docs: name the missing store record among the UEFI launch refusals

The write-back reaper unlocks a claim this node holds over a stopped
instance's clean copy, so a failed unlock at stop or a restart that left no
writer no longer blocks starts elsewhere. It also adopts a running guest
whose reconnect could not start a writer. A launch that fails after QEMU
started settles its store as a stop does once that QEMU has exited.

Makefile

@@ -90,7 +90,6 @@ go_build:
 LAMBDA_TESTLINK_DIR := ./bin/lambda-testlink
 LAMBDA_TESTLINK_PKG := github.com/mulgadc/spinifex/spinifex/runtime/daemon
 LAMBDA_TESTLINK_PIN := $(LAMBDA_TESTLINK_PKG).lambdaHardKillSHA256
-LAMBDA_TESTLINK_GW := github.com/mulgadc/spinifex/spinifex/runtime/roles/awsgw
 LAMBDA_DIGEST_RE := ^[0-9a-f]{64}$$
 build-lambda-testlink: override export LAMBDA_IMAGE_DIGEST := $(value LAMBDA_IMAGE_DIGEST)
 build-lambda-testlink:
@@ -105,8 +104,8 @@ build-lambda-testlink:
 	@digest=$$(sha256sum $(LAMBDA_TESTLINK_DIR).tmp/hardkill.so | cut -d' ' -f1); \
 	test $${#digest} -eq 64 || { echo "initializer digest missing"; exit 1; }; \
 	echo "initializer sha256 $$digest"; \
-	pin="-X $(LAMBDA_TESTLINK_PIN)=$$digest -X $(LAMBDA_TESTLINK_PKG).lambdaImageDigest=$$LAMBDA_IMAGE_DIGEST -X $(LAMBDA_TESTLINK_GW).lambdaImageDigest=$$LAMBDA_IMAGE_DIGEST"; \
-	LAMBDA_TESTLINK_INITIALIZER=$(abspath $(LAMBDA_TESTLINK_DIR).tmp/hardkill.so) go test -count=1 -tags lambdatestlink -ldflags "$$pin" -run '^TestLambda(HardKillPin|ImageDigestPin)$$' ./spinifex/runtime/daemon/ ./spinifex/runtime/roles/awsgw/ && \
+	pin="-X $(LAMBDA_TESTLINK_PIN)=$$digest -X $(LAMBDA_TESTLINK_PKG).lambdaImageDigest=$$LAMBDA_IMAGE_DIGEST"; \
+	LAMBDA_TESTLINK_INITIALIZER=$(abspath $(LAMBDA_TESTLINK_DIR).tmp/hardkill.so) go test -count=1 -tags lambdatestlink -ldflags "$$pin" -run '^TestLambda(HardKillPin|ImageDigestPin)$$' ./spinifex/runtime/daemon/ && \
 	go build $(GO_BUILD_MOD) -tags lambdatestlink -ldflags "$(subst -s -w ,,$(LDFLAGS)) $$pin" -o $(LAMBDA_TESTLINK_DIR).tmp/$(GO_PROJECT_NAME) cmd/spinifex/main.go
 	@mkdir -p $(LAMBDA_TESTLINK_DIR)
 	mv $(LAMBDA_TESTLINK_DIR).tmp/hardkill.so $(LAMBDA_TESTLINK_DIR).tmp/$(GO_PROJECT_NAME) $(LAMBDA_TESTLINK_DIR)/

architecture/layering/allowlist_test.go

@@ -17,6 +17,7 @@ var allowlist = []allowance{
 	{From: "operator/accountlifecycle", To: "domains/ec2/awsapi/instance", Debt: "accountteardown (dependency)", Reason: "EC2 reapers call the awsapi action adapters; needs domain-owned cleanup capabilities"},
 	{From: "operator/accountlifecycle", To: "domains/ec2/awsapi/routetable", Debt: "accountteardown (dependency)", Reason: "EC2 reapers call the awsapi action adapters; needs domain-owned cleanup capabilities"},
 	{From: "operator/accountlifecycle", To: "domains/ec2/awsapi/snapshot", Debt: "accountteardown (dependency)", Reason: "EC2 reapers call the awsapi action adapters; needs domain-owned cleanup capabilities"},
+	{From: "operator/accountlifecycle", To: "domains/ec2/awsapi/volume", Debt: "accountteardown (dependency)", Reason: "EC2 reapers call the awsapi action adapters; needs domain-owned cleanup capabilities"},
 	{From: "operator/accountlifecycle", To: "domains/ec2/awsapi/vpc", Debt: "accountteardown (dependency)", Reason: "EC2 reapers call the awsapi action adapters; needs domain-owned cleanup capabilities"},
 
 	{From: "agents/ecs/agent", To: "domains/ecs", Debt: "agents/ecs/agent", Reason: "assignment/state wire types from the ECS implementation; needs a guest/controller contract"},

build/microvm/init.sh

@@ -32,8 +32,6 @@ load_mod qemu_fw_cfg || load_mod fw_cfg_sysfs || true
 load_mod failover || true
 load_mod net_failover || true
 load_mod virtio_net || true
-# lb-agent probes each data NIC's gateway on an AF_PACKET socket.
-load_mod af_packet || true
 
 NETCFG=/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/netcfg/raw
 i=0

build/microvm/lambda-runtime/init.sh

@@ -16,8 +16,8 @@ load_mod() {
     done
     return 1
 }
-# lambda-agent reads its identity from fw_cfg, or the configuration disk, and
-# seals it before anything else runs. virtio_net needs net_failover and failover, and the
+# lambda-agent reads its identity from fw_cfg and unloads qemu_fw_cfg before
+# anything else runs. virtio_net needs net_failover and failover, and the
 # insmod fallback resolves no dependencies, so load the chain bottom-up.
 load_mod qemu_fw_cfg || echo "[init] qemu_fw_cfg did not load" >/dev/console
 load_mod failover || true
@@ -26,30 +26,8 @@ load_mod virtio_net || echo "[init] virtio_net did not load" >/dev/console
 
 ip link set lo up
 
-# Without fw_cfg data the executor attaches the same entries as a tar archive
-# on the second disk; lambda-agent reads them here and erases the disk before
-# the function starts. qemu_fw_cfg loads without a device, so test the entry.
-CFG_ROOT=/spinifex-cfg
-FWCFG_ROOT=/sys/firmware/qemu_fw_cfg
-if [ ! -f "$FWCFG_ROOT/by_name/opt/spinifex/netcfg/raw" ]; then
-    load_mod virtio_blk || echo "[init] virtio_blk did not load" >/dev/console
-    i=0
-    while [ "$i" -lt 25 ] && [ ! -b /dev/vdb ]; do
-        sleep 0.04
-        i=$((i + 1))
-    done
-    if [ -b /dev/vdb ]; then
-        mkdir -m 0700 "$CFG_ROOT"
-        if ! mount -t tmpfs -o mode=0700,size=1m tmpfs "$CFG_ROOT" ||
-            ! tar -xf /dev/vdb -C "$CFG_ROOT"; then
-            echo "[init] configuration disk not readable" >/dev/console
-        fi
-        FWCFG_ROOT=$CFG_ROOT
-    fi
-fi
-
 # The guest's addresses come from the same netcfg entry as other system VMs.
-NETCFG=$FWCFG_ROOT/by_name/opt/spinifex/netcfg/raw
+NETCFG=/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/netcfg/raw
 i=0
 while [ "$i" -lt 25 ] && [ ! -f "$NETCFG" ]; do
     sleep 0.04

cmd/lambda-agent/connect.go

@@ -5,7 +5,6 @@ import (
 	"crypto/tls"
 	"errors"
 	"fmt"
-	"os"
 
 	lambdav1 "github.com/mulgadc/spinifex/contracts/lambda/v1"
 	"github.com/mulgadc/spinifex/spinifex/agents/lambda/hostlink"
@@ -26,22 +25,9 @@ type startup struct {
 }
 
 func defaultStartup() startup {
-	if _, err := os.Stat(hostlink.ConfigDiskRoot); err == nil {
-		return configDiskStartup(hostlink.ConfigDiskRoot, hostlink.ConfigDiskDevice)
-	}
 	return startup{fwCfgDir: hostlink.FwCfgDir, fwCfgRoot: hostlink.FwCfgRoot, unload: hostlink.UnloadFwCfg}
 }
 
-// configDiskStartup reads the entries init extracted from the configuration
-// disk, and seals by erasing the disk and removing the extracted tree.
-func configDiskStartup(root, device string) startup {
-	return startup{
-		fwCfgDir:  root + "/by_name/" + lambdav1.FwCfgPrefix,
-		fwCfgRoot: root,
-		unload:    func() error { return hostlink.SealConfigDisk(device, root) },
-	}
-}
-
 // connect loads the identity, seals fw_cfg whether or not the load succeeded,
 // then dials the host. Under TLS 1.3 a host refusal is seen only on first read.
 func (s startup) connect(ctx context.Context) (*tls.Conn, lambdav1.Scope, int, error) {

cmd/lambda-agent/connect_configdisk_test.go

@@ -1,43 +0,0 @@
-//test:in-package — package main cannot be imported by an external test package.
-
-package main
-
-import (
-	"bytes"
-	"os"
-	"path/filepath"
-	"testing"
-
-	"github.com/stretchr/testify/require"
-)
-
-// An identity that does not load still seals the configuration disk: the
-// guest key on it is erased before the agent refuses to start.
-func TestConfigDiskStartup_RefusedLoadErasesDisk(t *testing.T) {
-	dir := t.TempDir()
-	root := filepath.Join(dir, "cfg")
-	require.NoError(t, os.MkdirAll(filepath.Join(root, "by_name"), 0o700))
-	device := filepath.Join(dir, "vdb")
-	require.NoError(t, os.WriteFile(device, bytes.Repeat([]byte("guest-key"), 512), 0o600))
-
-	_, _, code, err := configDiskStartup(root, device).connect(t.Context())
-	require.Error(t, err)
-	require.Equal(t, exitRefused, code)
-	got, rerr := os.ReadFile(device)
-	require.NoError(t, rerr)
-	require.Len(t, got, 9*512)
-	require.Equal(t, make([]byte, len(got)), got, "configuration disk erased")
-}
-
-// A disk that cannot be erased fails closed: the agent refuses, and the
-// extracted tree is kept so the seal check sees it.
-func TestConfigDiskStartup_UnerasableDiskRefuses(t *testing.T) {
-	dir := t.TempDir()
-	root := filepath.Join(dir, "cfg")
-	require.NoError(t, os.MkdirAll(filepath.Join(root, "by_name"), 0o700))
-
-	_, _, code, err := configDiskStartup(root, filepath.Join(dir, "missing")).connect(t.Context())
-	require.ErrorContains(t, err, "erase configuration disk")
-	require.Equal(t, exitRefused, code)
-	require.DirExists(t, root)
-}

cmd/lambda-agent/sdk_config_testlink.go

@@ -1,65 +0,0 @@
-//go:build lambdatestlink
-
-package main
-
-import (
-	"crypto/x509"
-	"errors"
-	"fmt"
-	"net"
-	"net/url"
-	"os"
-	"path/filepath"
-)
-
-const (
-	sdkEndpointEntry = "test-sdk-endpoint"
-	sdkCAEntry       = "test-sdk-ca"
-	sdkBundlePath    = "/run/lambda-sdk/ca-bundle.pem"
-	systemBundlePath = "/etc/ssl/certs/ca-certificates.crt"
-)
-
-type sdkConfig struct {
-	endpoint string
-	ca       []byte
-}
-
-// readSDKConfig reads optional public settings before the boot config is sealed.
-func readSDKConfig(dir string) (*sdkConfig, error) {
-	endpoint, endpointErr := os.ReadFile(filepath.Join(dir, sdkEndpointEntry, "raw"))
-	ca, caErr := os.ReadFile(filepath.Join(dir, sdkCAEntry, "raw"))
-	if errors.Is(endpointErr, os.ErrNotExist) && errors.Is(caErr, os.ErrNotExist) {
-		return nil, nil
-	}
-	if endpointErr != nil || caErr != nil {
-		return nil, errors.New("incomplete SDK gateway settings")
-	}
-	u, err := url.Parse(string(endpoint))
-	if err != nil || u.Scheme != "https" || u.User != nil || u.Path != "" || u.RawQuery != "" || u.Fragment != "" ||
-		net.ParseIP(u.Hostname()) == nil || u.Port() == "" {
-		return nil, errors.New("invalid SDK gateway endpoint")
-	}
-	if len(ca) > 64<<10 || !x509.NewCertPool().AppendCertsFromPEM(ca) {
-		return nil, errors.New("invalid SDK gateway CA")
-	}
-	return &sdkConfig{endpoint: string(endpoint), ca: ca}, nil
-}
-
-// writeBundle combines the system roots with the cluster's public root.
-func (c *sdkConfig) writeBundle(path, systemPath string) error {
-	system, err := os.ReadFile(systemPath)
-	if err != nil && !errors.Is(err, os.ErrNotExist) {
-		return fmt.Errorf("read system CA bundle: %w", err)
-	}
-	if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
-		return fmt.Errorf("create SDK CA directory: %w", err)
-	}
-	if err := os.Chmod(filepath.Dir(path), 0o755); err != nil {
-		return fmt.Errorf("make SDK CA directory readable: %w", err)
-	}
-	bundle := append(append(system, '\n'), c.ca...)
-	if err := os.WriteFile(path, bundle, 0o644); err != nil {
-		return fmt.Errorf("write SDK CA bundle: %w", err)
-	}
-	return os.Chmod(path, 0o644)
-}

cmd/lambda-agent/sdk_config_testlink_test.go

@@ -1,106 +0,0 @@
-//go:build lambdatestlink
-
-//test:in-package exercises the unexported boot configuration reader
-package main
-
-import (
-	"crypto/ecdsa"
-	"crypto/elliptic"
-	"crypto/rand"
-	"crypto/tls"
-	"crypto/x509"
-	"crypto/x509/pkix"
-	"encoding/pem"
-	"math/big"
-	"net/http"
-	"net/http/httptest"
-	"os"
-	"path/filepath"
-	"testing"
-	"time"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestSDKConfig_OptionalAndInvalid(t *testing.T) {
-	server := httptest.NewTLSServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) { w.WriteHeader(http.StatusOK) }))
-	defer server.Close()
-	ca := pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: server.Certificate().Raw})
-	for _, tc := range []struct {
-		name     string
-		endpoint []byte
-		ca       []byte
-		wantErr  bool
-		want     bool
-	}{
-		{name: "absent"},
-		{name: "valid", endpoint: []byte(server.URL), ca: ca, want: true},
-		{name: "missing ca", endpoint: []byte(server.URL), wantErr: true},
-		{name: "missing endpoint", ca: ca, wantErr: true},
-		{name: "invalid endpoint", endpoint: []byte("http://127.0.0.1:1234"), ca: ca, wantErr: true},
-		{name: "invalid ca", endpoint: []byte(server.URL), ca: []byte("not a certificate"), wantErr: true},
-	} {
-		t.Run(tc.name, func(t *testing.T) {
-			dir := t.TempDir()
-			for name, value := range map[string][]byte{sdkEndpointEntry: tc.endpoint, sdkCAEntry: tc.ca} {
-				if value != nil {
-					require.NoError(t, os.MkdirAll(filepath.Join(dir, name), 0o700))
-					require.NoError(t, os.WriteFile(filepath.Join(dir, name, "raw"), value, 0o600))
-				}
-			}
-			cfg, err := readSDKConfig(dir)
-			if tc.wantErr {
-				require.Error(t, err)
-				require.Nil(t, cfg)
-				return
-			}
-			require.NoError(t, err)
-			if !tc.want {
-				require.Nil(t, cfg)
-				return
-			}
-			bundle := filepath.Join(t.TempDir(), "sdk", "ca.pem")
-			require.NoError(t, cfg.writeBundle(bundle, filepath.Join(t.TempDir(), "absent")))
-			data, err := os.ReadFile(bundle)
-			require.NoError(t, err)
-			roots := x509.NewCertPool()
-			require.True(t, roots.AppendCertsFromPEM(data))
-			client := &http.Client{Transport: &http.Transport{TLSClientConfig: &tls.Config{RootCAs: roots}}}
-			resp, err := client.Get(cfg.endpoint)
-			require.NoError(t, err)
-			defer resp.Body.Close()
-			assert.Equal(t, http.StatusOK, resp.StatusCode)
-		})
-	}
-}
-
-func TestSDKConfig_BundleKeepsSystemRoots(t *testing.T) {
-	cluster := httptest.NewTLSServer(nil)
-	defer cluster.Close()
-	clusterCA := pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: cluster.Certificate().Raw})
-	key, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
-	require.NoError(t, err)
-	template := &x509.Certificate{SerialNumber: big.NewInt(1), Subject: pkix.Name{CommonName: "test system root"},
-		NotBefore: time.Now().Add(-time.Minute), NotAfter: time.Now().Add(time.Hour),
-		IsCA: true, BasicConstraintsValid: true, KeyUsage: x509.KeyUsageCertSign}
-	der, err := x509.CreateCertificate(rand.Reader, template, template, &key.PublicKey, key)
-	require.NoError(t, err)
-	systemCA := pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: der})
-	dir := t.TempDir()
-	systemPath, bundlePath := filepath.Join(dir, "system.pem"), filepath.Join(dir, "sdk", "ca.pem")
-	require.NoError(t, os.WriteFile(systemPath, systemCA, 0o600))
-	require.NoError(t, (&sdkConfig{ca: clusterCA}).writeBundle(bundlePath, systemPath))
-	bundle, err := os.ReadFile(bundlePath)
-	require.NoError(t, err)
-	roots := x509.NewCertPool()
-	require.True(t, roots.AppendCertsFromPEM(bundle))
-	client := &http.Client{Transport: &http.Transport{TLSClientConfig: &tls.Config{RootCAs: roots}}}
-	resp, err := client.Get(cluster.URL)
-	require.NoError(t, err)
-	require.NoError(t, resp.Body.Close())
-	systemCert, err := x509.ParseCertificate(der)
-	require.NoError(t, err)
-	_, err = systemCert.Verify(x509.VerifyOptions{Roots: roots})
-	require.NoError(t, err)
-}

cmd/lambda-agent/start_testlink.go

@@ -12,14 +12,11 @@ import (
 	"net/http"
 	"os"
 	"os/exec"
-	"sort"
-	"strings"
 	"time"
 
 	"github.com/mulgadc/spinifex/spinifex/agents/lambda/agent"
 	"github.com/mulgadc/spinifex/spinifex/agents/lambda/hostlink"
 	"github.com/mulgadc/spinifex/spinifex/agents/lambda/runtimeapi"
-	"github.com/mulgadc/spinifex/spinifex/foundation/aws/lambdaenv"
 )
 
 const (
@@ -34,17 +31,13 @@ type testlink struct {
 	pkg            taskPackage
 	runtimeAddr    string
 	startBootstrap func(ctx context.Context, env []string) (*exec.Cmd, error)
-	caBundlePath   string
-	systemCAPath   string
 }
 
 func start(ctx context.Context) int {
 	t := testlink{
-		startup:      defaultStartup(),
-		pkg:          devTaskPackage(),
-		runtimeAddr:  runtimeAddr,
-		caBundlePath: sdkBundlePath,
-		systemCAPath: systemBundlePath,
+		startup:     defaultStartup(),
+		pkg:         devTaskPackage(),
+		runtimeAddr: runtimeAddr,
 		startBootstrap: func(ctx context.Context, env []string) (*exec.Cmd, error) {
 			return agent.StartBootstrap(ctx, taskRoot+"/bootstrap", env, os.Stdout, os.Stderr)
 		},
@@ -60,7 +53,6 @@ func (t testlink) run(parent context.Context) (int, error) {
 	// The package values are fw_cfg entries, so read them before connect seals it.
 	spec, specErr := readPackageSpec(t.fwCfgDir)
 	grant, grantErr := readGrant(t.fwCfgDir)
-	sdk, sdkErr := readSDKConfig(t.fwCfgDir)
 	conn, scope, code, err := t.connect(parent)
 	if err != nil {
 		return code, errors.Join(err, specErr, grantErr)
@@ -93,18 +85,6 @@ func (t testlink) run(parent context.Context) (int, error) {
 	}
 	cred := *first.Credential
 	first.Credential = nil
-	if err := lambdaenv.Validate(cred.Environment); err != nil {
-		return exitRefused, errors.New("first frame carries an invalid function environment")
-	}
-	if sdkErr != nil {
-		slog.Warn("lambda-agent: SDK gateway settings unavailable", "err", sdkErr)
-	}
-	if sdk != nil {
-		if err := sdk.writeBundle(t.caBundlePath, t.systemCAPath); err != nil {
-			slog.Warn("lambda-agent: SDK gateway settings unavailable", "err", err)
-			sdk = nil
-		}
-	}
 	if err := t.pkg.install(spec); err != nil {
 		return exitRefused, fmt.Errorf("install package: %w", err)
 	}
@@ -120,8 +100,7 @@ func (t testlink) run(parent context.Context) (int, error) {
 
 	ctx, cancel := context.WithCancelCause(parent)
 	defer cancel(nil)
-	region, _, _ := strings.Cut(string(scope), "/")
-	cmd, err := t.startBootstrap(ctx, bootstrapEnv(ln.Addr().String(), spec.handler, cred, region, sdk, t.caBundlePath))
+	cmd, err := t.startBootstrap(ctx, bootstrapEnv(ln.Addr().String(), spec.handler, cred))
 	if err != nil {
 		return exitLink, fmt.Errorf("start bootstrap: %w", err)
 	}
@@ -145,35 +124,18 @@ func (t testlink) run(parent context.Context) (int, error) {
 	return exitClosed, nil
 }
 
-// bootstrapEnv supplies the runtime, execution-role and SDK settings.
-func bootstrapEnv(runtimeAPI, handler string, cred hostlink.Credential, region string, sdk *sdkConfig, caPath string) []string {
-	vars := map[string]string{
-		"AWS_LAMBDA_RUNTIME_API": runtimeAPI,
-		"_HANDLER":               handler,
-		"LAMBDA_TASK_ROOT":       taskRoot,
-		"PATH":                   "/bin:/sbin:/usr/bin",
-		"AWS_ACCESS_KEY_ID":      cred.AccessKeyID,
-		"AWS_SECRET_ACCESS_KEY":  cred.SecretAccessKey,
-		"AWS_SESSION_TOKEN":      cred.SessionToken,
-		"AWS_REGION":             region,
-		"AWS_DEFAULT_REGION":     region,
+// bootstrapEnv is the bootstrap's whole environment: the runtime variables and
+// the execution role's three credential variables.
+func bootstrapEnv(runtimeAPI, handler string, cred hostlink.Credential) []string {
+	return []string{
+		"AWS_LAMBDA_RUNTIME_API=" + runtimeAPI,
+		"_HANDLER=" + handler,
+		"LAMBDA_TASK_ROOT=" + taskRoot,
+		"PATH=/bin:/sbin:/usr/bin",
+		"AWS_ACCESS_KEY_ID=" + cred.AccessKeyID,
+		"AWS_SECRET_ACCESS_KEY=" + cred.SecretAccessKey,
+		"AWS_SESSION_TOKEN=" + cred.SessionToken,
 	}
-	if sdk != nil {
-		vars["AWS_ENDPOINT_URL"], vars["AWS_CA_BUNDLE"] = sdk.endpoint, caPath
-	}
-	for k, v := range cred.Environment {
-		vars[k] = v
-	}
-	keys := make([]string, 0, len(vars))
-	for k := range vars {
-		keys = append(keys, k)
-	}
-	sort.Strings(keys)
-	env := make([]string, 0, len(keys))
-	for _, k := range keys {
-		env = append(env, k+"="+vars[k])
-	}
-	return env
 }
 
