulting it (CLI completions,
// docs generator) can grep this test.
func TestRPCHandlersMatchDocumentedMethods(t *testing.T) {
expected := []string{
"image.delete",
"image.list",
"image.promote",
"image.pull",
"image.register",
"image.show",
"kernel.catalog",
"kernel.delete",
"kernel.import",
"kernel.list",
"kernel.pull",
"kernel.show",
"ping",
"shutdown",
"vm.create",
"vm.create.begin",
"vm.create.cancel",
"vm.create.status",
"vm.delete",
"vm.health",
"vm.kill",
"vm.list",
"vm.logs",
"vm.ping",
"vm.ports",
"vm.restart",
"vm.set",
"vm.show",
"vm.ssh",
"vm.start",
"vm.stats",
"vm.stop",
"vm.workspace.export",
"vm.workspace.prepare",
}
got := make([]string, 0, len(rpcHandlers))
for name := range rpcHandlers {
got = append(got, name)
}
sort.Strings(got)
sort.Strings(expected)
if len(got) != len(expected) {
t.Fatalf("method count: got %d, want %d\n got: %v\n want: %v", len(got), len(expected), got, expected)
}
for i := range expected {
if got[i] != expected[i] {
t.Fatalf("method[%d]: got %q, want %q\n full got: %v\n full want: %v", i, got[i], expected[i], got, expected)
}
}
}
// TestRPCHandlersAllNonNil catches a silly-but-possible footgun:
// registering a method with a nil function literal.
func TestRPCHandlersAllNonNil(t *testing.T) {
for name, h := range rpcHandlers {
if h == nil {
t.Errorf("rpcHandlers[%q] = nil", name)
}
}
}
// TestDispatchStampsOpIDOnError pins the contract that every error
// response leaving dispatch carries an op_id, even on the
// short-circuit paths (bad_version, unknown_method) that never
// reach a handler. Operators rely on this id to correlate a CLI
// failure to a daemon log line.
func TestDispatchStampsOpIDOnError(t *testing.T) {
d := &Daemon{}
t.Run("unknown_method", func(t *testing.T) {
resp := d.dispatch(context.Background(), rpc.Request{Version: rpc.Version, Method: "no.such.method"})
if resp.OK {
t.Fatalf("expected error response, got %+v", resp)
}
if resp.Error == nil || resp.Error.Code != "unknown_method" {
t.Fatalf("error = %+v, want unknown_method", resp.Error)
}
if !strings.HasPrefix(resp.Error.OpID, "op-") {
t.Fatalf("op_id = %q, want op-* prefix", resp.Error.OpID)
}
})
t.Run("bad_version", func(t *testing.T) {
resp := d.dispatch(context.Background(), rpc.Request{Version: rpc.Version + 99, Method: "ping"})
if resp.OK {
t.Fatalf("expected error response, got %+v", resp)
}
if resp.Error == nil || resp.Error.Code != "bad_version" {
t.Fatalf("error = %+v, want bad_version", resp.Error)
}
if !strings.HasPrefix(resp.Error.OpID, "op-") {
t.Fatalf("op_id = %q, want op-* prefix", resp.Error.OpID)
}
})
}
// TestDispatchPropagatesOpIDFromContext covers the case where a
// handler returns its own rpc.NewError with an empty op_id (most
// service errors do); the dispatch wrapper must stamp the
// dispatch-generated id on the way out.
func TestDispatchPropagatesOpIDFromContext(t *testing.T) {
d := &Daemon{
requestHandler: func(_ context.Context, _ rpc.Request) rpc.Response {
return rpc.NewError("operation_failed", "deliberate test failure")
},
}
resp := d.dispatch(context.Background(), rpc.Request{Version: rpc.Version, Method: "anything"})
if resp.OK || resp.Error == nil {
t.Fatalf("expected error response, got %+v", resp)
}
if !strings.HasPrefix(resp.Error.OpID, "op-") {
t.Fatalf("dispatch did not stamp op_id: %+v", resp.Error)
}
}