 // shutdownGrace is how long past the invocation deadline the agent waits for

cmd/lambda-agent/start_testlink_test.go

@@ -7,10 +7,8 @@ import (
 	"context"
 	"crypto/sha256"
 	"encoding/hex"
-	"encoding/pem"
 	"io"
 	"net/http"
-	"net/http/httptest"
 	"os"
 	"os/exec"
 	"path/filepath"
@@ -72,7 +70,6 @@ type bootstrapStarts struct {
 	sealed  bool
 	// handlerOut receives the _HANDLER the bootstrap process itself saw.
 	handlerOut string
-	caBundle   string
 	// env is the environment the agent gave the bootstrap.
 	env []string
 }
@@ -97,7 +94,6 @@ func (b *bootstrapStarts) testlink(t *testing.T, entries map[string][]byte) test
 	b.started = make(chan struct{}, 1)
 	b.task = t.TempDir()
 	b.handlerOut = filepath.Join(t.TempDir(), "handler")
-	b.caBundle = filepath.Join(t.TempDir(), "sdk", "ca-bundle.pem")
 	return testlink{
 		startup: g.startup(),
 		pkg: taskPackage{
@@ -106,9 +102,7 @@ func (b *bootstrapStarts) testlink(t *testing.T, entries map[string][]byte) test
 			mountDir:   func(string) error { return nil },
 			sealDir:    func(string) error { b.sealed = true; return nil },
 		},
-		runtimeAddr:  "127.0.0.1:0",
-		caBundlePath: b.caBundle,
-		systemCAPath: filepath.Join(t.TempDir(), "system-ca.pem"),
+		runtimeAddr: "127.0.0.1:0",
 		startBootstrap: func(ctx context.Context, env []string) (*exec.Cmd, error) {
 			b.n++
 			b.env = env
@@ -182,87 +176,10 @@ func TestTestlinkServesOneInvocationThenStops(t *testing.T) {
 		"AWS_ACCESS_KEY_ID=" + testCredential.AccessKeyID,
 		"AWS_SECRET_ACCESS_KEY=" + testCredential.SecretAccessKey,
 		"AWS_SESSION_TOKEN=" + testCredential.SessionToken,
-		"AWS_REGION=ap-southeast-2",
-		"AWS_DEFAULT_REGION=ap-southeast-2",
 	}, b.env, "the bootstrap gets exactly the runtime and credential variables")
 	assert.NotEmpty(t, envValue(b.env, "AWS_LAMBDA_RUNTIME_API"))
 }
 
-func TestTestlinkProvidesSDKSettingsToBootstrap(t *testing.T) {
-	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
-	defer cancel()
-	server := httptest.NewTLSServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) { w.WriteHeader(http.StatusOK) }))
-	defer server.Close()
-	h := newHost(t)
-	h.accept(ctx)
-	entries := h.entries()
-	entries[sdkEndpointEntry] = []byte(server.URL)
-	entries[sdkCAEntry] = pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: server.Certificate().Raw})
-	var b bootstrapStarts
-	done := runTestlinkAsync(ctx, b.testlink(t, entries))
-	conn := h.guest(t)
-	defer conn.Close()
-	hs := hostlink.NewTestStream(conn)
-	req := hostlink.Request{Credential: &testCredential, GrantID: "g-1", AttemptID: "a-1", RequestID: "r-1", DeadlineMs: time.Now().Add(5 * time.Second).UnixMilli(), Payload: []byte(`{}`)}
-	require.NoError(t, hs.HostSend(req))
-	_, err := hs.HostResult(ctx)
-	require.NoError(t, err)
-	res := wait(t, done, "agent to stop")
-	require.NoError(t, res.err)
-	assert.Equal(t, server.URL, envValue(b.env, "AWS_ENDPOINT_URL"))
-	assert.Equal(t, b.caBundle, envValue(b.env, "AWS_CA_BUNDLE"))
-	assert.Equal(t, "ap-southeast-2", envValue(b.env, "AWS_REGION"))
-	assert.Equal(t, "ap-southeast-2", envValue(b.env, "AWS_DEFAULT_REGION"))
-	ca, err := os.ReadFile(b.caBundle)
-	require.NoError(t, err)
-	assert.Contains(t, string(ca), string(entries[sdkCAEntry]))
-	info, err := os.Stat(b.caBundle)
-	require.NoError(t, err)
-	assert.Equal(t, os.FileMode(0o644), info.Mode().Perm())
-}
-
-func TestTestlinkProvidesFunctionEnvironment(t *testing.T) {
-	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
-	defer cancel()
-	h := newHost(t)
-	h.accept(ctx)
-	var b bootstrapStarts
-	done := runTestlinkAsync(ctx, b.testlink(t, h.entries()))
-	conn := h.guest(t)
-	defer conn.Close()
-	hs := hostlink.NewTestStream(conn)
-	cred := testCredential
-	cred.Environment = map[string]string{"GREETING": "hello $WORLD", "PATH": "/my/bin"}
-	req := hostlink.Request{Credential: &cred, GrantID: "g-1", AttemptID: "a-1", RequestID: "r-1", DeadlineMs: time.Now().Add(5 * time.Second).UnixMilli(), Payload: []byte(`{}`)}
-	require.NoError(t, hs.HostSend(req))
-	_, err := hs.HostResult(ctx)
-	require.NoError(t, err)
-	require.NoError(t, wait(t, done, "agent to stop").err)
-	assert.Equal(t, "hello $WORLD", envValue(b.env, "GREETING"))
-	assert.Equal(t, "/my/bin", envValue(b.env, "PATH"))
-	assert.Equal(t, testCredential.SecretAccessKey, envValue(b.env, "AWS_SECRET_ACCESS_KEY"))
-}
-
-func TestTestlinkRejectsReservedFunctionEnvironment(t *testing.T) {
-	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
-	defer cancel()
-	h := newHost(t)
-	h.accept(ctx)
-	var b bootstrapStarts
-	done := runTestlinkAsync(ctx, b.testlink(t, h.entries()))
-	conn := h.guest(t)
-	defer conn.Close()
-	cred := testCredential
-	cred.Environment = map[string]string{"AWS_SECRET_ACCESS_KEY": "forged"}
-	req := hostlink.Request{Credential: &cred, GrantID: "g-1", AttemptID: "a-1", RequestID: "r-1", DeadlineMs: time.Now().Add(5 * time.Second).UnixMilli(), Payload: []byte(`{}`)}
-	require.NoError(t, hostlink.NewTestStream(conn).HostSend(req))
-	result := wait(t, done, "agent to refuse the environment")
-	assert.Equal(t, exitRefused, result.code)
-	assert.Error(t, result.err)
-	assert.NotContains(t, result.err.Error(), "forged")
-	assert.Zero(t, b.n)
-}
-
 var testCredential = hostlink.Credential{AccessKeyID: "ASIAGUESTKEY", SecretAccessKey: "guest-secret", SessionToken: "guest-token"}
 
 func envValue(env []string, name string) string {

contracts/cluster/v1/heartbeat.go

@@ -1,6 +1,9 @@
 package clusterv1
 
-import "time"
+import (
+	"encoding/json"
+	"time"
+)
 
 // ClusterStateBucket is the KV bucket holding each daemon's heartbeat alongside
 // other cluster-wide state.
@@ -11,6 +14,27 @@ func HeartbeatKey(node string) string {
 	return "heartbeat." + node
 }
 
+// LivenessWindow is how recently a node must have heartbeated for a reader
+// deciding on a takeover to count it live. It is six heartbeats, compared
+// across node clocks, where NTP skew is small against it.
+const LivenessWindow = 60 * time.Second
+
+// LiveAt reports whether a heartbeat written at written still vouches for its
+// node at now: it does while it is under LivenessWindow old.
+func LiveAt(written, now time.Time) bool {
+	return now.Sub(written) < LivenessWindow
+}
+
+// HeartbeatWritten dates a stored heartbeat value. One that cannot be decoded
+// or dated vouches for nothing, which is an answer rather than a fault.
+func HeartbeatWritten(data []byte) (time.Time, bool) {
+	var beat HeartbeatTimestamp
+	if json.Unmarshal(data, &beat) != nil {
+		return time.Time{}, false
+	}
+	return beat.Time()
+}
+
 // HeartbeatTimestamp is the read side of a heartbeat for node liveness. It
 // decodes only the timestamp, so a reader does not depend on the capacity
 // figures the daemon writes beside it.

contracts/cluster/v1/heartbeat_test.go

@@ -33,3 +33,23 @@ func TestHeartbeatTimestamp_UndatedBeatVouchesForNothing(t *testing.T) {
 		assert.False(t, ok, "timestamp %q", ts)
 	}
 }
+
+// A beat is live strictly under the window, so one exactly the window old is
+// the first a takeover may treat as stopped.
+func TestLiveAt_IsStrictlyUnderTheWindow(t *testing.T) {
+	now := time.Date(2026, 11, 5, 12, 0, 0, 0, time.UTC)
+	assert.True(t, clusterv1.LiveAt(now.Add(-clusterv1.LivenessWindow+time.Nanosecond), now))
+	assert.False(t, clusterv1.LiveAt(now.Add(-clusterv1.LivenessWindow), now))
+	assert.True(t, clusterv1.LiveAt(now.Add(time.Second), now), "a beat from a clock running ahead is live")
+}
+
+func TestHeartbeatWritten_DatesAStoredBeatAndRefusesAnUndecodableOne(t *testing.T) {
+	written, ok := clusterv1.HeartbeatWritten([]byte(`{"node":"node-a","timestamp":"2026-10-30T01:02:03Z"}`))
+	require.True(t, ok)
+	assert.True(t, written.Equal(time.Date(2026, 10, 30, 1, 2, 3, 0, time.UTC)))
+
+	for _, raw := range []string{"not json", `{"node":"node-a"}`} {
+		_, ok := clusterv1.HeartbeatWritten([]byte(raw))
+		assert.False(t, ok, raw)
+	}
+}

contracts/ec2/v1/commands.go

@@ -12,11 +12,6 @@ const (
 	// this contract family. It is the NATS permission/subscription pattern for
 	// the per-instance command route, rather than the broader descendant match.
 	InstanceCommandSubjectWildcard = InstanceCommandSubjectPrefix + "*"
-
-	// HandledHeader is set to "1" on every reply an owner sends to a
-	// modify-instance-metadata-options or set-disable-api-termination command.
-	// An owner that predates those commands cannot set it.
-	HandledHeader = "Spinifex-Ec2-Cmd-Handled"
 )
 
 // InstanceCommandSubject returns the targeted command subject for instanceID.
@@ -28,9 +23,7 @@ func InstanceCommandSubject(instanceID string) string {
 }
 
 // EC2InstanceCommand is the NATS wire format for EC2 instance commands
-// (stop, terminate, start, reboot, attach/detach-volume, attach/detach-eni,
-// IAM profile, spot lineage, tags, monitoring, metadata options and
-// termination protection).
+// (stop, terminate, start, attach/detach-volume, attach/detach-eni).
 type EC2InstanceCommand struct {
 	ID                        string                     `json:"id"`
 	Attributes                EC2CommandAttributes       `json:"attributes"`
@@ -42,9 +35,6 @@ type EC2InstanceCommand struct {
 	SpotLineageData           *SpotLineageData           `json:"spot_lineage_data,omitempty"`
 	InstanceTagsData          *InstanceTagsData          `json:"instance_tags_data,omitempty"`
 	InstanceMonitoringData    *InstanceMonitoringData    `json:"instance_monitoring_data,omitempty"`
-
-	InstanceMetadataOptionsData *InstanceMetadataOptionsData `json:"instance_metadata_options_data,omitempty"`
-	DisableApiTerminationData   *DisableApiTerminationData   `json:"disable_api_termination_data,omitempty"`
 }
 
 // EC2CommandAttributes indicates which action the daemon should perform.
@@ -62,9 +52,6 @@ type EC2CommandAttributes struct {
 	SetInstanceTags             bool `json:"set_instance_tags,omitempty"`
 	RemoveInstanceTags          bool `json:"remove_instance_tags,omitempty"`
 	SetInstanceMonitoring       bool `json:"set_instance_monitoring,omitempty"`
-
-	ModifyInstanceMetadataOptions bool `json:"modify_instance_metadata_options,omitempty"`
-	SetDisableApiTermination      bool `json:"set_disable_api_termination,omitempty"`
 }
 
 // AttachVolumeData carries parameters for an attach-volume command.
@@ -116,23 +103,6 @@ type InstanceMonitoringData struct {
 	Enabled bool `json:"enabled"`
 }
 
-// InstanceMetadataOptionsData carries the requested metadata options for a
-// modify-instance-metadata-options command. Unset fields are left unchanged.
-type InstanceMetadataOptionsData struct {
-	HttpTokens              string `json:"http_tokens,omitempty"`
-	HttpEndpoint            string `json:"http_endpoint,omitempty"`
-	HttpProtocolIpv6        string `json:"http_protocol_ipv6,omitempty"`
-	InstanceMetadataTags    string `json:"instance_metadata_tags,omitempty"`
-	HttpPutResponseHopLimit *int64 `json:"http_put_response_hop_limit,omitempty"`
-}
-
-// DisableApiTerminationData carries the desired termination-protection value
-// for a set-disable-api-termination command. Carried as data so "false" is an
-// explicit value, not the absence of one.
-type DisableApiTerminationData struct {
-	Value bool `json:"value"`
-}
-
 // SpotLineageData carries the SIR id stamped onto a spot-launched instance.
 // Lifecycle is always "spot" for this command, so only the SIR id travels.
 type SpotLineageData struct {

contracts/ec2/v1/commands_test.go

@@ -24,9 +24,6 @@ func TestEC2InstanceCommandJSONContract(t *testing.T) {
 			SetInstanceTags:             true,
 			RemoveInstanceTags:          true,
 			SetInstanceMonitoring:       true,
-
-			ModifyInstanceMetadataOptions: true,
-			SetDisableApiTermination:      true,
 		},
 		AttachVolumeData:          &AttachVolumeData{VolumeID: "vol-1", Device: "/dev/sdf"},
 		DetachVolumeData:          &DetachVolumeData{VolumeID: "vol-2", Device: "/dev/sdg", Force: true},
@@ -36,15 +33,6 @@ func TestEC2InstanceCommandJSONContract(t *testing.T) {
 		SpotLineageData:           &SpotLineageData{SpotInstanceRequestId: "sir-1"},
 		InstanceTagsData:          &InstanceTagsData{Tags: map[string]string{"Name": "example"}, TagKeys: []string{"old"}},
 		InstanceMonitoringData:    &InstanceMonitoringData{Enabled: true},
-
-		InstanceMetadataOptionsData: &InstanceMetadataOptionsData{
-			HttpTokens:              "required",
-			HttpEndpoint:            "enabled",
-			HttpProtocolIpv6:        "disabled",
-			InstanceMetadataTags:    "enabled",
-			HttpPutResponseHopLimit: int64Ptr(3),
-		},
-		DisableApiTerminationData: &DisableApiTerminationData{Value: false},
 	}
 
 	data, err := json.Marshal(command)
@@ -64,9 +52,7 @@ func TestEC2InstanceCommandJSONContract(t *testing.T) {
 			"set_spot_lineage": true,
 			"set_instance_tags": true,
 			"remove_instance_tags": true,
-			"set_instance_monitoring": true,
-			"modify_instance_metadata_options": true,
-			"set_disable_api_termination": true
+			"set_instance_monitoring": true
 		},
 		"attach_volume_data": {"volume_id": "vol-1", "device": "/dev/sdf"},
 		"detach_volume_data": {"volume_id": "vol-2", "device": "/dev/sdg", "force": true},
@@ -75,62 +61,8 @@ func TestEC2InstanceCommandJSONContract(t *testing.T) {
 		"iam_profile_association_data": {"instance_profile_arn": "arn:aws:iam::123456789012:instance-profile/example"},
 		"spot_lineage_data": {"spot_instance_request_id": "sir-1"},
 		"instance_tags_data": {"tags": {"Name": "example"}, "tag_keys": ["old"]},
-		"instance_monitoring_data": {"enabled": true},
-		"instance_metadata_options_data": {
-			"http_tokens": "required",
-			"http_endpoint": "enabled",
-			"http_protocol_ipv6": "disabled",
-			"instance_metadata_tags": "enabled",
-			"http_put_response_hop_limit": 3
-		},
-		"disable_api_termination_data": {"value": false}
-	}`, string(data))
-}
-
-func int64Ptr(v int64) *int64 { return &v }
-
-// An owner that predates the metadata and termination-protection commands
-// sends and receives payloads without the new fields.
-func TestEC2InstanceCommandNewFieldsAreOptionalOnTheWire(t *testing.T) {
-	var n1 EC2InstanceCommand
-	require.NoError(t, json.Unmarshal([]byte(`{
-		"id": "i-0123456789abcdef0",
-		"attributes": {"stop_instance": true},
 		"instance_monitoring_data": {"enabled": true}
-	}`), &n1))
-	require.Equal(t, "i-0123456789abcdef0", n1.ID)
-	require.True(t, n1.Attributes.StopInstance)
-	require.False(t, n1.Attributes.ModifyInstanceMetadataOptions)
-	require.False(t, n1.Attributes.SetDisableApiTermination)
-	require.Nil(t, n1.InstanceMetadataOptionsData)
-	require.Nil(t, n1.DisableApiTerminationData)
-
-	unset, err := json.Marshal(EC2InstanceCommand{ID: "i-1", Attributes: EC2CommandAttributes{StopInstance: true}})
-	require.NoError(t, err)
-	require.NotContains(t, string(unset), "modify_instance_metadata_options")
-	require.NotContains(t, string(unset), "set_disable_api_termination")
-	require.NotContains(t, string(unset), "instance_metadata_options_data")
-	require.NotContains(t, string(unset), "disable_api_termination_data")
-}
-
-func TestInstanceMetadataOptionsDataOmitsUnsetFields(t *testing.T) {
-	data, err := json.Marshal(InstanceMetadataOptionsData{HttpPutResponseHopLimit: int64Ptr(1)})
-	require.NoError(t, err)
-	require.JSONEq(t, `{"http_put_response_hop_limit": 1}`, string(data))
-
-	data, err = json.Marshal(InstanceMetadataOptionsData{})
-	require.NoError(t, err)
-	require.JSONEq(t, `{}`, string(data))
-}
-
-func TestDisableApiTerminationDataKeepsExplicitFalse(t *testing.T) {
-	data, err := json.Marshal(DisableApiTerminationData{Value: false})
-	require.NoError(t, err)
-	require.JSONEq(t, `{"value": false}`, string(data))
-}
-
-func TestHandledHeaderName(t *testing.T) {
-	require.Equal(t, "Spinifex-Ec2-Cmd-Handled", HandledHeader)
+	}`, string(data))
 }
 
 func TestInstanceCommandSubjects(t *testing.T) {

contracts/lambda/v1/README.md

@@ -21,16 +21,8 @@ Tenant code may be able to reach the listener, but it cannot open the agent's co
 **Authentication:** mutual TLS 1.3, bound to the environment ID and realization generation, using material the user runtime cannot read (ADR-0009:S17).
 For each generation the executor creates ephemeral self-signed host and guest certificates whose single URI SAN is `spinifex-lambda://<Region>/<account>/<function ID>/<environment ID>/<generation>/<role>`, with role `host` or `guest`.
 Each end validates the peer certificate against the one pinned peer certificate and separately requires that URI SAN with the opposite role; hostname verification is not relied on.
-ALPN carries the L-GUEST major version, `spinifex-lambda-guest/1`, and an unsupported version fails the handshake. Builds with the `lambdatestlink` tag offer only `spinifex-lambda-testlink/3` on both host and guest, whose first frame carries the execution-role credential and function environment, so a test-link end and a release end fail the handshake; release builds are unchanged.
-In the same builds, the gateway reaches the host daemon on the local abstract socket `spinifex-lambda-testlink-control/4`, and each request carries the runtime-image digest the gateway was built with; the daemon refuses a missing or different digest before starting the attempt, and a gateway and daemon on different socket names never connect.
-On that socket a request's raw payload is at most 6 MiB, checked by the gateway before encoding and by the daemon after decoding.
-Its encoded frame is at most the base64 length of 6 MiB plus 64 KiB of metadata headroom and 64 bytes of field slack, and a receiver checks the frame length before allocating; metadata has no separate bound within the frame.
-The credential and environment travel as their own frame of at most 40 KiB, and the reply has the same payload and frame bounds as the request.
-The daemon holds at most two connections at once, from accept to close, and leaves the rest unaccepted in the listen backlog.
+ALPN carries the L-GUEST major version, `spinifex-lambda-guest/1`, and an unsupported version fails the handshake. Builds with the `lambdatestlink` tag offer only `spinifex-lambda-testlink/2` on both host and guest, whose first frame carries the execution-role credential, so a test-link end and a release end fail the handshake; release builds are unchanged.
 Under TLS 1.3 the guest's handshake can complete before the host refuses its certificate, so the guest treats the connection as accepted only after the host's first message.
-In the same builds, the function owner moves a created function with stored code from `Pending` to `Active` once this node's daemon serves the pinned image and the cluster configuration has exactly one node; a code copy over the test-link bound stays `Pending`, and `StateReason` names the unmet condition.
-Release builds never leave `Pending`.
-The Mulga deploy role installs a test-link build with `deploy_lambda_testlink=true`; this profile does not meet Q-199 fail-closed revocation, and its guests take management addresses whose record can expire (mulga-0hsub).
 
 **Bootstrap material:** the executor passes these fw_cfg entries under `opt/spinifex/lambda/`, and the agent reads them as root and unloads `qemu_fw_cfg` before user code starts.
 
@@ -44,8 +36,6 @@ The Mulga deploy role installs a test-link build with `deploy_lambda_testlink=tr
 
 The guest's own interface settings come from `opt/spinifex/netcfg`, the entry other system VMs use.
 The agent unloads `qemu_fw_cfg` even when an entry is missing, and refuses to start if the fw_cfg tree is still readable afterwards.
-On a VMM without fw_cfg (the test-link Firecracker executor), the executor attaches the same entries as a tar archive on the second, writable virtio-blk disk, laid out as `by_name/<entry>/raw`.
-The runtime image's init extracts it to a root-only tmpfs at `/spinifex-cfg`, and the agent reads the entries there, then overwrites the whole disk with zeros and removes the tree; it refuses to start if the tree is still present.
 Until the message format exists, the agent authenticates and holds the connection but accepts no host message and starts no user code.
 
 **Authority:** one environment's attempt only.

docs/DESIGN.md

@@ -111,9 +111,9 @@ Daemons (`spinifex/runtime/daemon/daemon.go`) subscribe to NATS topics and handl
 When a daemon handles `RunInstances`:
 
 1. **Resource Check**: Validates CPU/memory availability via `ResourceManager`
-2. **Volume Generation**: Creates boot and EFI volumes via Viperblock
-   A UEFI instance whose record carries a variable store pointer has no EFI volume: its `vars.fd` is fetched from the object under `uefi/` in the configured Predastore bucket that the pointer names, recorded in the `spinifex-uefi-stores` JetStream bucket (or reused when `seeded.json` on this node already names the same object) after a firmware architecture and VARS size check, and a record with both or neither is refused.
-   RunInstances still creates the EFI volume, nothing writes `vars.fd` back to the object store yet, and a terminated instance's store is removed by a cluster-wide reaper.
+2. **Volume Generation**: Creates the boot volume through the EBS provider and, for a UEFI guest, seeds its variable store as an object under `uefi/<instance-id>/` with a store record in `spinifex-uefi-stores` that points at it
+   At launch a UEFI instance's `vars.fd` is fetched from the object its store pointer names (or reused when `seeded.json` on this node already names the same object) after a firmware architecture and VARS size check, and a UEFI instance with no live store record naming a pointer, or one that still carries an EFI volume request, is refused.
+   The store is written back to the object store on write and at stop, and a terminated instance's store is removed by a cluster-wide reaper.
 3. **Volume Mount**: Publishes each volume through the EBS provider contract (`PublishVolume` on `ebs.provider.v1.{node}.mount`), receives its export
 4. **QEMU Launch**: Builds and executes QEMU command with NBD-backed disks
 5. **QMP Monitoring**: Establishes QEMU Machine Protocol connection for VM management

docs/testing/e2e-coverage.md

@@ -108,8 +108,6 @@ Three entry points, in the order they run:
 
 **Its own suite and its own cell, for instancerecovery's reasons plus one.** It takes a node away, so anything sharing the environment reports that outage as its own failure; it needs its own VPC, public subnet and a backend placement no shared fixture describes; and a failure here has to read as "the load balancer did not survive its host" rather than as one more red subtest in a suite about load balancer features.
 
-**A dead data NIC is the other half, and there the state does change.** `LBSelfHealsADeadDataNIC` disables one data NIC of a running load balancer guest through a second QMP client and runs for an internet-facing NLB and an internal ALB. The backends and an unrelated customer instance in the same subnet are launched more than 21 minutes before the fault, so a repair that reached a customer instance would find one the platform has long stopped expecting to change. A state watch samples `DescribeLoadBalancers` every two seconds and requires the sequence `active`, `active_impaired`, `active` with nothing in between, while the internet-facing run also watches the name continuously and serves traffic through it within five minutes. Afterwards the guest, its interfaces and its address are unchanged but its process is not, the backends and the bystander keep their instance ID, launch time and process (ADR-0003:S7), and the daemon journals exactly one `ELBv2 data-path repair started` line carrying the fields an operator needs. The internal ALB has no client inside the VPC, so it asserts the state sequence and target health rather than traffic through its name.
-
 ### `iam` — everything IAM and STS
 
 Policy evaluation against live requests rather than unit-level: a deny applying to the next request without a new credential, resource-scoped grants naming one key pair, a user rename needing `iam:UpdateUser` on both the old and the new name, `aws:SourceIp` seeing the real client address, assumed-role sessions carrying role policies, and revocation stopping one key alone.
@@ -131,7 +129,7 @@ The client reports on its serial console because the runner has no route into th
 
 ### `eks`, `ecs`, `ecr` — container services
 
-`eks`: cluster create/delete, access entries, addon delivery, `GetToken`, kubeconfig artifacts, the EBS CSI volume path, an IRSA web-identity exchange under the stock AWS trust policy (`StringEquals` on the issuer host-path `:sub`/`:aud`, a mismatched `:sub` denied with `AccessDenied`, a `StringLike` `:sub` granted), and an IRSA pod whose spec names only a ServiceAccount with the role-arn annotation, where the pod-identity webhook injects the token volume, `AWS_*` environment, STS endpoint and CA bundle and the pod assumes the role, while a pod of an unannotated ServiceAccount is left untouched. The unmodified aws-load-balancer-controller chart 1.11.0 installs with the values an AWS user passes plus the gateway endpoint override, the Shield and WAF switches and the single-subnet feature gate, and reconciles an Ingress into an active ALB. Worker IMDS is covered too: a managed nodegroup worker reports IMDSv2 required with hop limit 2, a pod with no IRSA resolves to the node role through IMDS, and a nodegroup built from a launch template takes the template's metadata options. The private-endpoint NIC of the cluster's endpoint NLB is disabled through QMP and every worker must be Ready again within ten minutes while the cluster stays `ACTIVE` and the NLB keeps its guest instance. Control-plane isolation is covered as well: no control-plane host registers as a Node, a DaemonSet tolerating every taint runs on workers only, metrics-server is Ready, and logs work for every kube-system pod. `ecs`: clusters, container instances, the three network modes (`awsvpc`, `bridge`, `host`), task-role credentials fetched from the container credentials endpoint, a service behind an ELB, and ENI release on terminate. `ecr`: control-plane CRUD plus a real OCI data plane — docker login/build/push/pull round trips, multi-layer images, and raw blob HTTP.
+`eks`: cluster create/delete, access entries, addon delivery, `GetToken`, kubeconfig artifacts, the EBS CSI volume path, an IRSA web-identity exchange under the stock AWS trust policy (`StringEquals` on the issuer host-path `:sub`/`:aud`, a mismatched `:sub` denied with `AccessDenied`, a `StringLike` `:sub` granted), and an IRSA pod whose spec names only a ServiceAccount with the role-arn annotation, where the pod-identity webhook injects the token volume, `AWS_*` environment, STS endpoint and CA bundle and the pod assumes the role, while a pod of an unannotated ServiceAccount is left untouched. The unmodified aws-load-balancer-controller chart 1.11.0 installs with the values an AWS user passes plus the gateway endpoint override, the Shield and WAF switches and the single-subnet feature gate, and reconciles an Ingress into an active ALB. Worker IMDS is covered too: a managed nodegroup worker reports IMDSv2 required with hop limit 2, a pod with no IRSA resolves to the node role through IMDS, and a nodegroup built from a launch template takes the template's metadata options. Control-plane isolation is covered as well: no control-plane host registers as a Node, a DaemonSet tolerating every taint runs on workers only, metrics-server is Ready, and logs work for every kube-system pod. `ecs`: clusters, container instances, the three network modes (`awsvpc`, `bridge`, `host`), task-role credentials fetched from the container credentials endpoint, a service behind an ELB, and ENI release on terminate. `ecr`: control-plane CRUD plus a real OCI data plane — docker login/build/push/pull round trips, multi-layer images, and raw blob HTTP.
 
 ### `rds` — managed databases
 

scripts/build-microvm-image.sh

@@ -209,7 +209,6 @@ echo "[build-microvm-image] nginx: binary + stream module present"
 #   virtio*.ko*    — virtio bus, net, blk, rng, console, and helpers
 #   *fw_cfg*.ko*   — QEMU fw_cfg sysfs driver (delivers boot blobs)
 #   *failover*.ko* — net_failover + failover (virtio_net deps for live migration)
-#   af_packet.ko*  — AF_PACKET sockets (lb-agent's per-NIC gateway ARP probe)
 # All other loadable modules are removed. Drivers compiled into the kernel
 # (CONFIG_VIRTIO_MMIO=y etc.) are unaffected — they live in vmlinuz, not here.
 echo "[build-microvm-image] stripping kernel modules to virtio+fw_cfg only..."
@@ -223,7 +222,6 @@ for kver_dir in "$CHROOT_DIR/lib/modules"/*/; do
     find "$kernel_dir" -name "virtio*.ko.gz" -o -name "virtio*.ko" \
                        -o -name "*fw_cfg*.ko.gz" -o -name "*fw_cfg*.ko" \
                        -o -name "*failover*.ko.gz" -o -name "*failover*.ko" \
-                       -o -name "af_packet.ko.gz" -o -name "af_packet.ko" \
         2>/dev/null | while read -r mod; do
         rel="${mod#"$kernel_dir/"}"
         dest_dir="$tmp_keep/$(dirname "$rel")"
@@ -242,15 +240,6 @@ for kver_dir in "$CHROOT_DIR/lib/modules"/*/; do
     done)
     rm -rf "$tmp_keep"
 
-    # lb-agent's data-path probe needs AF_PACKET; without it every LB would
-    # report a dead NIC. Accept a kernel with CONFIG_PACKET built in.
-    if ! find "$kernel_dir" -name "af_packet.ko*" | grep -q .; then
-        if ! grep -q "CONFIG_PACKET=y" "$KERNEL_CONFIG" 2>/dev/null; then
-            echo "ERROR: af_packet.ko missing from the kept modules — lb-agent's data-path probe cannot open its socket" >&2
-            exit 1
-        fi
-    fi
-
     # Decompress kept modules to plain .ko. Alpine ships .ko.gz, but a host
     # kmod that lacks gzip support (e.g. Debian trixie's kmod 34) silently
     # produces an EMPTY modules.dep from compressed modules — the guest's

spinifex/agents/elbv2/agent.go

@@ -21,7 +21,6 @@ import (
 	"path/filepath"
 	"strconv"
 	"strings"
-	"sync"
 	"syscall"
 	"time"
 
@@ -95,18 +94,6 @@ type Agent struct {
 	statsFn func(socketPath string) ([]ServerStatus, error)
 	// For testing: override the active health prober (nginx/NLB).
 	probeFn func(targets []HealthTarget) []ServerStatus
-
-	// Data-path probe: netcfg and sysfs locations, the ARP exchange, the
-	// instance ID lookup and per-NIC miss counters (overridable in tests).
-	netcfgPath   string
-	sysNetDir    string
-	probeWait    time.Duration
-	arpProbeFn   arpProbeFn
-	instanceIDFn func() (string, error)
-	instanceID   string
-	idSkip       int
-	probeMu      sync.Mutex
-	probeStates  map[string]*nicProbeState
 }
 
 // New creates a new LB agent for the given load balancer.
@@ -163,12 +150,6 @@ func New(lbID, gatewayURL, accessKey, secretKey, region string) (*Agent, error)
 		reloadNginxFn: reloadNginx,
 		statsFn:       queryHAProxyStats,
 		probeFn:       probeHealthTargets,
-		netcfgPath:    netcfgPath,
-		sysNetDir:     sysClassNet,
-		probeWait:     probeWait,
-		arpProbeFn:    arpProbe,
-		instanceIDFn:  fetchInstanceID,
-		probeStates:   map[string]*nicProbeState{},
 	}, nil
 }
 
@@ -226,7 +207,7 @@ func (a *Agent) tick() {
 		}
 	}
 
-	resp, err := a.sendHeartbeat(servers, a.dataPathEvidence())
+	resp, err := a.sendHeartbeat(servers)
 	if err != nil {
 		// Include the dial target so the serial console disambiguates a
 		// transport failure (never reached AWSGW — "send request: dial ...")
@@ -307,7 +288,7 @@ func (a *Agent) enginePaths(engine string) (configPath, pidPath, certDir string,
 }
 
 // sendHeartbeat sends a heartbeat with health report to the gateway.
-func (a *Agent) sendHeartbeat(servers []ServerStatus, ev dataPathEvidence) (*heartbeatResponse, error) {
+func (a *Agent) sendHeartbeat(servers []ServerStatus) (*heartbeatResponse, error) {
 	params := url.Values{}
 	params.Set("Action", "LBAgentHeartbeat")
 	params.Set("Version", "2015-12-01")
@@ -319,7 +300,6 @@ func (a *Agent) sendHeartbeat(servers []ServerStatus, ev dataPathEvidence) (*hea
 		params.Set("Servers.member."+idx+".Server", s.Server)
 		params.Set("Servers.member."+idx+".Status", s.Status)
 	}
-	ev.addTo(params)
 
 	body, err := a.signedPost(params)
 	if err != nil {

spinifex/agents/elbv2/arp_linux.go

@@ -1,62 +0,0 @@
-//go:build linux
-
-package elbv2
-
-import (
-	"errors"
-	"fmt"
-	"net"
-	"net/netip"
-	"syscall"
-	"time"
-)
-
-// arpProbe sends one ARP request for gateway out of iface on an AF_PACKET
-// socket bound to it and waits for the gateway's reply on the same socket.
-func arpProbe(iface string, mac net.HardwareAddr, src, gateway netip.Addr, wait time.Duration) error {
-	ifi, err := net.InterfaceByName(iface)
-	if err != nil {
-		return fmt.Errorf("interface %s: %w", iface, err)
-	}
-
-	proto := htons(syscall.ETH_P_ARP)
-	fd, err := syscall.Socket(syscall.AF_PACKET, syscall.SOCK_RAW, int(proto))
-	if err != nil {
-		return fmt.Errorf("packet socket: %w", err)
-	}
-	defer syscall.Close(fd)
-
-	addr := &syscall.SockaddrLinklayer{Protocol: proto, Ifindex: ifi.Index}
-	if err := syscall.Bind(fd, addr); err != nil {
-		return fmt.Errorf("bind %s: %w", iface, err)
-	}
-
-	deadline := time.Now().Add(wait)
-	if err := syscall.Sendto(fd, buildARPRequest(mac, src, gateway), 0, addr); err != nil {
-		return fmt.Errorf("send ARP request: %w", err)
-	}
-
-	buf := make([]byte, 128)
-	for {
-		left := time.Until(deadline)
-		if left <= 0 {
-			return errARPNoReply
-		}
-		tv := syscall.NsecToTimeval(int64(left))
-		if err := syscall.SetsockoptTimeval(fd, syscall.SOL_SOCKET, syscall.SO_RCVTIMEO, &tv); err != nil {
-			return fmt.Errorf("set receive deadline: %w", err)
-		}
-		n, _, err := syscall.Recvfrom(fd, buf, 0)
-		if err != nil {
-			if errors.Is(err, syscall.EAGAIN) || errors.Is(err, syscall.EINTR) {
-				continue
-			}
-			return fmt.Errorf("receive ARP reply: %w", err)
-		}
-		if isARPReplyFrom(buf[:n], mac, gateway) {
-			return nil
-		}
-	}
-}
-
-func htons(v uint16) uint16 { return v<<8 | v>>8 }

spinifex/agents/elbv2/arp_other.go

@@ -1,15 +0,0 @@
-//go:build !linux
-
-package elbv2
-
-import (
-	"errors"
-	"net"
-	"net/netip"
-	"time"
-)
-
-// arpProbe is only meaningful inside the Linux guest.
-func arpProbe(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error {
-	return errors.New("ARP probe is only supported on linux")
-}

spinifex/agents/elbv2/datapath.go

@@ -1,303 +0,0 @@
-package elbv2
-
-import (
-	"bufio"
-	"bytes"
-	"encoding/binary"
-	"errors"
-	"fmt"
-	"net"
-	"net/netip"
-	"os"
-	"path/filepath"
-	"sort"
-	"strconv"
-	"strings"
-	"sync"
-	"time"
-)
-
-const (
-	// netcfgPath is the fw_cfg blob the microVM init script reads to configure NICs.
-	netcfgPath = "/sys/firmware/qemu_fw_cfg/by_name/opt/spinifex/netcfg/raw"
-
-	sysClassNet = "/sys/class/net"
-
-	// probeFailureThreshold is the number of consecutive missed probes after
-	// which a NIC is reported as failing (about 15 s at the poll interval).
-	probeFailureThreshold = 3
-
-	// arpExchangeWait bounds one ARP exchange. probeWait bounds how long a tick
-	// waits for all probes, so a stuck probe never delays the heartbeat.
-	arpExchangeWait = time.Second
-	probeWait       = 1500 * time.Millisecond
-)
-
-// nicConfig is one NIC of the netcfg blob that has a gateway to probe.
-type nicConfig struct {
-	MAC     string
-	Addr    netip.Addr
-	Gateway netip.Addr
-}
-
-// NICReport is the per-NIC data-path evidence carried in the heartbeat.
-type NICReport struct {
-	MAC                 string
-	ProbeOK             bool
-	ConsecutiveFailures int64
-	RxPackets           int64
-	TxPackets           int64
-	RxDropped           int64
-	TxDropped           int64
-}
-
-// arpProbeFn sends one ARP request for gateway out of iface and returns nil
-// when the gateway's reply arrives within timeout.
-type arpProbeFn func(iface string, mac net.HardwareAddr, src, gateway netip.Addr, timeout time.Duration) error
-
-// parseNetcfg reads the KEY=value netcfg blob and returns the NICs that have a
-// gateway, in NIC order. The mgmt NIC carries no gateway and is skipped, as
-// is any NIC without a usable MAC, address or gateway.
-func parseNetcfg(blob []byte) []nicConfig {
-	type raw struct{ mac, cidr, gw string }
-	byIndex := map[int]*raw{}
-
-	sc := bufio.NewScanner(bytes.NewReader(blob))
-	for sc.Scan() {
-		line := strings.TrimSpace(sc.Text())
-		key, val, ok := strings.Cut(line, "=")
-		if !ok || !strings.HasPrefix(key, "NIC") {
-			continue
-		}
-		idxStr, field, ok := strings.Cut(strings.TrimPrefix(key, "NIC"), "_")
-		if !ok {
-			continue
-		}
-		idx, err := strconv.Atoi(idxStr)
-		if err != nil {
-			continue
-		}
-		r := byIndex[idx]
-		if r == nil {
-			r = &raw{}
-			byIndex[idx] = r
-		}
-		switch field {
-		case "MAC":
-			r.mac = strings.ToLower(val)
-		case "CIDR":
-			r.cidr = val
-		case "GW":
-			r.gw = val
-		}
-	}
-
-	idxs := make([]int, 0, len(byIndex))
-	for i := range byIndex {
-		idxs = append(idxs, i)
-	}
-	sort.Ints(idxs)
-
-	var out []nicConfig
-	for _, i := range idxs {
-		r := byIndex[i]
-		if r.mac == "" || r.cidr == "" || r.gw == "" {
-			continue
-		}
-		prefix, err := netip.ParsePrefix(r.cidr)
-		if err != nil || !prefix.Addr().Is4() {
-			continue
-		}
-		gw, err := netip.ParseAddr(r.gw)
-		if err != nil || !gw.Is4() {
-			continue
-		}
-		out = append(out, nicConfig{MAC: r.mac, Addr: prefix.Addr(), Gateway: gw})
-	}
-	return out
-}
-
-// ifaceByMAC returns the interface under sysNetDir whose address is mac.
-func ifaceByMAC(sysNetDir, mac string) (string, error) {
-	entries, err := os.ReadDir(sysNetDir)
-	if err != nil {
-		return "", err
-	}
-	for _, e := range entries {
-		b, err := os.ReadFile(filepath.Join(sysNetDir, e.Name(), "address"))
-		if err != nil {
-			continue
-		}
-		if strings.EqualFold(strings.TrimSpace(string(b)), mac) {
-			return e.Name(), nil
-		}
-	}
-	return "", fmt.Errorf("no interface with MAC %s", mac)
-}
-
-// readCounters fills the sysfs statistics of iface into r.
-func readCounters(sysNetDir, iface string, r *NICReport) {
-	read := func(name string) int64 {
-		b, err := os.ReadFile(filepath.Join(sysNetDir, iface, "statistics", name))
-		if err != nil {
-			return 0
-		}
-		v, _ := strconv.ParseInt(strings.TrimSpace(string(b)), 10, 64)
-		return v
-	}
-	r.RxPackets = read("rx_packets")
-	r.TxPackets = read("tx_packets")
-	r.RxDropped = read("rx_dropped")
-	r.TxDropped = read("tx_dropped")
-}
-
-var errARPNoReply = errors.New("no ARP reply from the gateway")
-
-// buildARPRequest returns the Ethernet frame asking who has target.
-func buildARPRequest(mac net.HardwareAddr, src, target netip.Addr) []byte {
-	f := make([]byte, 0, 42)
-	f = append(f, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff)
-	f = append(f, mac...)
-	f = binary.BigEndian.AppendUint16(f, 0x0806)
-	f = binary.BigEndian.AppendUint16(f, 1)      // Ethernet
-	f = binary.BigEndian.AppendUint16(f, 0x0800) // IPv4
-	f = append(f, 6, 4)
-	f = binary.BigEndian.AppendUint16(f, 1) // request
-	f = append(f, mac...)
-	s := src.As4()
-	f = append(f, s[:]...)
-	f = append(f, 0, 0, 0, 0, 0, 0)
-	t := target.As4()
-	f = append(f, t[:]...)
-	return f
-}
-
-// isARPReplyFrom reports whether frame is an ARP reply sent by sender to mac.
-func isARPReplyFrom(frame []byte, mac net.HardwareAddr, sender netip.Addr) bool {
-	if len(frame) < 42 || binary.BigEndian.Uint16(frame[12:14]) != 0x0806 {
-		return false
-	}
-	if binary.BigEndian.Uint16(frame[20:22]) != 2 || !bytes.Equal(frame[32:38], mac) {
-		return false
-	}
-	s := sender.As4()
-	return bytes.Equal(frame[28:32], s[:])
-}
-
-// nicProbeState carries the miss counter of one NIC across ticks. A probe that
-// has not returned is still in flight, and counts as a miss without starting
-// another goroutine behind it.
-type nicProbeState struct {
-	mu       sync.Mutex
-	failures int64
-	inflight bool
-}
-
-type probeResult struct {
-	ok  bool
-	set bool
-}
-
-// probeNICs probes every gateway-bearing data NIC once, in parallel, and
-// returns within probeWait however long a probe blocks.
-func (a *Agent) probeNICs() []NICReport {
-	blob, err := os.ReadFile(a.netcfgPath)
-	if err != nil {
-		return nil
-	}
-	nics := parseNetcfg(blob)
-	if len(nics) == 0 {
-		return nil
-	}
-
-	reports := make([]NICReport, len(nics))
-	results := make([]chan probeResult, len(nics))
-	for i, n := range nics {
-		results[i] = make(chan probeResult, 1)
-		st := a.probeState(n.MAC)
-
-		st.mu.Lock()
-		busy := st.inflight
-		if !busy {
-			st.inflight = true
-		}
-		st.mu.Unlock()
-		if busy {
-			results[i] <- probeResult{set: true}
-			continue
-		}
-
-		go func(n nicConfig, st *nicProbeState, out chan<- probeResult) {
-			ok := a.probeOne(n)
-			st.mu.Lock()
-			st.inflight = false
-			st.mu.Unlock()
-			out <- probeResult{ok: ok, set: true}
-		}(n, st, results[i])
-	}
-
-	deadline := time.NewTimer(a.probeWait)
-	defer deadline.Stop()
-	expired := false
-	for i, n := range nics {
-		var res probeResult
-		if expired {
-			select {
-			case res = <-results[i]:
-			default:
-			}
-		} else {
-			select {
-			case res = <-results[i]:
-			case <-deadline.C:
-				// Out of time: an unfinished probe is a miss for this tick.
-				expired = true
-			}
-		}
-
-		st := a.probeState(n.MAC)
-		st.mu.Lock()
-		if res.ok {
-			st.failures = 0
-		} else {
-			st.failures++
-		}
-		failures := st.failures
-		st.mu.Unlock()
-
-		r := NICReport{
-			MAC:                 n.MAC,
-			ProbeOK:             failures < probeFailureThreshold,
-			ConsecutiveFailures: failures,
-		}
-		if iface, err := ifaceByMAC(a.sysNetDir, n.MAC); err == nil {
-			readCounters(a.sysNetDir, iface, &r)
-		}
-		reports[i] = r
-	}
-	return reports
-}
-
-// probeOne runs one ARP exchange for n, false on any failure.
-func (a *Agent) probeOne(n nicConfig) bool {
-	iface, err := ifaceByMAC(a.sysNetDir, n.MAC)
-	if err != nil {
-		return false
-	}
-	hw, err := net.ParseMAC(n.MAC)
-	if err != nil {
-		return false
-	}
-	return a.arpProbeFn(iface, hw, n.Addr, n.Gateway, arpExchangeWait) == nil
-}
-
-func (a *Agent) probeState(mac string) *nicProbeState {
-	a.probeMu.Lock()
-	defer a.probeMu.Unlock()
-	st := a.probeStates[mac]
-	if st == nil {
-		st = &nicProbeState{}
-		a.probeStates[mac] = st
-	}
-	return st
-}

spinifex/agents/elbv2/datapath_test.go

@@ -1,484 +0,0 @@
-package elbv2
-
-import (
-	"errors"
-	"fmt"
-	"net"
-	"net/http"
-	"net/http/httptest"
-	"net/netip"
-	"net/url"
-	"os"
-	"path/filepath"
-	"sync"
-	"sync/atomic"
-	"testing"
-	"time"
-)
-
-const (
-	testDataMAC = "02:00:00:aa:00:01"
-	testMgmtMAC = "02:00:00:aa:00:02"
-	testNetcfg  = "NIC0_MAC=" + testDataMAC + "\nNIC0_CIDR=10.0.1.5/24\nNIC0_GW=10.0.1.1\nNIC0_DEFAULT=1\n" +
-		"NIC1_MAC=" + testMgmtMAC + "\nNIC1_CIDR=10.255.0.9/24\nNIC1_DEFAULT=0\n"
-)
-
-// datapathAgent returns a test agent whose netcfg and sysfs live in a temp dir.
-func datapathAgent(t *testing.T, gwURL, netcfg string) *Agent {
-	t.Helper()
-	a := newTestAgent(t, gwURL)
-	dir := t.TempDir()
-
-	a.netcfgPath = filepath.Join(dir, "netcfg")
-	if err := os.WriteFile(a.netcfgPath, []byte(netcfg), 0o600); err != nil {
-		t.Fatal(err)
-	}
-
-	a.sysNetDir = filepath.Join(dir, "net")
-	for name, mac := range map[string]string{"ens3": testDataMAC, "ens4": testMgmtMAC} {
-		stats := filepath.Join(a.sysNetDir, name, "statistics")
-		if err := os.MkdirAll(stats, 0o755); err != nil {
-			t.Fatal(err)
-		}
-		if err := os.WriteFile(filepath.Join(a.sysNetDir, name, "address"), []byte(mac+"\n"), 0o600); err != nil {
-			t.Fatal(err)
-		}
-		for file, v := range map[string]string{"rx_packets": "100", "tx_packets": "200", "rx_dropped": "3", "tx_dropped": "4"} {
-			if err := os.WriteFile(filepath.Join(stats, file), []byte(v+"\n"), 0o600); err != nil {
-				t.Fatal(err)
-			}
-		}
-	}
-
-	a.instanceIDFn = func() (string, error) { return "i-0123456789abcdef0", nil }
-	return a
-}
-
-func TestParseNetcfg_SkipsMgmtNICAndMissingGateway(t *testing.T) {
-	blob := testNetcfg +
-		"NIC2_MAC=02:00:00:aa:00:03\nNIC2_CIDR=10.0.2.5/24\nNIC2_GW=10.0.2.1\nNIC2_DEFAULT=0\n" +
-		"NIC3_MAC=02:00:00:aa:00:04\nNIC3_DHCP=1\nNIC3_DEFAULT=0\n" +
-		"NIC4_CIDR=10.0.4.5/24\nNIC4_GW=10.0.4.1\n" +
-		"NIC5_MAC=02:00:00:aa:00:06\nNIC5_CIDR=not-a-cidr\nNIC5_GW=10.0.5.1\n"
-
-	got := parseNetcfg([]byte(blob))
-	if len(got) != 2 {
-		t.Fatalf("got %d NICs, want the 2 with a gateway: %+v", len(got), got)
-	}
-	if got[0].MAC != testDataMAC || got[0].Gateway != netip.MustParseAddr("10.0.1.1") || got[0].Addr != netip.MustParseAddr("10.0.1.5") {
-		t.Errorf("first NIC = %+v", got[0])
-	}
-	if got[1].MAC != "02:00:00:aa:00:03" || got[1].Gateway != netip.MustParseAddr("10.0.2.1") {
-		t.Errorf("second NIC = %+v", got[1])
-	}
-}
-
-func TestParseNetcfg_OrdersByNICIndexAndLowercasesMAC(t *testing.T) {
-	blob := "NIC10_MAC=02:00:00:AA:00:0A\nNIC10_CIDR=10.0.9.5/24\nNIC10_GW=10.0.9.1\n" +
-		"NIC2_MAC=02:00:00:AA:00:02\nNIC2_CIDR=10.0.2.5/24\nNIC2_GW=10.0.2.1\n"
-	got := parseNetcfg([]byte(blob))
-	if len(got) != 2 || got[0].MAC != "02:00:00:aa:00:02" || got[1].MAC != "02:00:00:aa:00:0a" {
-		t.Fatalf("order or case wrong: %+v", got)
-	}
-}
-
-func TestIfaceByMACAndCounters(t *testing.T) {
-	a := datapathAgent(t, "http://unused", testNetcfg)
-
-	iface, err := ifaceByMAC(a.sysNetDir, "02:00:00:AA:00:01")
-	if err != nil || iface != "ens3" {
-		t.Fatalf("iface = %q, err = %v", iface, err)
-	}
-	if _, err := ifaceByMAC(a.sysNetDir, "02:00:00:ff:ff:ff"); err == nil {
-		t.Error("an unknown MAC must not resolve to an interface")
-	}
-
-	var r NICReport
-	readCounters(a.sysNetDir, iface, &r)
-	if r.RxPackets != 100 || r.TxPackets != 200 || r.RxDropped != 3 || r.TxDropped != 4 {
-		t.Errorf("counters = %+v", r)
-	}
-}
-
-func TestARPFrames_RequestShapeAndReplyMatching(t *testing.T) {
-	mac, _ := net.ParseMAC(testDataMAC)
-	src, gw := netip.MustParseAddr("10.0.1.5"), netip.MustParseAddr("10.0.1.1")
-
-	req := buildARPRequest(mac, src, gw)
-	if len(req) != 42 || req[12] != 0x08 || req[13] != 0x06 || req[21] != 1 {
-		t.Fatalf("request frame malformed: % x", req)
-	}
-	if string(req[38:42]) != string(gw.AsSlice()) {
-		t.Errorf("target IP not the gateway: % x", req[38:42])
-	}
-
-	reply := make([]byte, 42)
-	copy(reply[0:6], mac)
-	reply[12], reply[13] = 0x08, 0x06
-	reply[21] = 2
-	copy(reply[28:32], gw.AsSlice())
-	copy(reply[32:38], mac)
-	if !isARPReplyFrom(reply, mac, gw) {
-		t.Error("a reply from the gateway to our MAC must match")
-	}
-	if isARPReplyFrom(reply, mac, netip.MustParseAddr("10.0.1.99")) {
-		t.Error("a reply from another host must not match")
-	}
-	reply[21] = 1
-	if isARPReplyFrom(reply, mac, gw) {
-		t.Error("a request must not count as a reply")
-	}
-	if isARPReplyFrom(reply[:20], mac, gw) {
-		t.Error("a short frame must not match")
-	}
-}
-
-func TestProbeNICs_ProbesOnlyTheGatewayNIC(t *testing.T) {
-	a := datapathAgent(t, "http://unused", testNetcfg)
-	var mu sync.Mutex
-	var probed []string
-	a.arpProbeFn = func(iface string, _ net.HardwareAddr, src, gw netip.Addr, _ time.Duration) error {
-		mu.Lock()
-		defer mu.Unlock()
-		probed = append(probed, fmt.Sprintf("%s %s>%s", iface, src, gw))
-		return nil
-	}
-
-	reports := a.probeNICs()
-	if len(reports) != 1 || reports[0].MAC != testDataMAC || !reports[0].ProbeOK || reports[0].ConsecutiveFailures != 0 {
-		t.Fatalf("reports = %+v", reports)
-	}
-	if len(probed) != 1 || probed[0] != "ens3 10.0.1.5>10.0.1.1" {
-		t.Errorf("probed = %v, want only the data NIC", probed)
-	}
-	if reports[0].RxPackets != 100 || reports[0].TxDropped != 4 {
-		t.Errorf("counters missing from report: %+v", reports[0])
-	}
-}
-
-func TestProbeNICs_TwoMissesThenSuccessNeverReportsFailing(t *testing.T) {
-	a := datapathAgent(t, "http://unused", testNetcfg)
-	results := []error{errARPNoReply, errARPNoReply, nil, errARPNoReply, errARPNoReply, nil}
-	var n atomic.Int32
-	a.arpProbeFn = func(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error {
-		return results[n.Add(1)-1]
-	}
-
-	for i, want := range []int64{1, 2, 0, 1, 2, 0} {
-		r := a.probeNICs()
-		if len(r) != 1 || !r[0].ProbeOK || r[0].ConsecutiveFailures != want {
-			t.Fatalf("tick %d: report = %+v, want ok with %d failures", i, r, want)
-		}
-	}
-}
-
-func TestProbeNICs_ThreeMissesReportFailingAndRecoveryClears(t *testing.T) {
-	a := datapathAgent(t, "http://unused", testNetcfg)
-	var fail atomic.Bool
-	fail.Store(true)
-	a.arpProbeFn = func(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error {
-		if fail.Load() {
-			return errARPNoReply
-		}
-		return nil
-	}
-
-	for i := 1; i <= 4; i++ {
-		r := a.probeNICs()
-		wantOK := i < 3 // failing from the third consecutive miss
-		if r[0].ProbeOK != wantOK || r[0].ConsecutiveFailures != int64(i) {
-			t.Fatalf("miss %d: report = %+v, want ProbeOK=%v", i, r[0], wantOK)
-		}
-	}
-
-	fail.Store(false)
-	if r := a.probeNICs(); !r[0].ProbeOK || r[0].ConsecutiveFailures != 0 {
-		t.Errorf("a successful probe must clear the failure: %+v", r[0])
-	}
-}
-
-func TestProbeNICs_BlockedProbeIsBoundedAndCountsAsMiss(t *testing.T) {
-	a := datapathAgent(t, "http://unused", testNetcfg)
-	a.probeWait = 50 * time.Millisecond
-	release := make(chan struct{})
-	defer close(release)
-	var calls atomic.Int32
-	a.arpProbeFn = func(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error {
-		calls.Add(1)
-		<-release
-		return nil
-	}
-
-	for i := 1; i <= 3; i++ {
-		start := time.Now()
-		r := a.probeNICs()
-		if d := time.Since(start); d > time.Second {
-			t.Fatalf("tick %d waited %s on a stuck probe", i, d)
-		}
-		if len(r) != 1 || r[0].ConsecutiveFailures != int64(i) {
-			t.Fatalf("tick %d: report = %+v, want %d misses", i, r, i)
-		}
-	}
-	if got := calls.Load(); got != 1 {
-		t.Errorf("a stuck probe was started %d times, want 1 (no goroutine pile-up)", got)
-	}
-}
-
-func TestTick_BlockedProbeDoesNotDelayHeartbeatBeyondOneTick(t *testing.T) {
-	var got atomic.Pointer[url.Values]
-	gw := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		_ = r.ParseForm()
-		if r.FormValue("Action") == "LBAgentHeartbeat" {
-			v := r.Form
-			got.Store(&v)
-		}
-		fmt.Fprint(w, `<LBAgentHeartbeatResponse><LBAgentHeartbeatResult><Status>active</Status><ConfigHash>h1</ConfigHash></LBAgentHeartbeatResult></LBAgentHeartbeatResponse>`)
-	}))
-	defer gw.Close()
-
-	a := datapathAgent(t, gw.URL, testNetcfg)
-	a.localConfigHash = "h1"
-	a.probeWait = 50 * time.Millisecond
-	release := make(chan struct{})
-	defer close(release)
-	a.arpProbeFn = func(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error {
-		<-release
-		return nil
-	}
-
-	start := time.Now()
-	a.tick()
-	if d := time.Since(start); d >= pollInterval {
-		t.Fatalf("tick took %s with a stuck probe, want under one %s tick", d, pollInterval)
-	}
-	if got.Load() == nil {
-		t.Fatal("the heartbeat was never sent")
-	}
-}
-
-func TestHeartbeat_EncodesEvidenceWhenPresent(t *testing.T) {
-	var got atomic.Pointer[url.Values]
-	gw := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		_ = r.ParseForm()
-		v := r.Form
-		got.Store(&v)
-		fmt.Fprint(w, `<LBAgentHeartbeatResponse><LBAgentHeartbeatResult><Status>active</Status><ConfigHash>h1</ConfigHash></LBAgentHeartbeatResult></LBAgentHeartbeatResponse>`)
-	}))
-	defer gw.Close()
-
-	a := datapathAgent(t, gw.URL, testNetcfg)
-	a.localConfigHash = "h1"
-	a.arpProbeFn = func(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error { return errARPNoReply }
-
-	a.tick()
-	v := got.Load()
-	if v == nil {
-		t.Fatal("no heartbeat received")
-	}
-	want := map[string]string{
-		"InstanceID":                        "i-0123456789abcdef0",
-		"NICs.member.1.MAC":                 testDataMAC,
-		"NICs.member.1.ProbeOK":             "true",
-		"NICs.member.1.ConsecutiveFailures": "1",
-		"NICs.member.1.RxPackets":           "100",
-		"NICs.member.1.TxPackets":           "200",
-		"NICs.member.1.RxDropped":           "3",
-		"NICs.member.1.TxDropped":           "4",
-	}
-	for k, w := range want {
-		if g := v.Get(k); g != w {
-			t.Errorf("%s = %q, want %q", k, g, w)
-		}
-	}
-	if v.Get("NICs.member.2.MAC") != "" {
-		t.Error("the mgmt NIC must not be reported")
-	}
-}
-
-func TestHeartbeat_OmitsEvidenceWithoutNetcfg(t *testing.T) {
-	var got atomic.Pointer[url.Values]
-	gw := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		_ = r.ParseForm()
-		v := r.Form
-		got.Store(&v)
-		fmt.Fprint(w, `<LBAgentHeartbeatResponse><LBAgentHeartbeatResult><Status>active</Status><ConfigHash>h1</ConfigHash></LBAgentHeartbeatResult></LBAgentHeartbeatResponse>`)
-	}))
-	defer gw.Close()
-
-	a := datapathAgent(t, gw.URL, testNetcfg)
-	a.localConfigHash = "h1"
-	a.netcfgPath = filepath.Join(t.TempDir(), "absent")
-	a.instanceIDFn = func() (string, error) { return "", errors.New("must not be asked without NICs") }
-
-	a.tick()
-	v := got.Load()
-	if v == nil {
-		t.Fatal("no heartbeat received")
-	}
-	for k := range *v {
-		if k == "InstanceID" || len(k) >= 4 && k[:4] == "NICs" {
-			t.Errorf("unexpected evidence parameter %s", k)
-		}
-	}
-}
-
-func TestDataPathEvidence_InstanceIDRetriedUntilKnown(t *testing.T) {
-	a := datapathAgent(t, "http://unused", testNetcfg)
-	a.arpProbeFn = func(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error { return nil }
-	var calls atomic.Int32
-	a.instanceIDFn = func() (string, error) {
-		if calls.Add(1) < 3 {
-			return "", errors.New("metadata not ready")
-		}
-		return "i-late", nil
-	}
-
-	// A failure backs the lookup off for imdsRetryTicks-1 ticks.
-	ticks := func(n int) string {
-		var id string
-		for range n {
-			id = a.dataPathEvidence().InstanceID
-		}
-		return id
-	}
-	if got := ticks(imdsRetryTicks); got != "" || calls.Load() != 1 {
-		t.Fatalf("after %d ticks: id %q, calls %d; want no id and one lookup", imdsRetryTicks, got, calls.Load())
-	}
-	if got := ticks(imdsRetryTicks); got != "" || calls.Load() != 2 {
-		t.Fatalf("second window: id %q, calls %d; want no id and two lookups", got, calls.Load())
-	}
-	if got := ticks(imdsRetryTicks); got != "i-late" || calls.Load() != 3 {
-		t.Fatalf("third window: id %q, calls %d; want i-late and three lookups", got, calls.Load())
-	}
-	ticks(imdsRetryTicks)
-	if calls.Load() != 3 {
-		t.Errorf("metadata asked %d times, want 3 (stop once known)", calls.Load())
-	}
-}
-
-func TestTick_BlockedInstanceIDLookupDoesNotDelayHeartbeat(t *testing.T) {
-	var got atomic.Pointer[url.Values]
-	gw := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		_ = r.ParseForm()
-		if r.FormValue("Action") == "LBAgentHeartbeat" {
-			v := r.Form
-			got.Store(&v)
-		}
-		fmt.Fprint(w, `<LBAgentHeartbeatResponse><LBAgentHeartbeatResult><Status>active</Status><ConfigHash>h1</ConfigHash></LBAgentHeartbeatResult></LBAgentHeartbeatResponse>`)
-	}))
-	defer gw.Close()
-
-	a := datapathAgent(t, gw.URL, testNetcfg)
-	a.localConfigHash = "h1"
-	a.arpProbeFn = func(string, net.HardwareAddr, netip.Addr, netip.Addr, time.Duration) error { return nil }
-	release := make(chan struct{})
-	defer close(release)
-	var calls atomic.Int32
-	a.instanceIDFn = func() (string, error) {
-		calls.Add(1)
-		<-release
-		return "", errors.New("unreachable")
-	}
-
-	for i := range 3 {
-		start := time.Now()
-		a.tick()
-		if d := time.Since(start); d >= pollInterval {
-			t.Fatalf("tick %d took %s with IMDS blocked, want under one %s tick", i, d, pollInterval)
-		}
-	}
-	v := got.Load()
-	if v == nil {
-		t.Fatal("the heartbeat was never sent")
-	}
-	if v.Get("InstanceID") != "" {
-		t.Error("InstanceID must be omitted while the lookup fails")
-	}
-	if v.Get("NICs.member.1.MAC") == "" {
-		t.Error("NIC evidence must still be sent")
-	}
-	if calls.Load() != 1 {
-		t.Errorf("blocked lookup started %d times in 3 ticks, want 1 (backoff)", calls.Load())
-	}
-}
-
-func TestFetchInstanceID_SharesOneDeadline(t *testing.T) {
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		time.Sleep(3 * imdsWait)
-	}))
-	defer srv.Close()
-	old := imdsBase
-	imdsBase = srv.URL + "/latest"
-	defer func() { imdsBase = old }()
-
-	start := time.Now()
-	if _, err := fetchInstanceID(); err == nil {
-		t.Fatal("a hung metadata service must fail the lookup")
-	}
-	if d := time.Since(start); d > imdsWait+imdsWait/2 {
-		t.Fatalf("lookup took %s, want about one %s deadline", d, imdsWait)
-	}
-}
-
-func TestARPProbe_NoResponderFailsClosed(t *testing.T) {
-	mac, _ := net.ParseMAC(testDataMAC)
-	// Nobody answers an ARP request sent on loopback, so a probe that returned
-	// nil here would mask a dead data path. It fails either on the socket (no
-	// CAP_NET_RAW) or on the missing reply.
-	err := arpProbe("lo", mac, netip.MustParseAddr("10.9.9.2"), netip.MustParseAddr("10.9.9.1"), 100*time.Millisecond)
-	if err == nil {
-		t.Fatal("a probe with no responder must fail")
-	}
-	if err := arpProbe("no-such-iface0", mac, netip.MustParseAddr("10.9.9.2"), netip.MustParseAddr("10.9.9.1"), 100*time.Millisecond); err == nil {
-		t.Fatal("a probe on a missing interface must fail")
-	}
-}
-
-func TestFetchInstanceID(t *testing.T) {
-	serve := func(tokenStatus, idStatus int, wantToken string) *httptest.Server {
-		return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-			switch {
-			case r.Method == http.MethodPut && r.URL.Path == "/latest/api/token":
-				w.WriteHeader(tokenStatus)
-				fmt.Fprint(w, "tok123")
-			case r.URL.Path == "/latest/meta-data/instance-id":
-				if got := r.Header.Get("X-Aws-Ec2-Metadata-Token"); got != wantToken {
-					t.Errorf("token header = %q, want %q", got, wantToken)
-				}
-				w.WriteHeader(idStatus)
-				fmt.Fprint(w, " i-abc123 \n")
-			default:
-				http.NotFound(w, r)
-			}
-		}))
-	}
-	saved := imdsBase
-	defer func() { imdsBase = saved }()
-
-	for _, tc := range []struct {
-		name                  string
-		tokenStatus, idStatus int
-		wantToken             string
-		wantID                string
-		wantErr               bool
-	}{
-		{"token then id", http.StatusOK, http.StatusOK, "tok123", "i-abc123", false},
-		{"token refused falls back to v1", http.StatusNotFound, http.StatusOK, "", "i-abc123", false},
-		{"id endpoint error", http.StatusOK, http.StatusServiceUnavailable, "tok123", "", true},
-	} {
-		t.Run(tc.name, func(t *testing.T) {
-			srv := serve(tc.tokenStatus, tc.idStatus, tc.wantToken)
-			defer srv.Close()
-			imdsBase = srv.URL + "/latest"
-			id, err := fetchInstanceID()
-			if (err != nil) != tc.wantErr || id != tc.wantID {
-				t.Fatalf("id = %q, err = %v", id, err)
-			}
-		})
-	}
-
-	imdsBase = "http://127.0.0.1:1/latest"
-	if _, err := fetchInstanceID(); err == nil {
-		t.Error("an unreachable metadata service must fail")
-	}
-}

spinifex/agents/elbv2/evidence.go

@@ -1,133 +0,0 @@
-package elbv2
-
-import (
-	"context"
-	"fmt"
-	"io"
-	"net/http"
-	"net/url"
-	"strconv"
-	"strings"
-	"time"
-)
-
-const (
-	imdsWait       = time.Second
-	imdsRetryTicks = 6
-)
-
-// imdsBase is the in-VM metadata endpoint (a variable so tests can point it at a fake).
-var imdsBase = "http://169.254.169.254/latest"
-
-// dataPathEvidence is the optional part of a heartbeat: which VM the agent runs
-// in and what each data NIC's gateway probe saw. An empty value adds nothing,
-// which is what an agent without a netcfg blob sends.
-type dataPathEvidence struct {
-	InstanceID string
-	NICs       []NICReport
-}
-
-// addTo encodes the evidence as additional heartbeat query parameters.
-func (e dataPathEvidence) addTo(params url.Values) {
-	if e.InstanceID != "" {
-		params.Set("InstanceID", e.InstanceID)
-	}
-	for i, n := range e.NICs {
-		p := "NICs.member." + strconv.Itoa(i+1) + "."
-		params.Set(p+"MAC", n.MAC)
-		params.Set(p+"ProbeOK", strconv.FormatBool(n.ProbeOK))
-		params.Set(p+"ConsecutiveFailures", strconv.FormatInt(n.ConsecutiveFailures, 10))
-		params.Set(p+"RxPackets", strconv.FormatInt(n.RxPackets, 10))
-		params.Set(p+"TxPackets", strconv.FormatInt(n.TxPackets, 10))
-		params.Set(p+"RxDropped", strconv.FormatInt(n.RxDropped, 10))
-		params.Set(p+"TxDropped", strconv.FormatInt(n.TxDropped, 10))
-	}
-}
-
-// dataPathEvidence gathers the probe results and instance ID for one heartbeat.
-// Neither step can fail the heartbeat: missing evidence is simply omitted.
-func (a *Agent) dataPathEvidence() dataPathEvidence {
-	nics := a.probeNICs()
-	if len(nics) == 0 {
-		return dataPathEvidence{}
-	}
-	if a.instanceID == "" {
-		a.lookupInstanceID()
-	}
-	return dataPathEvidence{InstanceID: a.instanceID, NICs: nics}
-}
-
-// lookupInstanceID asks for the instance ID without ever holding up a tick: the
-// call is bounded by imdsWait and, after a failure, retried only every
-// imdsRetryTicks ticks.
-func (a *Agent) lookupInstanceID() {
-	if a.idSkip > 0 {
-		a.idSkip--
-		return
-	}
-	type result struct {
-		id  string
-		err error
-	}
-	ch := make(chan result, 1)
-	go func() {
-		id, err := a.instanceIDFn()
-		ch <- result{id, err}
-	}()
-	timer := time.NewTimer(imdsWait)
-	defer timer.Stop()
-	select {
-	case r := <-ch:
-		if r.err == nil && r.id != "" {
-			a.instanceID = r.id
-			return
-		}
-	case <-timer.C:
-	}
-	a.idSkip = imdsRetryTicks - 1
-}
-
-// fetchInstanceID reads the instance ID from the in-VM metadata service. The
-// token request and the read share one deadline.
-func fetchInstanceID() (string, error) {
-	ctx, cancel := context.WithTimeout(context.Background(), imdsWait)
-	defer cancel()
-	c := &http.Client{}
-
-	var token string
-	if req, err := http.NewRequestWithContext(ctx, http.MethodPut, imdsBase+"/api/token", nil); err == nil {
-		req.Header.Set("X-Aws-Ec2-Metadata-Token-Ttl-Seconds", "60")
-		if resp, err := c.Do(req); err == nil {
-			b, _ := io.ReadAll(io.LimitReader(resp.Body, 256))
-			resp.Body.Close()
-			if resp.StatusCode == http.StatusOK {
-				token = strings.TrimSpace(string(b))
-			}
-		}
-	}
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodGet, imdsBase+"/meta-data/instance-id", nil)
-	if err != nil {
-		return "", err
-	}
-	if token != "" {
-		req.Header.Set("X-Aws-Ec2-Metadata-Token", token)
-	}
-	resp, err := c.Do(req)
-	if err != nil {
-		return "", err
-	}
-	defer resp.Body.Close()
-	if resp.StatusCode != http.StatusOK {
-		return "", fmt.Errorf("instance-id: status %d", resp.StatusCode)
-	}
-	b, err := io.ReadAll(io.LimitReader(resp.Body, 256))
-	if err != nil {
-		return "", err
-	}
-	id := strings.TrimSpace(string(b))
-	if id == "" {
-		return "", fmt.Errorf("instance-id: empty")
-	}
-	return id, nil
-}

spinifex/agents/lambda/hostlink/configdisk.go

@@ -1,71 +0,0 @@
-package hostlink
-
-import (
-	"errors"
-	"fmt"
-	"io"
-	"os"
-
-	lambdav1 "github.com/mulgadc/spinifex/contracts/lambda/v1"
-	"golang.org/x/sys/unix"
-)
-
-// On a VMM without fw_cfg the executor attaches the same entries as a tar
-// archive on a second disk, which the runtime image's init extracts here in
-// the fw_cfg sysfs layout, so the fw_cfg readers work unchanged.
-const (
-	ConfigDiskRoot   = "/spinifex-cfg"
-	ConfigDiskDir    = ConfigDiskRoot + "/by_name/" + lambdav1.FwCfgPrefix
-	ConfigDiskDevice = "/dev/vdb"
-)
-
-// zeroChunk bounds each write while erasing the configuration disk.
-const zeroChunk = 64 << 10
-
-// unmount is a var so tests without privilege can stand in for the tmpfs.
-var unmount = unix.Unmount
-
-// SealConfigDisk overwrites the whole configuration disk with zeros, then
-// removes the extracted tree. The tree stays if the disk cannot be erased, so
-// SealFwCfg's check on root fails closed.
-func SealConfigDisk(device, root string) error {
-	if err := zeroDevice(device); err != nil {
-		return fmt.Errorf("hostlink: erase configuration disk: %w", err)
-	}
-	if err := unmount(root, 0); err != nil && !errors.Is(err, unix.EINVAL) {
-		return fmt.Errorf("hostlink: unmount %s: %w", root, err)
-	}
-	if err := os.RemoveAll(root); err != nil {
-		return fmt.Errorf("hostlink: remove %s: %w", root, err)
-	}
-	return nil
-}
-
-func zeroDevice(device string) error {
-	f, err := os.OpenFile(device, os.O_WRONLY, 0)
-	if err != nil {
-		return err
-	}
-	size, err := f.Seek(0, io.SeekEnd)
-	if err == nil {
-		_, err = f.Seek(0, io.SeekStart)
-	}
-	if err != nil {
-		_ = f.Close()
-		return err
-	}
-	zeros := make([]byte, zeroChunk)
-	for left := size; left > 0; {
-		n := min(left, int64(len(zeros)))
-		if _, err := f.Write(zeros[:n]); err != nil {
-			_ = f.Close()
-			return err
-		}
-		left -= n
-	}
-	if err := f.Sync(); err != nil {
-		_ = f.Close()
-		return err
-	}
-	return f.Close()
-}

spinifex/agents/lambda/hostlink/configdisk_test.go

@@ -1,63 +0,0 @@
-//test:in-package reaches the unexported configDisk and fwCfgTree readers to seal and read a real archive
-
-package hostlink
-
-import (
-	"bytes"
-	"os"
-	"path/filepath"
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-	"golang.org/x/sys/unix"
-)
-
-// configDisk writes a fake disk holding secret and an extracted tree; the
-// stand-in unmount reports the tree is not a mount, as on a plain directory.
-func configDisk(t *testing.T, secret []byte) (device, root string) {
-	t.Helper()
-	dir := t.TempDir()
-	device = filepath.Join(dir, "vdb")
-	require.NoError(t, os.WriteFile(device, append(bytes.Repeat([]byte{0xa5}, 3*zeroChunk+7), secret...), 0o600))
-	root = fwCfgTree(t)
-	require.NoError(t, os.WriteFile(filepath.Join(root, "by_name", "key"), secret, 0o600))
-	old := unmount
-	unmount = func(string, int) error { return unix.EINVAL }
-	t.Cleanup(func() { unmount = old })
-	return device, root
-}
-
-func TestSealConfigDisk(t *testing.T) {
-	secret := []byte("guest-key-material")
-
-	t.Run("erases the disk and removes the tree", func(t *testing.T) {
-		device, root := configDisk(t, secret)
-		before, err := os.Stat(device)
-		require.NoError(t, err)
-		require.NoError(t, SealFwCfg(root, func() error { return SealConfigDisk(device, root) }))
-
-		disk, err := os.ReadFile(device)
-		require.NoError(t, err)
-		assert.Len(t, disk, int(before.Size()))
-		assert.Equal(t, make([]byte, len(disk)), disk, "configuration disk not fully erased")
-		_, err = os.Stat(root)
-		assert.ErrorIs(t, err, os.ErrNotExist)
-	})
-	t.Run("an unwritable disk keeps the tree and fails closed", func(t *testing.T) {
-		device, root := configDisk(t, secret)
-		require.NoError(t, os.Chmod(device, 0o400))
-		err := SealFwCfg(root, func() error { return SealConfigDisk(device, root) })
-		require.Error(t, err)
-		assert.Contains(t, err.Error(), "still readable")
-		_, err = os.Stat(root)
-		assert.NoError(t, err)
-	})
-	t.Run("a failed unmount keeps the tree", func(t *testing.T) {
-		device, root := configDisk(t, secret)
-		unmount = func(string, int) error { return unix.EBUSY }
-		require.Error(t, SealFwCfg(root, func() error { return SealConfigDisk(device, root) }))
-		_, err := os.Stat(root)
-		assert.NoError(t, err)
-	})
-}

spinifex/agents/lambda/hostlink/link.go

@@ -28,12 +28,11 @@ type Request struct {
 	Credential *Credential `json:"credential,omitempty"`
 }
 
-// Credential holds one execution-role session and its function environment.
+// Credential is one execution-role session. It is never logged or stored.
 type Credential struct {
-	AccessKeyID     string            `json:"access_key_id"`
-	SecretAccessKey string            `json:"secret_access_key"`
-	SessionToken    string            `json:"session_token"`
-	Environment     map[string]string `json:"environment,omitempty"`
+	AccessKeyID     string `json:"access_key_id"`
+	SecretAccessKey string `json:"secret_access_key"`
+	SessionToken    string `json:"session_token"`
 }
 
 // LogValue keeps every value out of logs.

spinifex/agents/lambda/hostlink/protocol_testlink.go

@@ -4,5 +4,5 @@ package hostlink
 
 // Protocol is the only ALPN protocol a test-link host and guest offer. It
 // differs from the release protocol because the first frame carries the
-// execution credential and function environment, which an older guest would ignore.
-const Protocol = "spinifex-lambda-testlink/3"
+// execution credential, which an older guest would ignore.
+const Protocol = "spinifex-lambda-testlink/2"

spinifex/agents/lambda/hostlink/test_stream.go

@@ -321,19 +321,6 @@ func (s *TestStream) HostSend(req Request) error {
 	return s.write(frame{Type: typeInvoke, Request: &req})
 }
 
-// CheckInvokeFits encodes req as HostSend would and returns ErrFrameTooLarge
-// when the frame exceeds the bound, so a caller can refuse before launch.
-func CheckInvokeFits(req Request) error {
-	b, err := json.Marshal(frame{Type: typeInvoke, Request: &req})
-	if err != nil {
-		return err
-	}
-	if len(b) > maxFrame {
-		return ErrFrameTooLarge
-	}
-	return nil
-}
-
 // HostShutdown writes a shutdown frame, as the host side of the carrier.
 func (s *TestStream) HostShutdown() error {
 	return s.write(frame{Type: typeShutdown})

spinifex/domains/acm/store.go

@@ -14,7 +14,6 @@ import (
 	iamdomain "github.com/mulgadc/spinifex/spinifex/domains/iam"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/arn"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
 )
@@ -184,7 +183,7 @@ func NewStore(ctx context.Context, nc *nats.Conn, masterKey []byte) (*Store, err
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	kv, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketACM, KVBucketACMVersion)
+	kv, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketACM, KVBucketACMVersion)
 	if err != nil {
 		return nil, err
 	}

spinifex/domains/admission/quota/bedrock.go

@@ -10,7 +10,7 @@ import (
 	"time"
 
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 	"golang.org/x/time/rate"
 )
@@ -171,7 +171,7 @@ func (s *Service) RunBedrockRPMSync(ctx context.Context, js jetstream.JetStream)
 	if s == nil || !s.limits.RequestsPerMinuteEnabled {
 		return
 	}
-	kv, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketBedrockRPMUsage, 1)
+	kv, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketBedrockRPMUsage, 1)
 	if err != nil {
 		slog.Warn("bedrock rpm sync: open usage bucket failed, sync disabled", "err", err)
 		return

spinifex/domains/admission/quota/override.go

@@ -9,7 +9,7 @@ import (
 	"time"
 
 	awserrors "github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -187,7 +187,7 @@ func (s *Service) PutAccountQuota(ctx context.Context, accountID string, over Ov
 	over.UpdatedAt = time.Now().UTC().Format(time.RFC3339)
 	over.UpdatedBy = updatedBy
 
-	_, err := kvutil.Update(ctx, s.overrides, accountID, kvutil.CASConfig{
+	_, err := kvstore.Update(ctx, s.overrides, accountID, kvstore.CASConfig{
 		CreateIfAbsent: true,
 		Attempts:       overrideMaxAttempts,
 		Exhausted:      func(string, int) error { return errOverridesContended },
@@ -230,7 +230,7 @@ func (s *Service) clearAccountQuota(ctx context.Context, accountID string) error
 			return nil
 		case errors.Is(err, jetstream.ErrKeyNotFound):
 			return nil
-		case kvutil.IsRevisionConflict(err):
+		case kvstore.IsRevisionConflict(err):
 			// A concurrent write landed between the read and the guarded
 			// delete; retry against whatever revision now holds.
 			continue

spinifex/domains/admission/quota/quota.go

@@ -22,10 +22,10 @@ const ReconcileInterval = 30 * time.Second
 // dimension; the live-counted dimensions need no stored state.
 const KVBucketAccountUsage = "spinifex-account-usage"
 
-// KVBucketQuotaReconcile is the leader-lock bucket for the gateway quota
-// reconcile. It is deliberately distinct from vpcd's network-reconcile lock so
-// the two independent loops never share one mutex and block each other.
-const KVBucketQuotaReconcile = "spinifex-quota-reconcile"
+// KeyReconcile is the sync-bucket leader lease for the gateway quota reconcile.
+// It is deliberately distinct from vpcd's network-reconcile lease so the two
+// independent loops never share one mutex and block each other.
+const KeyReconcile = "quota.reconcile"
 
 // KVBucketAccountQuota holds the per-account limit overrides, keyed by
 // accountID. It is separate from the accounts bucket so a quota write never

spinifex/domains/admission/quota/seal.go

@@ -8,7 +8,7 @@ import (
 
 	awserrors "github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -73,7 +73,7 @@ func (s *Service) sealBucket(ctx context.Context) (jetstream.KeyValue, error) {
 		slog.ErrorContext(ctx, "Quota override seal has no JetStream client wired to open the override bucket")
 		return nil, errors.New(awserrors.ErrorInternalError)
 	}
-	kv, err := kvutil.GetOrCreateBucket(ctx, s.js, KVBucketAccountQuota, 1)
+	kv, err := kvstore.GetOrCreateBucket(ctx, s.js, KVBucketAccountQuota, 1)
 	if err != nil {
 		slog.ErrorContext(ctx, "Failed to open quota override bucket for seal", "err", err)
 		return nil, errors.New(awserrors.ErrorInternalError)
@@ -136,7 +136,7 @@ func (s *Service) SealAccountOverrides(ctx context.Context, accountID string) (*
 		switch {
 		case writeErr == nil:
 			return &OverrideSeal{Revision: newRev}, nil
-		case kvutil.IsRevisionConflict(writeErr):
+		case kvstore.IsRevisionConflict(writeErr):
 			continue
 		default:
 			slog.ErrorContext(ctx, "Failed to write sealed quota override record", "accountID", accountID, "err", writeErr)

spinifex/domains/dns/delegation.go

@@ -8,7 +8,7 @@ import (
 	"time"
 
 	nsconfig "github.com/mulgadc/northstar/pkg/config"
-	"github.com/nats-io/nats.go"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvsync"
 
 	reconcilelock "github.com/mulgadc/spinifex/spinifex/domains/network/reconcile"
 )
@@ -48,7 +48,7 @@ func (w *Writer) PlatformDelegation(name string) (servers []string, zone string,
 // Delegate writes, or with remove withdraws, an NS set in the base zone that
 // delegates each name to this cluster's nameservers. It takes the zone-writer
 // election first, so it never races the reconcile's own writes to the zone.
-func (w *Writer) Delegate(ctx context.Context, nc *nats.Conn, holder string, names []string, remove bool) (*ChangeResult, error) {
+func (w *Writer) Delegate(ctx context.Context, lease *kvsync.Lease, names []string, remove bool) (*ChangeResult, error) {
 	if !w.Enabled() {
 		return nil, errors.New("northstar is not configured on this node")
 	}
@@ -77,7 +77,7 @@ func (w *Writer) Delegate(ctx context.Context, nc *nats.Conn, holder string, nam
 	if len(batch.Changes) == 0 {
 		return &ChangeResult{}, nil
 	}
-	release, err := w.awaitWriterElection(ctx, nc, holder)
+	release, err := w.awaitWriterElection(ctx, lease)
 	if err != nil {
 		return nil, err
 	}
@@ -85,19 +85,19 @@ func (w *Writer) Delegate(ctx context.Context, nc *nats.Conn, holder string, nam
 	return w.ApplyBatch(batch)
 }
 
-func (w *Writer) awaitWriterElection(ctx context.Context, nc *nats.Conn, holder string) (func(), error) {
-	if nc == nil {
+func (w *Writer) awaitWriterElection(ctx context.Context, lease *kvsync.Lease) (func(), error) {
+	if lease == nil {
 		return func() {}, nil
 	}
 	tick := time.NewTicker(delegateLockPoll)
 	defer tick.Stop()
 	for {
-		if release, ok := reconcilelock.AcquireLeader(ctx, nc, KVBucketDNSReconcile, holder); ok {
+		if release, ok := reconcilelock.AcquireLeader(ctx, lease); ok {
 			return release, nil
 		}
 		select {
 		case <-ctx.Done():
-			return nil, fmt.Errorf("another node is writing the zone: %w", ctx.Err())
+			return nil, fmt.Errorf("another writer has the zone: %w", ctx.Err())
 		case <-tick.C:
 		}
 	}

spinifex/domains/dns/delegation_test.go

@@ -14,7 +14,7 @@ import (
 
 func TestDelegateWritesAndWithdrawsPlatformDelegations(t *testing.T) {
 	w, _ := newMultiNodeTestWriter(t)
-	_, err := w.Delegate(t.Context(), nil, "node1", []string{"a.r53e2e.spx3.net", "B.r53e2e.spx3.net."}, false)
+	_, err := w.Delegate(t.Context(), nil, []string{"a.r53e2e.spx3.net", "B.r53e2e.spx3.net."}, false)
 	require.NoError(t, err)
 
 	for _, name := range []string{"a.r53e2e.spx3.net.", "b.r53e2e.spx3.net"} {
@@ -29,7 +29,7 @@ func TestDelegateWritesAndWithdrawsPlatformDelegations(t *testing.T) {
 	assert.True(t, ok, "a name below the base domain is decided by the base zone")
 	assert.Empty(t, servers)
 
-	_, err = w.Delegate(t.Context(), nil, "node1", []string{"a.r53e2e.spx3.net"}, true)
+	_, err = w.Delegate(t.Context(), nil, []string{"a.r53e2e.spx3.net"}, true)
 	require.NoError(t, err)
 	servers, _, _, err = w.PlatformDelegation("a.r53e2e.spx3.net")
 	require.NoError(t, err)
@@ -43,11 +43,11 @@ func TestDelegateWritesAndWithdrawsPlatformDelegations(t *testing.T) {
 // servers written while the cluster published a larger set.
 func TestDelegateRemoveWithdrawsServersFromAnEarlierSet(t *testing.T) {
 	w, _ := newMultiNodeTestWriter(t)
-	_, err := w.Delegate(t.Context(), nil, "node1", []string{"a.r53e2e.spx3.net"}, false)
+	_, err := w.Delegate(t.Context(), nil, []string{"a.r53e2e.spx3.net"}, false)
 	require.NoError(t, err)
 
 	w.nameservers = w.nameservers[:1]
-	_, err = w.Delegate(t.Context(), nil, "node1", []string{"a.r53e2e.spx3.net"}, true)
+	_, err = w.Delegate(t.Context(), nil, []string{"a.r53e2e.spx3.net"}, true)
 	require.NoError(t, err)
 	servers, _, _, err := w.PlatformDelegation("a.r53e2e.spx3.net")
 	require.NoError(t, err)
@@ -66,7 +66,7 @@ func TestPlatformDelegationIgnoresNamesOutsideTheBaseDomain(t *testing.T) {
 func TestDelegateRefusesPlatformAndForeignNames(t *testing.T) {
 	w, _ := newMultiNodeTestWriter(t)
 	for _, name := range []string{"ns1.spx3.net", "x.elb.spx3.net", "compute.spx3.net", "example.com", "spx3.net"} {
-		_, err := w.Delegate(t.Context(), nil, "node1", []string{"ok.spx3.net", name}, false)
+		_, err := w.Delegate(t.Context(), nil, []string{"ok.spx3.net", name}, false)
 		require.Error(t, err, name)
 	}
 	servers, _, _, err := w.PlatformDelegation("ok.spx3.net")
@@ -76,22 +76,45 @@ func TestDelegateRefusesPlatformAndForeignNames(t *testing.T) {
 
 func TestDelegateWaitsForTheZoneWriterElection(t *testing.T) {
 	w, _ := newMultiNodeTestWriter(t)
-	_, nc, _ := testutil.StartTestJetStream(t)
+	_, _, js := testutil.StartTestJetStream(t)
 	prev := delegateLockPoll
 	delegateLockPoll = 10 * time.Millisecond
 	t.Cleanup(func() { delegateLockPoll = prev })
 
-	release, ok := reconcilelock.AcquireLeader(t.Context(), nc, KVBucketDNSReconcile, "node2")
+	release, ok := reconcilelock.AcquireLeader(t.Context(), dnsLease(js, "node2"))
 	require.True(t, ok)
 	ctx, cancel := context.WithTimeout(t.Context(), 100*time.Millisecond)
 	defer cancel()
-	_, err := w.Delegate(ctx, nc, "node1", []string{"a.spx3.net"}, false)
-	require.ErrorContains(t, err, "another node is writing the zone")
+	node1 := dnsLease(js, "node1")
+	_, err := w.Delegate(ctx, node1, []string{"a.spx3.net"}, false)
+	require.ErrorContains(t, err, "another writer has the zone")
 
 	release()
-	_, err = w.Delegate(t.Context(), nc, "node1", []string{"a.spx3.net"}, false)
+	_, err = w.Delegate(t.Context(), node1, []string{"a.spx3.net"}, false)
 	require.NoError(t, err)
 	servers, _, _, err := w.PlatformDelegation("a.spx3.net")
 	require.NoError(t, err)
 	assert.Len(t, servers, 2)
 }
+
+// The reconcile pass and a delegation on one node share the DNS lease, so a
+// delegation waits for this node's own pass as it would for another node's.
+func TestDelegateWaitsForThisNodesOwnReconcilePass(t *testing.T) {
+	w, _ := newMultiNodeTestWriter(t)
+	_, _, js := testutil.StartTestJetStream(t)
+	prev := delegateLockPoll
+	delegateLockPoll = 10 * time.Millisecond
+	t.Cleanup(func() { delegateLockPoll = prev })
+
+	shared := dnsLease(js, "node1")
+	release, ok := reconcilelock.AcquireLeader(t.Context(), shared)
+	require.True(t, ok, "the reconcile pass holds the lease")
+	ctx, cancel := context.WithTimeout(t.Context(), 100*time.Millisecond)
+	defer cancel()
+	_, err := w.Delegate(ctx, shared, []string{"a.spx3.net"}, false)
+	require.ErrorContains(t, err, "another writer has the zone")
+
+	release()
+	_, err = w.Delegate(t.Context(), shared, []string{"a.spx3.net"}, false)
+	require.NoError(t, err)
+}

spinifex/domains/dns/reconcile.go

@@ -11,7 +11,7 @@ import (
 	"github.com/mulgadc/spinifex/spinifex/bootstrap/config"
 	reconcilelock "github.com/mulgadc/spinifex/spinifex/domains/network/reconcile"
 	"github.com/mulgadc/spinifex/spinifex/foundation/lifecycle/reconciler"
-	"github.com/nats-io/nats.go"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvsync"
 )
 
 // DefaultReconcileInterval is how often the drift backstop rebuilds managed
@@ -33,6 +33,12 @@ const DefaultReconcileInterval = reconciler.DefaultResync
 // costs a cycle for nothing.
 const DefaultComputeTimeout = 2 * time.Minute
 
+// NewReconcileLease returns the zone writer's election lease, shared by the
+// drift reconcile and Delegate so the two never write the zone together.
+func NewReconcileLease(store kvsync.KVStore, holder string) *kvsync.Lease {
+	return reconcilelock.NewLeaderLease(store, KeyReconcile, holder)
+}
+
 // DesiredFunc returns the full desired managed record set built from the live
 // resource inventory across all tenants. The daemon supplies it by enumerating
 // instances, load balancers, and EKS clusters.
@@ -102,7 +108,8 @@ type Reconciler struct {
 	// lands in. Distinct from both baseDomain and internalDomain, so pruning it
 	// needs its own read the same way internalDomain does for EC2.
 	serviceZone string
-	nc          *nats.Conn
+	// lease elects the one node that writes the zones; nil skips the election.
+	lease *kvsync.Lease
 	// writer applies the converged batch. The reconcile owns the zone object on
 	// the node that won the election, so the batch never leaves this process.
 	writer   *Writer
@@ -111,7 +118,6 @@ type Reconciler struct {
 	// computeTimeout bounds how long one pass spends building the desired set.
 	// A field rather than a constant so a test can shrink it.
 	computeTimeout time.Duration
-	holder         string
 	sources        []reconciler.Source
 }
 
@@ -121,18 +127,15 @@ type Reconciler struct {
 // changes should wake the loop. They are optional: with none supplied the loop
 // falls back to the interval alone, which is the behaviour that predates the
 // watch.
-func NewReconciler(cfg *config.Config, cluster *config.ClusterConfig, nc *nats.Conn, writer *Writer, desired DesiredFunc, sources ...reconciler.Source) *Reconciler {
+func NewReconciler(cfg *config.Config, cluster *config.ClusterConfig, lease *kvsync.Lease, writer *Writer, desired DesiredFunc, sources ...reconciler.Source) *Reconciler {
 	r := &Reconciler{
-		nc:             nc,
+		lease:          lease,
 		writer:         writer,
 		desired:        desired,
 		interval:       DefaultReconcileInterval,
 		computeTimeout: DefaultComputeTimeout,
 		sources:        sources,
 	}
-	if cfg != nil {
-		r.holder = cfg.Node
-	}
 	zoneCfg, ok := zoneS3Config(cfg)
 	if !ok || desired == nil || !writer.Enabled() {
 		return r
@@ -166,25 +169,24 @@ func (r *Reconciler) Run(ctx context.Context) {
 		// No revisit deadline: the desired set is a function of the buckets
 		// being watched, so a change is the only reason to run early.
 		Reconcile: func(ctx context.Context) (time.Duration, error) {
-			r.reconcileOnce(ctx)
-			return 0, nil
+			return r.reconcileOnce(ctx), nil
 		},
 		Resync: r.interval,
 	})
 }
 
 // reconcileOnce computes the converging batch and applies it, on whichever node
-// wins this cycle's election. The read that finds the stale records and the
+// wins this cycle's election, and returns the election's revisit when it lost. The read that finds the stale records and the
 // write that removes them now happen in one process, so a record re-added
 // between them is no longer deleted by a batch that predates it.
-func (r *Reconciler) reconcileOnce(ctx context.Context) {
+func (r *Reconciler) reconcileOnce(ctx context.Context) time.Duration {
 	if !r.enabled {
-		return
+		return 0
 	}
-	if r.nc != nil {
-		release, elected := reconcilelock.AcquireLeader(ctx, r.nc, KVBucketDNSReconcile, r.holder)
+	if r.lease != nil {
+		release, retry, elected := reconcilelock.ElectPass(ctx, r.lease)
 		if !elected {
-			return
+			return retry
 		}
 		defer release()
 	}
@@ -193,18 +195,19 @@ func (r *Reconciler) reconcileOnce(ctx context.Context) {
 		// A corrupt zone is handled in readZone; anything reaching here left the
 		// desired state unapplied, so surface it rather than burying it in WARN.
 		slog.Error("dns reconcile: compute batch failed, retrying next cycle", "error", err)
-		return
+		return 0
 	}
 	if len(batch) == 0 {
-		return
+		return 0
 	}
 	slog.Debug("dns reconcile: converging", "changes", len(batch))
 	res, err := r.writer.ApplyBatch(&ChangeBatch{Changes: batch})
 	if err != nil {
 		slog.Warn("dns reconcile: apply failed, retrying next cycle", "changes", len(batch), "error", err)
-		return
+		return 0
 	}
 	slog.Debug("dns reconcile: converged", "changes", res.Applied, "zones", res.Zones)
+	return 0
 }
 
 // computeBatchBounded runs computeBatch with a deadline, so a pass that never

spinifex/domains/dns/reconcile_election_test.go

@@ -8,41 +8,50 @@ import (
 
 	"github.com/mulgadc/spinifex/internal/testkit"
 	reconcilelock "github.com/mulgadc/spinifex/spinifex/domains/network/reconcile"
-	"github.com/nats-io/nats.go"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvsync"
+	"github.com/nats-io/nats.go/jetstream"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
 
-// electingReconcilerFor is reconcilerFor with a live connection, so a pass runs
-// the election instead of skipping it.
-func electingReconcilerFor(w *Writer, nc *nats.Conn, holder string, desired func() DesiredSet) *Reconciler {
+// electingReconcilerFor is reconcilerFor with a lease, so a pass runs the
+// election instead of skipping it.
+func electingReconcilerFor(w *Writer, lease *kvsync.Lease, desired func() DesiredSet) *Reconciler {
 	r := reconcilerFor(w, desired)
-	r.nc = nc
-	r.holder = holder
+	r.lease = lease
 	return r
 }
 
+// dnsLease returns holder's DNS lease on a store of its own, standing in for
+// a separate process.
+func dnsLease(js jetstream.JetStream, holder string) *kvsync.Lease {
+	return NewReconcileLease(kvstore.OpenSyncBucket(js), holder)
+}
+
 // The zone lock is gone, and what replaces it is that a zone has one writer:
 // the node holding the reconcile lease. A node that loses the election must
 // therefore not touch the object at all — if it did, two nodes could be
 // read-modify-writing the same zone with nothing serialising them.
 func TestReconcileOnce_ALoserOfTheElectionDoesNotWriteTheZone(t *testing.T) {
-	_, nc, _ := testutil.StartTestJetStream(t)
+	_, _, js := testutil.StartTestJetStream(t)
 	w, objects := newTestWriter(t)
 	before := objects[testBase+".toml"]
 	require.NotEmpty(t, before)
 
 	// A peer holds the lease for the whole pass, so this node cannot win it.
-	release, elected := reconcilelock.AcquireLeader(t.Context(), nc, KVBucketDNSReconcile, "peer")
+	release, elected := reconcilelock.AcquireLeader(t.Context(), dnsLease(js, "peer"))
 	require.True(t, elected, "the peer has to hold the lease or this test proves nothing")
 
-	r := electingReconcilerFor(w, nc, "node1", func() DesiredSet {
+	r := electingReconcilerFor(w, dnsLease(js, "node1"), func() DesiredSet {
 		return DesiredSet{Changes: []Change{upsert("lb-1.elb."+testBase, "10.200.1.9")}}
 	})
-	r.reconcileOnce(t.Context())
+	revisit := r.reconcileOnce(t.Context())
 
 	assert.Equal(t, before, objects[testBase+".toml"],
 		"a node that did not win the election must leave the zone object exactly as it found it")
+	assert.Positive(t, revisit, "the loser asks to retry so a dead holder is taken over")
+	assert.LessOrEqual(t, revisit, reconcilelock.LeaderTTL/3, "the retry comes within a renewal interval")
 
 	release()
 }
@@ -50,10 +59,10 @@ func TestReconcileOnce_ALoserOfTheElectionDoesNotWriteTheZone(t *testing.T) {
 // The counterpart, so the test above is not passing because the pass was broken
 // for some other reason: with the lease free, the same reconciler writes.
 func TestReconcileOnce_TheWinnerOfTheElectionWritesTheZone(t *testing.T) {
-	_, nc, _ := testutil.StartTestJetStream(t)
+	_, _, js := testutil.StartTestJetStream(t)
 	w, objects := newTestWriter(t)
 
-	r := electingReconcilerFor(w, nc, "node1", func() DesiredSet {
+	r := electingReconcilerFor(w, dnsLease(js, "node1"), func() DesiredSet {
 		return DesiredSet{Changes: []Change{upsert("lb-1.elb."+testBase, "10.200.1.9")}}
 	})
 	r.reconcileOnce(t.Context())

spinifex/domains/dns/reconcile_test.go

@@ -224,7 +224,7 @@ func TestReconciler_RunReturnsWhenDisabled(t *testing.T) {
 // path of Run: the startup pass must fire before any watch or resync activity,
 // and cancelling ctx must still let the loop return.
 func TestReconciler_RunPerformsStartupPassThenStopsOnCancel(t *testing.T) {
-	_, nc, _ := testutil.StartTestJetStream(t)
+	_, _, js := testutil.StartTestJetStream(t)
 	endpoint, _ := fakeS3(t, "northstar")
 
 	var passes atomic.Int64
@@ -235,8 +235,7 @@ func TestReconciler_RunPerformsStartupPassThenStopsOnCancel(t *testing.T) {
 			Endpoint: endpoint, Bucket: "northstar", Region: "us-east-1",
 			AccessKey: "SYSTEM", SecretKey: "SYSTEMSECRET",
 		},
-		nc:     nc,
-		holder: "test-node",
+		lease: dnsLease(js, "test-node"),
 		desired: func() DesiredSet {
 			passes.Add(1)
 			return DesiredSet{}
@@ -268,7 +267,7 @@ func TestReconciler_RunPerformsStartupPassThenStopsOnCancel(t *testing.T) {
 // acquirable afterwards, not merely on the call returning, since it is the
 // leadership and not the pass that the rest of the cluster waits on.
 func TestReconcileOnce_AStuckPassSurrendersLeadership(t *testing.T) {
-	_, nc, _ := testutil.StartTestJetStream(t)
+	_, _, js := testutil.StartTestJetStream(t)
 	endpoint, _ := fakeS3(t, "northstar")
 
 	blocked := make(chan struct{})
@@ -281,8 +280,7 @@ func TestReconcileOnce_AStuckPassSurrendersLeadership(t *testing.T) {
 			Endpoint: endpoint, Bucket: "northstar", Region: "us-east-1",
 			AccessKey: "SYSTEM", SecretKey: "SYSTEMSECRET",
 		},
-		nc:             nc,
-		holder:         "stuck-node",
+		lease:          dnsLease(js, "stuck-node"),
 		computeTimeout: 100 * time.Millisecond,
 		desired: func() DesiredSet {
 			<-blocked
@@ -302,7 +300,7 @@ func TestReconcileOnce_AStuckPassSurrendersLeadership(t *testing.T) {
 		t.Fatal("reconcileOnce never returned: a stuck pass still holds the lease forever")
 	}
 
-	release, elected := reconcilelock.AcquireLeader(t.Context(), nc, KVBucketDNSReconcile, "next-node")
+	release, elected := reconcilelock.AcquireLeader(t.Context(), dnsLease(js, "next-node"))
 	require.True(t, elected, "another node must be able to take the next cycle")
 	release()
 }

spinifex/domains/dns/types.go

@@ -11,9 +11,9 @@ const (
 	// PUT so every northstar instance reloads just that zone immediately, instead
 	// of waiting for the next S3 sync poll. No queue group: all servers consume.
 	SubjectZoneReload = "dns.zone.reload"
-	// KVBucketDNSReconcile elects the one node that writes the zones for a drift
-	// cycle. Its own bucket, so the election never blocks another reconcile loop.
-	KVBucketDNSReconcile = "spinifex-dns-reconcile"
+	// KeyReconcile elects the one node that writes the zones for a drift cycle.
+	// Its own key, so the election never blocks another reconcile loop.
+	KeyReconcile = "dns.reconcile"
 	// PrivateZone is the fixed AWS-parity private DNS zone (IMDS synthHostname).
 	PrivateZone = "compute.internal"
 	// DefaultTTL is applied when a change omits a TTL.

spinifex/domains/ec2/account/seal.go

@@ -8,7 +8,7 @@ import (
 
 	awserrors "github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -109,7 +109,7 @@ func (s *AccountSettingsServiceImpl) SealAccount(ctx context.Context, accountID
 		switch {
 		case writeErr == nil:
 			return &AccountSettingsSeal{Revision: newRev}, nil
-		case kvutil.IsRevisionConflict(writeErr):
+		case kvstore.IsRevisionConflict(writeErr):
 			continue
 		default:
 			slog.ErrorContext(ctx, "Failed to write sealed account settings record", "accountID", accountID, "err", writeErr)

spinifex/domains/ec2/account/service_impl.go

@@ -13,7 +13,7 @@ import (
 	ebsencryption "github.com/mulgadc/spinifex/spinifex/domains/ec2/ebs/encryption"
 	awserrors "github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -57,7 +57,7 @@ func NewAccountSettingsServiceImplWithNATS(ctx context.Context, cfg *config.Conf
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	settingsKV, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketAccountSettings, 10)
+	settingsKV, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketAccountSettings, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create account settings KV bucket: %w", err)
 	}
@@ -115,13 +115,13 @@ func mapSettingsError(ctx context.Context, accountID string, err error) error {
 		return errors.New(awserrors.ErrorBlocked)
 	case errors.Is(err, errSettingsContended):
 		slog.ErrorContext(ctx, "Account settings CAS retries exhausted under contention", "accountID", accountID)
-	case errors.Is(err, kvutil.ErrRead):
+	case errors.Is(err, kvstore.ErrRead):
 		slog.ErrorContext(ctx, "Failed to read account settings from KV", "accountID", accountID, "err", err)
-	case errors.Is(err, kvutil.ErrDecode):
+	case errors.Is(err, kvstore.ErrDecode):
 		slog.ErrorContext(ctx, "Corrupt account settings record in KV", "accountID", accountID, "err", err)
-	case errors.Is(err, kvutil.ErrEncode):
+	case errors.Is(err, kvstore.ErrEncode):
 		slog.ErrorContext(ctx, "Failed to marshal account settings record", "accountID", accountID, "err", err)
-	case errors.Is(err, kvutil.ErrWrite):
+	case errors.Is(err, kvstore.ErrWrite):
 		slog.ErrorContext(ctx, "Failed to write account settings to KV", "accountID", accountID, "err", err)
 	default:
 		slog.ErrorContext(ctx, "Account settings operation failed", "accountID", accountID, "err", err)
@@ -133,7 +133,7 @@ func mapSettingsError(ctx context.Context, accountID string, err error) error {
 // changed is always true, so every success is a leader-accepted write over an
 // active value; a seal landing between the read and the write fails the CAS.
 func (s *AccountSettingsServiceImpl) mutateSettings(ctx context.Context, accountID string, set func(*AccountSettingsRecord)) error {
-	_, err := kvutil.Update(ctx, s.settingsKV, settingsKey(accountID), kvutil.CASConfig{
+	_, err := kvstore.Update(ctx, s.settingsKV, settingsKey(accountID), kvstore.CASConfig{
 		CreateIfAbsent: true,
 		Exhausted:      func(string, int) error { return errSettingsContended },
 	}, func(rec *AccountSettingsRecord) (bool, error) {

spinifex/domains/ec2/accountcleanup/remainder.go

@@ -1,155 +0,0 @@
-package accountcleanup
-
-import (
-	"context"
-	"log/slog"
-	"slices"
-
-	"github.com/aws/aws-sdk-go/aws"
-	"github.com/aws/aws-sdk-go/service/ec2"
-	awserrors "github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
-)
-
-// ReasonRecordRemains marks a resource whose record the owner still holds.
-const ReasonRecordRemains = "record-remains"
-
-// ComputeRemainder reads what compute the account still holds, changing nothing.
-type ComputeRemainder interface {
-	ComputeRemaining(ctx context.Context, accountID string) (*Outcome, error)
-}
-
-// StorageRemainder reads what volumes and tag residue the account still holds,
-// changing nothing.
-type StorageRemainder interface {
-	StorageRemaining(ctx context.Context, accountID string) (*Outcome, error)
-}
-
-var (
-	_ ComputeRemainder = (*ComputeCleanup)(nil)
-	_ StorageRemainder = (*StorageCleanup)(nil)
-)
-
-func validRemainderAccount(accountID string) error {
-	if accountID == "" {
-		return awserrors.Errorf(awserrors.ErrorMissingParameter, "The request must specify an account ID.")
-	}
-	if !awsidentifiers.IsAccountID(accountID) {
-		return awserrors.Errorf(awserrors.ErrorInvalidParameterValue, "The account ID must be 12 digits.")
-	}
-	return nil
-}
-
-// ComputeRemaining lists every instance the account still holds and every spot
-// request that can still launch. A read that does not complete is an incomplete
-// reason, never a shorter list. The result is a snapshot, not evidence of absence.
-func (c *ComputeCleanup) ComputeRemaining(ctx context.Context, accountID string) (*Outcome, error) {
-	if err := validRemainderAccount(accountID); err != nil {
-		return nil, err
-	}
-	if c == nil || c.spot == nil || c.instances == nil || c.teardown == nil {
-		return nil, awserrors.Errorf(awserrors.ErrorServerInternal, "Compute cleanup is not fully wired.")
-	}
-	res := &Outcome{AccountID: accountID}
-	defer res.finish()
-
-	scan, err := c.spot.ScanAccount(ctx, accountID)
-	if err != nil {
-		slog.WarnContext(ctx, "ComputeRemaining: spot scan failed", "account_id", accountID, "err", err)
-		res.incomplete(ReasonSpotInventoryIncomplete)
-	} else {
-		for _, r := range slices.Concat(scan.Active, scan.Terminal) {
-			if !terminalSpotState(r.State) {
-				res.Items = append(res.Items, Item{ResourceID: r.RequestID, Kind: KindSpotRequest, State: r.State, Disposition: DispositionResidue, Reason: ReasonRecordRemains})
-			}
-		}
-	}
-
-	inventory, err := c.instances.AccountInstances(ctx, accountID)
-	if err != nil || !inventoryUsable(inventory) {
-		slog.WarnContext(ctx, "ComputeRemaining: instance inventory incomplete", "account_id", accountID, "err", err)
-		res.incomplete(ReasonInstanceInventoryIncomplete)
-	} else {
-		for _, reservation := range inventory.Reservations {
-			for _, in := range reservation.Instances {
-				state := ""
-				if in.State != nil {
-					state = aws.StringValue(in.State.Name)
-				}
-				res.Items = append(res.Items, Item{ResourceID: aws.StringValue(in.InstanceId), Kind: KindInstance, State: state, Disposition: DispositionResidue, Reason: ReasonRecordRemains})
-			}
-		}
-	}
-
-	outstanding, err := c.teardown.AccountTeardownOutstanding(accountID)
-	if err != nil {
-		slog.WarnContext(ctx, "ComputeRemaining: teardown unread", "account_id", accountID, "err", err)
-		res.incomplete(ReasonTeardownUnread)
-	} else {
-		for _, o := range outstanding {
-			if !res.hasInstance(o.InstanceID) {
-				res.Items = append(res.Items, Item{ResourceID: o.InstanceID, Kind: KindInstance, State: o.State, Disposition: DispositionResidue, Reason: ReasonTerminating})
-			}
-		}
-	}
-	return res, nil
-}
-
-// StorageRemaining lists every volume record the account still holds, deleting
-// ones included, and the tag entries of resources whose owner holds no record.
-func (c *StorageCleanup) StorageRemaining(ctx context.Context, accountID string) (*Outcome, error) {
-	if err := validRemainderAccount(accountID); err != nil {
-		return nil, err
-	}
-	if c == nil || c.volumes == nil || c.instances == nil || c.teardown == nil || c.spot == nil || c.tags == nil {
-		return nil, awserrors.Errorf(awserrors.ErrorServerInternal, "Storage cleanup is not fully wired.")
-	}
-	res := &Outcome{AccountID: accountID}
-	defer res.finish()
-
-	volumes, err := c.volumes.VolumeInventory(ctx, accountID)
-	if err != nil {
-		slog.WarnContext(ctx, "StorageRemaining: volume inventory incomplete", "account_id", accountID, "err", err)
-		res.incomplete(ReasonVolumeInventoryIncomplete)
-	}
-	for _, v := range volumes {
-		reason := ReasonRecordRemains
-		if v.State == ec2.VolumeStateDeleting {
-			reason = ReasonVolumeDeleting
-		}
-		res.Items = append(res.Items, Item{ResourceID: v.VolumeID, Kind: KindVolume, State: v.State, Disposition: DispositionResidue, Reason: reason})
-	}
-
-	live, err := c.instances.AccountInstances(ctx, accountID)
-	if err != nil || !inventoryUsable(live) {
-		slog.WarnContext(ctx, "StorageRemaining: no complete instance view", "account_id", accountID, "err", err)
-		res.incomplete(ReasonHolderProofIncomplete)
-		return res, nil
-	}
-	outstanding, err := c.teardown.AccountTeardownOutstanding(accountID)
-	if err != nil {
-		slog.WarnContext(ctx, "StorageRemaining: teardown unread", "account_id", accountID, "err", err)
-		res.incomplete(ReasonHolderProofIncomplete)
-		return res, nil
-	}
-
-	tagged, err := c.tags.ScanAccountResourceIDs(ctx, accountID, storageTagPrefixes...)
-	if err != nil {
-		slog.WarnContext(ctx, "StorageRemaining: tag index read failed", "account_id", accountID, "err", err)
-		res.incomplete(ReasonTagIndexUnread)
-		return res, nil
-	}
-	owned, resolved := c.ownedResourceIDs(ctx, accountID, live, outstanding)
-	for _, id := range tagged {
-		switch {
-		case !resolved[tagPrefix(id)]:
-			res.incomplete(ReasonTagOwnerUnresolved)
-			res.Items = append(res.Items, Item{ResourceID: id, Kind: KindTag, Disposition: DispositionResidue, Reason: ReasonOwnerRecordsUnresolved})
-		case owned[id]:
-			// The resource itself is reported; its tags are not residue.
-		default:
-			res.Items = append(res.Items, Item{ResourceID: id, Kind: KindTag, Disposition: DispositionResidue, Reason: ReasonRecordRemains})
-		}
-	}
-	return res, nil
-}

spinifex/domains/ec2/accountcleanup/remainder_test.go

@@ -1,97 +0,0 @@
-package accountcleanup
-
-import (
-	"errors"
-	"testing"
-
-	"github.com/aws/aws-sdk-go/service/ec2"
-	"github.com/mulgadc/spinifex/spinifex/domains/ec2/accountinventory"
-	ec2instance "github.com/mulgadc/spinifex/spinifex/domains/ec2/instance"
-	ec2spotinstance "github.com/mulgadc/spinifex/spinifex/domains/ec2/spotinstance"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestComputeRemainingReportsEveryHeldInstanceAndOpenSpot(t *testing.T) {
-	r := newComputeRig()
-	r.spot.scans = []*ec2spotinstance.AccountScan{{
-		Active:   []ec2spotinstance.AccountSpotRequest{{RequestID: "sir-0aaaaaaaaaaaaaaa1", State: ec2.SpotInstanceStateOpen}},
-		Terminal: []ec2spotinstance.AccountSpotRequest{{RequestID: "sir-0aaaaaaaaaaaaaaa2", State: ec2.SpotInstanceStateClosed}},
-	}}
-	r.teardown.outstanding = []ec2instance.OutstandingInstance{{InstanceID: "i-0aaaaaaaaaaaaaa99", State: ec2.InstanceStateNameTerminated}}
-
-	res, err := r.cleanup().ComputeRemaining(t.Context(), testAccount)
-	require.NoError(t, err)
-
-	assert.Equal(t, StatusIncomplete, res.Status)
-	assert.NotNil(t, itemFor(res, runningID))
-	assert.NotNil(t, itemFor(res, stoppedID))
-	assert.NotNil(t, itemFor(res, "i-0aaaaaaaaaaaaaa99"))
-	assert.NotNil(t, itemFor(res, "sir-0aaaaaaaaaaaaaaa1"))
-	assert.Nil(t, itemFor(res, "sir-0aaaaaaaaaaaaaaa2"), "a closed request launches nothing")
-	assert.Empty(t, r.terminator.terminated, "a re-read changes nothing")
-}
-
-func TestComputeRemainingCleanWhenNothingHeld(t *testing.T) {
-	r := newComputeRig()
-	r.inventory.out = listing(nil)
-	res, err := r.cleanup().ComputeRemaining(t.Context(), testAccount)
-	require.NoError(t, err)
-	assert.Equal(t, StatusClean, res.Status)
-}
-
-func TestComputeRemainingUnreadStoreIsIncompleteNeverClean(t *testing.T) {
-	for name, mutate := range map[string]func(*computeRig){
-		"instances": func(r *computeRig) { r.inventory.out, r.inventory.err = nil, errors.New("node silent") },
-		"spot":      func(r *computeRig) { r.spot.scanErr = errors.New("spot unreadable") },
-		"teardown":  func(r *computeRig) { r.teardown.err = errors.New("no responder") },
-	} {
-		t.Run(name, func(t *testing.T) {
-			r := newComputeRig()
-			r.inventory.out = listing(nil)
-			mutate(r)
-			res, err := r.cleanup().ComputeRemaining(t.Context(), testAccount)
-			require.NoError(t, err)
-			assert.Equal(t, StatusIncomplete, res.Status)
-			assert.NotEmpty(t, res.IncompleteReasons)
-		})
-	}
-}
-
-func TestStorageRemainingReportsVolumeRecordsAndTagResidue(t *testing.T) {
-	r := newStorageRig()
-	r.volumes.vols = []accountinventory.VolumeItem{
-		volume(freeVol, ec2.VolumeStateAvailable, ""),
-		volume("vol-0aaaaaaaaaaaaaaa2", ec2.VolumeStateDeleting, ""),
-	}
-	r.tags.ids = []string{"i-0aaaaaaaaaaaaaaaf", freeVol}
-
-	res, err := r.cleanup().StorageRemaining(t.Context(), testAccount)
-	require.NoError(t, err)
-
-	assert.Equal(t, StatusIncomplete, res.Status)
-	assert.Equal(t, ReasonRecordRemains, itemFor(res, freeVol).Reason)
-	assert.Equal(t, ReasonVolumeDeleting, itemFor(res, "vol-0aaaaaaaaaaaaaaa2").Reason)
-	assert.Equal(t, KindTag, itemFor(res, "i-0aaaaaaaaaaaaaaaf").Kind, "tags of a resource with no record are residue")
-	assert.Empty(t, r.deleter.deleted, "a re-read deletes nothing")
-	assert.Empty(t, r.tags.deleted, "a re-read clears no tag")
-}
-
-func TestStorageRemainingUnreadStoreIsIncompleteNeverClean(t *testing.T) {
-	for name, mutate := range map[string]func(*storageRig){
-		"volumes":   func(r *storageRig) { r.volumes.failAt = -1 },
-		"instances": func(r *storageRig) { r.instances.out, r.instances.err = nil, errors.New("node silent") },
-		"teardown":  func(r *storageRig) { r.teardown.err = errors.New("no responder") },
-		"tags":      func(r *storageRig) { r.tags.scanErr = errors.New("tag index unreadable") },
-	} {
-		t.Run(name, func(t *testing.T) {
-			r := newStorageRig()
-			r.volumes.vols = nil
-			mutate(r)
-			res, err := r.cleanup().StorageRemaining(t.Context(), testAccount)
-			require.NoError(t, err)
-			assert.Equal(t, StatusIncomplete, res.Status)
-			assert.NotEmpty(t, res.IncompleteReasons)
-		})
-	}
-}

spinifex/domains/ec2/awsapi/instance/AccountTeardownOutstanding.go

@@ -1,28 +0,0 @@
-package instance
-
-import (
-	"context"
-	"fmt"
-	"time"
-
-	ec2instance "github.com/mulgadc/spinifex/spinifex/domains/ec2/instance"
-	natsmsg "github.com/mulgadc/spinifex/spinifex/foundation/messaging/nats"
-	"github.com/nats-io/nats.go"
-)
-
-// AccountTeardownOutstandingSubject serves the instances of an account whose
-// teardown is not complete, from the shared record store.
-const AccountTeardownOutstandingSubject = "ec2.AccountTeardownOutstanding"
-
-var accountTeardownOutstandingTimeout = 15 * time.Second
-
-// AccountTeardownOutstanding reads the account's unfinished instance teardowns.
-// No responder, a timeout or an error reply is an error, never an empty list.
-func AccountTeardownOutstanding(ctx context.Context, natsConn *nats.Conn, accountID string) ([]ec2instance.OutstandingInstance, error) {
-	out, err := natsmsg.NATSRequest[ec2instance.TeardownOutstandingOutput](ctx, natsConn, AccountTeardownOutstandingSubject,
-		&ec2instance.TeardownOutstandingInput{}, accountTeardownOutstandingTimeout, accountID)
-	if err != nil {
-		return nil, fmt.Errorf("read outstanding teardown: %w", err)
-	}
-	return out.Instances, nil
-}

spinifex/domains/ec2/awsapi/instance/AccountTeardownOutstanding_test.go

@@ -1,88 +0,0 @@
-package instance
-
-import (
-	"context"
-	"encoding/json"
-	"testing"
-	"time"
-
-	ec2instance "github.com/mulgadc/spinifex/spinifex/domains/ec2/instance"
-	awserrors "github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/nats-io/nats.go"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func serveTeardownOutstanding(t *testing.T, nc *nats.Conn, reply []byte) {
-	t.Helper()
-	_, err := nc.Subscribe(AccountTeardownOutstandingSubject, func(msg *nats.Msg) { _ = msg.Respond(reply) })
-	require.NoError(t, err)
-	require.NoError(t, nc.Flush())
-}
-
-func TestAccountTeardownOutstanding_ReturnsTheServedInstances(t *testing.T) {
-	t.Parallel()
-	_, nc := startTestNATSServer(t)
-	reply, err := json.Marshal(ec2instance.TeardownOutstandingOutput{Instances: []ec2instance.OutstandingInstance{{InstanceID: "i-1"}}})
-	require.NoError(t, err)
-	serveTeardownOutstanding(t, nc, reply)
-
-	got, err := AccountTeardownOutstanding(context.Background(), nc, "123456789012")
-
-	require.NoError(t, err)
-	require.Len(t, got, 1)
-	assert.Equal(t, "i-1", got[0].InstanceID)
-}
-
-// A read that cannot be answered must never look like an empty list, because an
-// empty list lets the storage cleanup believe no QEMU still holds a volume.
-func TestAccountTeardownOutstanding_UnansweredReadIsAnError(t *testing.T) {
-	t.Parallel()
-
-	t.Run("no responder", func(t *testing.T) {
-		t.Parallel()
-		_, nc := startTestNATSServer(t)
-
-		got, err := AccountTeardownOutstanding(context.Background(), nc, "123456789012")
-
-		require.Error(t, err)
-		assert.Nil(t, got)
-		assert.ErrorIs(t, err, nats.ErrNoResponders)
-	})
-	t.Run("error reply", func(t *testing.T) {
-		t.Parallel()
-		_, nc := startTestNATSServer(t)
-		serveTeardownOutstanding(t, nc, awserrors.GenerateErrorPayload(awserrors.ErrorServerInternal))
-
-		got, err := AccountTeardownOutstanding(context.Background(), nc, "123456789012")
-
-		require.Error(t, err)
-		assert.Nil(t, got)
-	})
-	t.Run("unreadable reply", func(t *testing.T) {
-		t.Parallel()
-		_, nc := startTestNATSServer(t)
-		serveTeardownOutstanding(t, nc, []byte("not json"))
-
-		got, err := AccountTeardownOutstanding(context.Background(), nc, "123456789012")
-
-		require.Error(t, err)
-		assert.Nil(t, got)
-	})
-}
-
-// Not parallel: it shortens the package timeout.
-func TestAccountTeardownOutstanding_TimeoutIsAnError(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	_, err := nc.Subscribe(AccountTeardownOutstandingSubject, func(*nats.Msg) {})
-	require.NoError(t, err)
-	require.NoError(t, nc.Flush())
-	previous := accountTeardownOutstandingTimeout
-	accountTeardownOutstandingTimeout = 200 * time.Millisecond
-	t.Cleanup(func() { accountTeardownOutstandingTimeout = previous })
-
-	got, err := AccountTeardownOutstanding(context.Background(), nc, "123456789012")
-
-	require.Error(t, err)
-	assert.Nil(t, got)
-}

spinifex/domains/ec2/awsapi/instance/ModifyInstanceAttribute.go

@@ -11,7 +11,6 @@ import (
 	"time"
 
 	"github.com/aws/aws-sdk-go/service/ec2"
-	"github.com/mulgadc/spinifex/contracts/ec2/v1"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	"github.com/nats-io/nats.go"
 )
@@ -66,20 +65,6 @@ func ModifyInstanceAttribute(ctx context.Context, input *ec2.ModifyInstanceAttri
 
 	slog.InfoContext(ctx, "ModifyInstanceAttribute: Processing request", "instance_id", *input.InstanceId)
 
-	// Termination protection on a running instance lives on its owner, so it
-	// goes there first. A nil value is not a value and keeps the queue path.
-	if input.DisableApiTermination != nil && input.DisableApiTermination.Value != nil {
-		cmd := ec2v1.EC2InstanceCommand{
-			ID:                        *input.InstanceId,
-			Attributes:                ec2v1.EC2CommandAttributes{SetDisableApiTermination: true},
-			DisableApiTerminationData: &ec2v1.DisableApiTerminationData{Value: *input.DisableApiTermination.Value},
-		}
-		if _, err := routeInstanceMutation[ec2.ModifyInstanceAttributeOutput](ctx, natsConn, *input.InstanceId, accountID, cmd, "ec2.ModifyInstanceAttribute", input); err != nil {
-			return ec2.ModifyInstanceAttributeOutput{}, err
-		}
-		return ec2.ModifyInstanceAttributeOutput{}, nil
-	}
-
 	jsonData, err := json.Marshal(input)
 	if err != nil {
 		slog.ErrorContext(ctx, "ModifyInstanceAttribute: Failed to marshal request", "instance_id", *input.InstanceId, "err", err)

spinifex/domains/ec2/awsapi/instance/ModifyInstanceAttribute_test.go

@@ -3,13 +3,10 @@ package instance
 import (
 	"context"
 	"encoding/json"
-	"sync"
-	"sync/atomic"
 	"testing"
 
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
-	"github.com/mulgadc/spinifex/contracts/ec2/v1"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	"github.com/nats-io/nats.go"
 	"github.com/stretchr/testify/assert"
@@ -229,53 +226,3 @@ func TestModifyInstanceAttribute_ValidationFailure(t *testing.T) {
 	require.Error(t, err)
 	assert.Equal(t, awserrors.ErrorInvalidParameterValue, err.Error())
 }
-
-func TestModifyInstanceAttribute_DisableApiTerminationRouting(t *testing.T) {
-	const id = "i-0123456789abcdef0"
-	_, nc := startTestNATSServer(t)
-
-	var gotValues []bool
-	var mu sync.Mutex
-	_, err := nc.Subscribe(ec2v1.InstanceCommandSubject(id), func(msg *nats.Msg) {
-		var cmd ec2v1.EC2InstanceCommand
-		if json.Unmarshal(msg.Data, &cmd) == nil && cmd.Attributes.SetDisableApiTermination && cmd.DisableApiTerminationData != nil {
-			mu.Lock()
-			gotValues = append(gotValues, cmd.DisableApiTerminationData.Value)
-			mu.Unlock()
-		}
-		reply := nats.NewMsg(msg.Reply)
-		reply.Data = []byte(`{}`)
-		reply.Header.Set(ec2v1.HandledHeader, "1")
-		_ = msg.RespondMsg(reply)
-	})
-	require.NoError(t, err)
-	var queue atomic.Int32
-	_, err = nc.Subscribe("ec2.ModifyInstanceAttribute", func(msg *nats.Msg) {
-		queue.Add(1)
-		_ = msg.Respond([]byte(`{}`))
-	})
-	require.NoError(t, err)
-
-	for _, v := range []bool{true, false} {
-		_, err := ModifyInstanceAttribute(context.Background(), &ec2.ModifyInstanceAttributeInput{
-			InstanceId:            aws.String(id),
-			DisableApiTermination: &ec2.AttributeBooleanValue{Value: aws.Bool(v)},
-		}, nc, "acct")
-		require.NoError(t, err)
-	}
-	mu.Lock()
-	assert.Equal(t, []bool{true, false}, gotValues, "explicit true and false both reach the owner")
-	mu.Unlock()
-	assert.Zero(t, queue.Load())
-
-	// A nil value is not a value: queue path, no owner command.
-	_, err = ModifyInstanceAttribute(context.Background(), &ec2.ModifyInstanceAttributeInput{
-		InstanceId:            aws.String(id),
-		DisableApiTermination: &ec2.AttributeBooleanValue{},
-	}, nc, "acct")
-	require.NoError(t, err)
-	assert.Equal(t, int32(1), queue.Load())
-	mu.Lock()
-	assert.Len(t, gotValues, 2, "no owner command for a nil value")
-	mu.Unlock()
-}

spinifex/domains/ec2/awsapi/instance/ModifyInstanceMetadataOptions.go

@@ -3,11 +3,11 @@ package instance
 import (
 	"context"
 	"errors"
+	"github.com/mulgadc/spinifex/spinifex/foundation/messaging/nats"
 	"strings"
+	"time"
 
-	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
-	"github.com/mulgadc/spinifex/contracts/ec2/v1"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	"github.com/nats-io/nats.go"
 )
@@ -22,16 +22,5 @@ func ModifyInstanceMetadataOptions(ctx context.Context, input *ec2.ModifyInstanc
 	if !strings.HasPrefix(*input.InstanceId, "i-") {
 		return nil, errors.New(awserrors.ErrorInvalidInstanceIDMalformed)
 	}
-	cmd := ec2v1.EC2InstanceCommand{
-		ID:         *input.InstanceId,
-		Attributes: ec2v1.EC2CommandAttributes{ModifyInstanceMetadataOptions: true},
-		InstanceMetadataOptionsData: &ec2v1.InstanceMetadataOptionsData{
-			HttpTokens:              aws.StringValue(input.HttpTokens),
-			HttpEndpoint:            aws.StringValue(input.HttpEndpoint),
-			HttpProtocolIpv6:        aws.StringValue(input.HttpProtocolIpv6),
-			InstanceMetadataTags:    aws.StringValue(input.InstanceMetadataTags),
-			HttpPutResponseHopLimit: input.HttpPutResponseHopLimit,
-		},
-	}
-	return routeInstanceMutation[ec2.ModifyInstanceMetadataOptionsOutput](ctx, natsConn, *input.InstanceId, accountID, cmd, "ec2.ModifyInstanceMetadataOptions", input)
+	return natsmsg.NATSRequest[ec2.ModifyInstanceMetadataOptionsOutput](ctx, natsConn, "ec2.ModifyInstanceMetadataOptions", input, 30*time.Second, accountID)
 }

spinifex/domains/ec2/awsapi/instance/ModifyInstanceMetadataOptions_test.go

@@ -2,16 +2,11 @@ package instance
 
 import (
 	"context"
-	"encoding/json"
-	"sync/atomic"
 	"testing"
-	"time"
 
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
-	"github.com/mulgadc/spinifex/contracts/ec2/v1"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/nats-io/nats.go"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -35,179 +30,3 @@ func TestModifyInstanceMetadataOptions_RejectsMalformedID(t *testing.T) {
 		})
 	}
 }
-
-const metaTestID = "i-0123456789abcdef0"
-
-func metaInput(hop int64) *ec2.ModifyInstanceMetadataOptionsInput {
-	return &ec2.ModifyInstanceMetadataOptionsInput{
-		InstanceId:              aws.String(metaTestID),
-		HttpPutResponseHopLimit: aws.Int64(hop),
-	}
-}
-
-// ownerReply answers one ec2.cmd request, optionally carrying the handled marker.
-func ownerReply(t *testing.T, nc *nats.Conn, handled bool, body []byte, calls *atomic.Int32) {
-	t.Helper()
-	_, err := nc.Subscribe(ec2v1.InstanceCommandSubject(metaTestID), func(msg *nats.Msg) {
-		calls.Add(1)
-		reply := nats.NewMsg(msg.Reply)
-		reply.Data = body
-		if handled {
-			reply.Header.Set(ec2v1.HandledHeader, "1")
-		}
-		_ = msg.RespondMsg(reply)
-	})
-	require.NoError(t, err)
-}
-
-func queueReply(t *testing.T, nc *nats.Conn, body []byte, calls *atomic.Int32) {
-	t.Helper()
-	_, err := nc.QueueSubscribe("ec2.ModifyInstanceMetadataOptions", "spinifex-workers", func(msg *nats.Msg) {
-		calls.Add(1)
-		_ = msg.Respond(body)
-	})
-	require.NoError(t, err)
-}
-
-func metaOutputJSON(t *testing.T, hop int64) []byte {
-	t.Helper()
-	b, err := json.Marshal(&ec2.ModifyInstanceMetadataOptionsOutput{
-		InstanceId:              aws.String(metaTestID),
-		InstanceMetadataOptions: &ec2.InstanceMetadataOptionsResponse{HttpPutResponseHopLimit: aws.Int64(hop)},
-	})
-	require.NoError(t, err)
-	return b
-}
-
-func TestModifyInstanceMetadataOptions_OwnerAnswerReturnedAsIs(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	var owner, queue atomic.Int32
-	ownerReply(t, nc, true, metaOutputJSON(t, 3), &owner)
-	queueReply(t, nc, []byte(`{}`), &queue)
-
-	out, err := ModifyInstanceMetadataOptions(context.Background(), metaInput(3), nc, "acct")
-	require.NoError(t, err)
-	assert.Equal(t, int64(3), aws.Int64Value(out.InstanceMetadataOptions.HttpPutResponseHopLimit))
-	assert.Equal(t, int32(1), owner.Load())
-	assert.Zero(t, queue.Load())
-}
-
-func TestModifyInstanceMetadataOptions_OwnerErrorWithHeaderIsNotRetried(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	var owner, queue atomic.Int32
-	ownerReply(t, nc, true, awserrors.GenerateErrorPayload(awserrors.ErrorServerInternal), &owner)
-	queueReply(t, nc, metaOutputJSON(t, 3), &queue)
-
-	_, err := ModifyInstanceMetadataOptions(context.Background(), metaInput(3), nc, "acct")
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
-	assert.Zero(t, queue.Load())
-}
-
-func TestModifyInstanceMetadataOptions_OwnerValidationErrorRelayed(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	var owner, queue atomic.Int32
-	ownerReply(t, nc, true, awserrors.GenerateErrorPayload(awserrors.ErrorUnsupportedOperation), &owner)
-	queueReply(t, nc, metaOutputJSON(t, 3), &queue)
-
-	_, err := ModifyInstanceMetadataOptions(context.Background(), metaInput(3), nc, "acct")
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorUnsupportedOperation, err.Error())
-	assert.Zero(t, queue.Load())
-}
-
-func TestModifyInstanceMetadataOptions_HeaderlessOtherErrorReturnedAsIs(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	var owner, queue atomic.Int32
-	ownerReply(t, nc, false, awserrors.GenerateErrorPayload(awserrors.ErrorInvalidInstanceIDNotFound), &owner)
-	queueReply(t, nc, metaOutputJSON(t, 3), &queue)
-
-	_, err := ModifyInstanceMetadataOptions(context.Background(), metaInput(3), nc, "acct")
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorInvalidInstanceIDNotFound, err.Error())
-	assert.Zero(t, queue.Load())
-}
-
-func TestS9_N1OwnerRejectionFallsBackOnceToQueueGroup(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	var owner, queue atomic.Int32
-	ownerReply(t, nc, false, awserrors.GenerateErrorPayload(awserrors.ErrorServerInternal), &owner)
-	queueReply(t, nc, metaOutputJSON(t, 3), &queue)
-
-	out, err := ModifyInstanceMetadataOptions(context.Background(), metaInput(3), nc, "acct")
-	require.NoError(t, err)
-	assert.Equal(t, int64(3), aws.Int64Value(out.InstanceMetadataOptions.HttpPutResponseHopLimit))
-	assert.Equal(t, int32(1), owner.Load(), "exactly one owner request")
-	assert.Equal(t, int32(1), queue.Load(), "exactly one queue-group request")
-}
-
-func TestModifyInstanceMetadataOptions_NoOwnerUsesQueueGroupAfterThreeAttempts(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	var sleeps atomic.Int32
-	prev := cmdRetrySleep
-	cmdRetrySleep = func(time.Duration) { sleeps.Add(1) }
-	t.Cleanup(func() { cmdRetrySleep = prev })
-
-	var queue atomic.Int32
-	queueReply(t, nc, metaOutputJSON(t, 2), &queue)
-
-	out, err := ModifyInstanceMetadataOptions(context.Background(), metaInput(2), nc, "acct")
-	require.NoError(t, err)
-	assert.Equal(t, int64(2), aws.Int64Value(out.InstanceMetadataOptions.HttpPutResponseHopLimit))
-	assert.Equal(t, int32(2), sleeps.Load(), "three attempts means two backoffs")
-	assert.Equal(t, int32(1), queue.Load())
-}
-
-func TestModifyInstanceMetadataOptions_AbsentIsNotFound(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	var queue atomic.Int32
-	queueReply(t, nc, awserrors.GenerateErrorPayload(awserrors.ErrorInvalidInstanceIDNotFound), &queue)
-
-	_, err := ModifyInstanceMetadataOptions(context.Background(), metaInput(2), nc, "acct")
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorInvalidInstanceIDNotFound, err.Error())
-}
-
-func TestModifyInstanceMetadataOptions_OwnerTimeoutIsServerInternalWithoutFallback(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	prev := instanceCmdTimeout
-	instanceCmdTimeout = 50 * time.Millisecond
-	t.Cleanup(func() { instanceCmdTimeout = prev })
-
-	// The owner applies the change, then withholds its reply.
-	var applied, queue atomic.Int32
-	_, err := nc.Subscribe(ec2v1.InstanceCommandSubject(metaTestID), func(*nats.Msg) { applied.Add(1) })
-	require.NoError(t, err)
-	queueReply(t, nc, metaOutputJSON(t, 3), &queue)
-
-	_, err = ModifyInstanceMetadataOptions(context.Background(), metaInput(3), nc, "acct")
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
-	assert.Equal(t, int32(1), applied.Load(), "the owner saw the command, so the outcome is unknown")
-	assert.Zero(t, queue.Load(), "a timeout never falls back to the queue group")
-}
-
-func TestModifyInstanceMetadataOptions_TimeoutAfterApplyIsUnknownNotRolledBack(t *testing.T) {
-	_, nc := startTestNATSServer(t)
-	prev := instanceCmdTimeout
-	instanceCmdTimeout = 50 * time.Millisecond
-	t.Cleanup(func() { instanceCmdTimeout = prev })
-
-	// The owner persists the change, then the reply is lost.
-	var persisted atomic.Int64
-	_, err := nc.Subscribe(ec2v1.InstanceCommandSubject(metaTestID), func(msg *nats.Msg) {
-		var cmd ec2v1.EC2InstanceCommand
-		if json.Unmarshal(msg.Data, &cmd) == nil && cmd.InstanceMetadataOptionsData != nil {
-			persisted.Store(aws.Int64Value(cmd.InstanceMetadataOptionsData.HttpPutResponseHopLimit))
-		}
-	})
-	require.NoError(t, err)
-	var queue atomic.Int32
-	queueReply(t, nc, metaOutputJSON(t, 1), &queue)
-
-	_, err = ModifyInstanceMetadataOptions(context.Background(), metaInput(3), nc, "acct")
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
-	assert.Zero(t, queue.Load(), "no fallback that could race a live owner")
-	assert.Equal(t, int64(3), persisted.Load(), "the gateway error does not undo what the owner stored")
-}

spinifex/domains/ec2/awsapi/instance/instanceid.go

@@ -99,46 +99,3 @@ func classifyUnownedInstance(ctx context.Context, natsConn *nats.Conn, instanceI
 	}
 	return bucketStateAbsent, allOK
 }
-
-// routeInstanceMutation sends cmd to the instance's owner first and falls back
-// to the queue-group subject only when that is safe: no daemon runs the
-// instance (ErrNoResponders after every retry), or the owner is N-1 and
-// rejected the command without writing (ServerInternal with no HandledHeader).
-// A timeout or other transport error is an unknown outcome and never falls back.
-func routeInstanceMutation[Out any](ctx context.Context, natsConn *nats.Conn, instanceID, accountID string, cmd ec2v1.EC2InstanceCommand, queueSubject string, queueInput any) (*Out, error) {
-	queue := func() (*Out, error) {
-		return natsmsg.NATSRequest[Out](ctx, natsConn, queueSubject, queueInput, 30*time.Second, accountID)
-	}
-
-	data, err := json.Marshal(cmd)
-	if err != nil {
-		return nil, errors.New(awserrors.ErrorServerInternal)
-	}
-	msg, err := requestInstanceCmd(ctx, natsConn, instanceID, accountID, data)
-	if err != nil {
-		if errors.Is(err, nats.ErrNoResponders) {
-			return queue()
-		}
-		slog.ErrorContext(ctx, "Owner request failed, outcome unknown", "instance_id", instanceID, "subject", queueSubject, "err", err)
-		return nil, errors.New(awserrors.ErrorServerInternal)
-	}
-
-	responseError, parseErr := awserrors.ValidateErrorPayload(msg.Data)
-	if parseErr != nil {
-		handled := msg.Header.Get(ec2v1.HandledHeader) == "1"
-		if !handled && *responseError.Code == awserrors.ErrorServerInternal {
-			slog.WarnContext(ctx, "Owner did not recognise the command, using the queue group", "instance_id", instanceID, "subject", queueSubject)
-			return queue()
-		}
-		if responseError.Message != nil && *responseError.Message != "" {
-			return nil, awserrors.Errorf(*responseError.Code, "%s", *responseError.Message)
-		}
-		return nil, errors.New(*responseError.Code)
-	}
-
-	var out Out
-	if err := json.Unmarshal(msg.Data, &out); err != nil {
-		return nil, errors.New(awserrors.ErrorServerInternal)
-	}
-	return &out, nil
-}

spinifex/domains/ec2/eigw/service_impl.go

@@ -19,7 +19,7 @@ import (
 	awsfilters "github.com/mulgadc/spinifex/spinifex/foundation/aws/filters"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
 	awstags "github.com/mulgadc/spinifex/spinifex/foundation/aws/tags"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -56,7 +56,7 @@ func NewEgressOnlyIGWServiceImplWithNATS(ctx context.Context, cfg *config.Config
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	eigwKV, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketEgressOnlyIGW, 10)
+	eigwKV, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketEgressOnlyIGW, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketEgressOnlyIGW, err)
 	}
@@ -65,7 +65,7 @@ func NewEgressOnlyIGWServiceImplWithNATS(ctx context.Context, cfg *config.Config
 	}
 
 	// Get or create VPC KV bucket for cross-resource ownership validation
-	vpcKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
+	vpcKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get VPC KV bucket: %w", err)
 	}
@@ -90,7 +90,7 @@ func (s *EgressOnlyIGWServiceImpl) CreateEgressOnlyInternetGateway(ctx context.C
 		slog.ErrorContext(ctx, "VPC KV unavailable, cannot verify VPC ownership")
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
-	if _, err := s.vpcKV.Get(ctx, kvutil.AccountKey(accountID, *input.VpcId)); err != nil {
+	if _, err := s.vpcKV.Get(ctx, kvstore.AccountKey(accountID, *input.VpcId)); err != nil {
 		slog.WarnContext(ctx, "CreateEgressOnlyInternetGateway: VPC not found for account", "vpcId", *input.VpcId, "accountID", accountID)
 		return nil, awserrors.IDNotFound(awserrors.ErrorInvalidVpcIDNotFound, "vpc", *input.VpcId)
 	}
@@ -109,7 +109,7 @@ func (s *EgressOnlyIGWServiceImpl) CreateEgressOnlyInternetGateway(ctx context.C
 	if err != nil {
 		return nil, fmt.Errorf("failed to marshal Egress-only IGW record: %w", err)
 	}
-	if _, err := s.eigwKV.Put(ctx, kvutil.AccountKey(accountID, eigwID), data); err != nil {
+	if _, err := s.eigwKV.Put(ctx, kvstore.AccountKey(accountID, eigwID), data); err != nil {
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
 
@@ -127,7 +127,7 @@ func (s *EgressOnlyIGWServiceImpl) DeleteEgressOnlyInternetGateway(ctx context.C
 	}
 
 	eigwID := *input.EgressOnlyInternetGatewayId
-	key := kvutil.AccountKey(accountID, eigwID)
+	key := kvstore.AccountKey(accountID, eigwID)
 
 	// Verify the EIGW exists before deleting
 	if _, err := s.eigwKV.Get(ctx, key); err != nil {
@@ -178,7 +178,7 @@ func (s *EgressOnlyIGWServiceImpl) DescribeEgressOnlyInternetGateways(ctx contex
 	}
 
 	for _, key := range keys {
-		if key == kvutil.VersionKey {
+		if key == kvstore.VersionKey {
 			continue
 		}
 		if !strings.HasPrefix(key, prefix) {

spinifex/domains/ec2/eigw/service_impl_test.go

@@ -9,7 +9,7 @@ import (
 	"github.com/mulgadc/spinifex/internal/testkit"
 	ec2vpc "github.com/mulgadc/spinifex/spinifex/domains/ec2/vpc"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -23,7 +23,7 @@ func setupTestEIGWService(t *testing.T) *EgressOnlyIGWServiceImpl {
 	// Create VPC KV bucket and register test VPCs so fail-closed ownership checks pass
 	vpcEntries := map[string][]byte{}
 	for _, vpcID := range []string{"vpc-test123", "vpc-tagged", "vpc-tagged2"} {
-		vpcEntries[kvutil.AccountKey(testAccountID, vpcID)] = []byte(`{"vpc_id":"` + vpcID + `","state":"available"}`)
+		vpcEntries[kvstore.AccountKey(testAccountID, vpcID)] = []byte(`{"vpc_id":"` + vpcID + `","state":"available"}`)
 	}
 	testutil.SeedKV(t, js, ec2vpc.KVBucketVPCs, vpcEntries)
 
@@ -312,7 +312,7 @@ func TestCreateEgressOnlyInternetGateway_CrossAccountVPCRejected(t *testing.T) {
 	// Create VPC KV bucket with a VPC owned by testAccountID
 	vpcID := "vpc-alpha456"
 	testutil.SeedKV(t, js, ec2vpc.KVBucketVPCs, map[string][]byte{
-		kvutil.AccountKey(testAccountID, vpcID): []byte(`{"vpc_id":"vpc-alpha456","state":"available"}`),
+		kvstore.AccountKey(testAccountID, vpcID): []byte(`{"vpc_id":"vpc-alpha456","state":"available"}`),
 	})
 
 	var err error

spinifex/domains/ec2/eip/allocate_idempotency.go

@@ -12,7 +12,7 @@ import (
 
 	"github.com/aws/aws-sdk-go/service/ec2"
 	"github.com/mulgadc/spinifex/spinifex/foundation/lifecycle/idempotency"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -36,7 +36,7 @@ func (s *EIPServiceImpl) allocateOnce(ctx context.Context, accountID string, all
 	if key == "" || s.idemKV == nil {
 		return alloc()
 	}
-	kvKey := kvutil.AccountKey(accountID, idempotencyKeyHash(key))
+	kvKey := kvstore.AccountKey(accountID, idempotencyKeyHash(key))
 
 	if out := s.cachedAllocation(ctx, kvKey); out != nil {
 		slog.WarnContext(ctx, "AllocateAddress: returning first result for a retried request",

spinifex/domains/ec2/eip/lease_rebind.go

@@ -10,7 +10,7 @@ import (
 
 	"github.com/mulgadc/spinifex/spinifex/domains/network/projection"
 	"github.com/mulgadc/spinifex/spinifex/domains/network/topology"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -80,7 +80,7 @@ func (s *EIPServiceImpl) AllocationExists(ctx context.Context, allocationID stri
 	}
 	suffix := "." + allocationID
 	for _, k := range keys {
-		if k != kvutil.VersionKey && strings.HasSuffix(k, suffix) {
+		if k != kvstore.VersionKey && strings.HasSuffix(k, suffix) {
 			return true, nil
 		}
 	}
@@ -98,7 +98,7 @@ func (s *EIPServiceImpl) findByAllocationIDAnyAccount(ctx context.Context, alloc
 
 	suffix := "." + allocationID
 	for _, k := range keys {
-		if k == kvutil.VersionKey || !strings.HasSuffix(k, suffix) {
+		if k == kvstore.VersionKey || !strings.HasSuffix(k, suffix) {
 			continue
 		}
 		entry, err := s.eipKV.Get(ctx, k)

spinifex/domains/ec2/eip/service_impl.go

@@ -20,7 +20,7 @@ import (
 	awsfilters "github.com/mulgadc/spinifex/spinifex/foundation/aws/filters"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
 	awstags "github.com/mulgadc/spinifex/spinifex/foundation/aws/tags"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -66,7 +66,7 @@ func NewEIPServiceImpl(ctx context.Context, natsConn *nats.Conn, externalIPAM *e
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	eipKV, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketEIPs, 10)
+	eipKV, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketEIPs, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketEIPs, err)
 	}
@@ -76,7 +76,7 @@ func NewEIPServiceImpl(ctx context.Context, natsConn *nats.Conn, externalIPAM *e
 
 	// TTL'd, so a failure to create it is not worth refusing to serve EIPs over:
 	// a nil bucket only costs post-completion dedupe.
-	idemKV, err := kvutil.GetOrCreateBucketWithTTL(ctx, js, KVBucketEIPIdempotency, 1, eipIdempotencyTTL)
+	idemKV, err := kvstore.GetOrCreateBucketWithTTL(ctx, js, KVBucketEIPIdempotency, 1, eipIdempotencyTTL)
 	if err != nil {
 		slog.Warn("EIP service: idempotency bucket unavailable; a retried AllocateAddress may allocate twice",
 			"bucket", KVBucketEIPIdempotency, "err", err)
@@ -146,7 +146,7 @@ func (s *EIPServiceImpl) allocateAddress(ctx context.Context, input *ec2.Allocat
 	if err != nil {
 		return nil, fmt.Errorf("failed to marshal EIP record: %w", err)
 	}
-	if _, err := s.eipKV.Put(ctx, kvutil.AccountKey(accountID, allocID), data); err != nil {
+	if _, err := s.eipKV.Put(ctx, kvstore.AccountKey(accountID, allocID), data); err != nil {
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
 
@@ -167,7 +167,7 @@ func (s *EIPServiceImpl) ReleaseAddress(ctx context.Context, input *ec2.ReleaseA
 	}
 
 	allocID := *input.AllocationId
-	key := kvutil.AccountKey(accountID, allocID)
+	key := kvstore.AccountKey(accountID, allocID)
 
 	entry, err := s.eipKV.Get(ctx, key)
 	if err != nil {
@@ -206,7 +206,7 @@ func (s *EIPServiceImpl) AssociateAddress(ctx context.Context, input *ec2.Associ
 	}
 
 	allocID := *input.AllocationId
-	key := kvutil.AccountKey(accountID, allocID)
+	key := kvstore.AccountKey(accountID, allocID)
 
 	entry, err := s.eipKV.Get(ctx, key)
 	if err != nil {
@@ -536,7 +536,7 @@ func (s *EIPServiceImpl) findByENI(ctx context.Context, accountID, eniID string)
 		if err := ctx.Err(); err != nil {
 			return nil, err
 		}
-		if k == kvutil.VersionKey || !strings.HasPrefix(k, prefix) {
+		if k == kvstore.VersionKey || !strings.HasPrefix(k, prefix) {
 			continue
 		}
 		entry, err := s.eipKV.Get(ctx, k)
@@ -590,7 +590,7 @@ func (s *EIPServiceImpl) DescribeAddresses(ctx context.Context, input *ec2.Descr
 
 	var addresses []*ec2.Address
 	for _, k := range keys {
-		if k == kvutil.VersionKey {
+		if k == kvstore.VersionKey {
 			continue
 		}
 		if !strings.HasPrefix(k, prefix) {
@@ -699,7 +699,7 @@ func (s *EIPServiceImpl) DescribeAddressesAttribute(ctx context.Context, input *
 			if id == nil {
 				continue
 			}
-			key := kvutil.AccountKey(accountID, *id)
+			key := kvstore.AccountKey(accountID, *id)
 			entry, err := s.eipKV.Get(ctx, key)
 			if err != nil {
 				continue // not found, skip
@@ -722,7 +722,7 @@ func (s *EIPServiceImpl) DescribeAddressesAttribute(ctx context.Context, input *
 			return nil, errors.New(awserrors.ErrorServerInternal)
 		}
 		for _, k := range keys {
-			if k == kvutil.VersionKey {
+			if k == kvstore.VersionKey {
 				continue
 			}
 			if !strings.HasPrefix(k, prefix) {
@@ -819,7 +819,7 @@ func (s *EIPServiceImpl) findByAssociationID(ctx context.Context, accountID, ass
 	}
 
 	for _, k := range keys {
-		if k == kvutil.VersionKey {
+		if k == kvstore.VersionKey {
 			continue
 		}
 		if !strings.HasPrefix(k, prefix) {
@@ -857,7 +857,7 @@ func (s *EIPServiceImpl) AssociatedPublicIPForInstance(ctx context.Context, acco
 		return "", false
 	}
 	for _, k := range keys {
-		if k == kvutil.VersionKey || !strings.HasPrefix(k, prefix) {
+		if k == kvstore.VersionKey || !strings.HasPrefix(k, prefix) {
 			continue
 		}
 		entry, err := s.eipKV.Get(ctx, k)

spinifex/domains/ec2/igw/service_impl.go

@@ -22,7 +22,7 @@ import (
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/paging"
 	awstags "github.com/mulgadc/spinifex/spinifex/foundation/aws/tags"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -93,7 +93,7 @@ func NewIGWServiceImplWithNATS(ctx context.Context, cfg *config.Config, natsConn
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	igwKV, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketIGW, 10)
+	igwKV, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketIGW, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketIGW, err)
 	}
@@ -102,13 +102,13 @@ func NewIGWServiceImplWithNATS(ctx context.Context, cfg *config.Config, natsConn
 	}
 
 	// Get or create VPC KV bucket for cross-resource ownership validation
-	vpcKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
+	vpcKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get VPC KV bucket: %w", err)
 	}
 
 	// Read EIP associations so detach can refuse while addresses are mapped.
-	eipKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2eip.KVBucketEIPs, 10)
+	eipKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2eip.KVBucketEIPs, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get EIP KV bucket: %w", err)
 	}
@@ -142,7 +142,7 @@ func (s *IGWServiceImpl) createIGW(ctx context.Context, input *ec2.CreateInterne
 	if err != nil {
 		return nil, fmt.Errorf("failed to marshal IGW record: %w", err)
 	}
-	if _, err := s.igwKV.Put(ctx, kvutil.AccountKey(accountID, igwID), data); err != nil {
+	if _, err := s.igwKV.Put(ctx, kvstore.AccountKey(accountID, igwID), data); err != nil {
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
 
@@ -160,7 +160,7 @@ func (s *IGWServiceImpl) DeleteInternetGateway(ctx context.Context, input *ec2.D
 	}
 
 	igwID := *input.InternetGatewayId
-	key := kvutil.AccountKey(accountID, igwID)
+	key := kvstore.AccountKey(accountID, igwID)
 
 	entry, err := s.igwKV.Get(ctx, key)
 	if err != nil {
@@ -236,7 +236,7 @@ func (s *IGWServiceImpl) DescribeInternetGateways(ctx context.Context, input *ec
 	foundIDs := make(map[string]bool)
 
 	for _, key := range keys {
-		if key == kvutil.VersionKey {
+		if key == kvstore.VersionKey {
 			continue
 		}
 		if !strings.HasPrefix(key, prefix) {
@@ -326,7 +326,7 @@ func (s *IGWServiceImpl) AttachInternetGateway(ctx context.Context, input *ec2.A
 
 	igwID := *input.InternetGatewayId
 	vpcID := *input.VpcId
-	key := kvutil.AccountKey(accountID, igwID)
+	key := kvstore.AccountKey(accountID, igwID)
 
 	entry, err := s.igwKV.Get(ctx, key)
 	if err != nil {
@@ -347,7 +347,7 @@ func (s *IGWServiceImpl) AttachInternetGateway(ctx context.Context, input *ec2.A
 		slog.ErrorContext(ctx, "VPC KV unavailable, cannot verify VPC ownership")
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
-	if _, err := s.vpcKV.Get(ctx, kvutil.AccountKey(accountID, vpcID)); err != nil {
+	if _, err := s.vpcKV.Get(ctx, kvstore.AccountKey(accountID, vpcID)); err != nil {
 		slog.WarnContext(ctx, "AttachInternetGateway: VPC not found for account", "vpcId", vpcID, "accountID", accountID)
 		return nil, awserrors.IDNotFound(awserrors.ErrorInvalidVpcIDNotFound, "vpc", vpcID)
 	}
@@ -388,7 +388,7 @@ func (s *IGWServiceImpl) DetachInternetGateway(ctx context.Context, input *ec2.D
 
 	igwID := *input.InternetGatewayId
 	vpcID := *input.VpcId
-	key := kvutil.AccountKey(accountID, igwID)
+	key := kvstore.AccountKey(accountID, igwID)
 
 	entry, err := s.igwKV.Get(ctx, key)
 	if err != nil {
@@ -449,7 +449,7 @@ func (s *IGWServiceImpl) DetachInternetGateway(ctx context.Context, input *ec2.D
 // associated into vpcID, whether to an interface or a NAT gateway. The listing
 // is filtered server-side to the account's keys.
 func (s *IGWServiceImpl) vpcHasMappedAddress(ctx context.Context, accountID, vpcID string) (bool, error) {
-	lister, err := s.eipKV.ListKeysFiltered(ctx, kvutil.AccountKey(accountID, "*"))
+	lister, err := s.eipKV.ListKeysFiltered(ctx, kvstore.AccountKey(accountID, "*"))
 	if err != nil {
 		return false, err
 	}
@@ -493,11 +493,11 @@ func (s *IGWServiceImpl) CreateAttachedInternetGateway(ctx context.Context, acco
 	if s.vpcKV == nil {
 		return false, errors.New(awserrors.ErrorServerInternal)
 	}
-	if _, err := s.vpcKV.Get(ctx, kvutil.AccountKey(accountID, vpcID)); err != nil {
+	if _, err := s.vpcKV.Get(ctx, kvstore.AccountKey(accountID, vpcID)); err != nil {
 		return false, awserrors.IDNotFound(awserrors.ErrorInvalidVpcIDNotFound, "vpc", vpcID)
 	}
 
-	key := kvutil.AccountKey(accountID, igwID)
+	key := kvstore.AccountKey(accountID, igwID)
 	data, err := json.Marshal(IGWRecord{
 		InternetGatewayId: igwID,
 		VpcId:             vpcID,
@@ -558,7 +558,7 @@ func (s *IGWServiceImpl) AttachmentIntent(ctx context.Context, accountID, vpcID
 	}
 
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := s.igwKV.Get(ctx, key)
@@ -599,7 +599,7 @@ func (s *IGWServiceImpl) AttachmentIntent(ctx context.Context, accountID, vpcID
 // Writing only on the pending transition keeps a pass from waking the drift loop.
 func MarkAttached(ctx context.Context, kv jetstream.KeyValue, recordKey, vpcID string) error {
 	confirmed := false
-	record, err := kvutil.Update(ctx, kv, recordKey, kvutil.CASConfig{}, func(r *IGWRecord) (bool, error) {
+	record, err := kvstore.Update(ctx, kv, recordKey, kvstore.CASConfig{}, func(r *IGWRecord) (bool, error) {
 		// Reset per attempt: a retried CAS may see a record another writer has
 		// already moved out of pending.
 		confirmed = r.VpcId == vpcID && r.AttachState == AttachStatePending

spinifex/domains/ec2/igw/service_impl_test.go

@@ -13,7 +13,7 @@ import (
 	ec2eip "github.com/mulgadc/spinifex/spinifex/domains/ec2/eip"
 	ec2vpc "github.com/mulgadc/spinifex/spinifex/domains/ec2/vpc"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -55,7 +55,7 @@ func setupTestIGWService(t *testing.T) (*IGWServiceImpl, *nats.Conn) {
 	// Create VPC KV bucket and register test VPCs so fail-closed ownership checks pass
 	vpcEntries := map[string][]byte{}
 	for _, vpcID := range []string{"vpc-test123", "vpc-other", "vpc-lifecycle", "vpc-event-test"} {
-		vpcEntries[kvutil.AccountKey(testAccountID, vpcID)] = []byte(`{"vpc_id":"` + vpcID + `","state":"available"}`)
+		vpcEntries[kvstore.AccountKey(testAccountID, vpcID)] = []byte(`{"vpc_id":"` + vpcID + `","state":"available"}`)
 	}
 	testutil.SeedKV(t, js, ec2vpc.KVBucketVPCs, vpcEntries)
 
@@ -68,7 +68,7 @@ func setupTestIGWService(t *testing.T) (*IGWServiceImpl, *nats.Conn) {
 // gateway come up. Without it an attach stays pending and reports no attachment.
 func confirmAttach(t *testing.T, svc *IGWServiceImpl, igwID, vpcID string) {
 	t.Helper()
-	require.NoError(t, MarkAttached(context.Background(), svc.igwKV, kvutil.AccountKey(testAccountID, igwID), vpcID))
+	require.NoError(t, MarkAttached(context.Background(), svc.igwKV, kvstore.AccountKey(testAccountID, igwID), vpcID))
 }
 
 func createTestIGW(t *testing.T, svc *IGWServiceImpl) string {
@@ -331,7 +331,7 @@ func TestDescribeInternetGateways_LegacyRecordReportsAttachment(t *testing.T) {
 	svc, _ := setupTestIGWService(t)
 	igwID := createTestIGW(t, svc)
 
-	key := kvutil.AccountKey(testAccountID, igwID)
+	key := kvstore.AccountKey(testAccountID, igwID)
 	entry, err := svc.igwKV.Get(context.Background(), key)
 	require.NoError(t, err)
 	var record IGWRecord
@@ -398,7 +398,7 @@ func TestMarkAttached_NoRewriteWhenAlreadyAttached(t *testing.T) {
 	}, testAccountID)
 	require.NoError(t, err)
 
-	key := kvutil.AccountKey(testAccountID, igwID)
+	key := kvstore.AccountKey(testAccountID, igwID)
 	confirmAttach(t, svc, igwID, "vpc-test123")
 	first, err := svc.igwKV.Get(context.Background(), key)
 	require.NoError(t, err)
@@ -605,7 +605,7 @@ func TestDetachInternetGateway_MappedPublicAddress(t *testing.T) {
 	putEIP := func(rec ec2eip.EIPRecord) {
 		data, err := json.Marshal(rec)
 		require.NoError(t, err)
-		_, err = svc.eipKV.Put(context.Background(), kvutil.AccountKey(testAccountID, rec.AllocationId), data)
+		_, err = svc.eipKV.Put(context.Background(), kvstore.AccountKey(testAccountID, rec.AllocationId), data)
 		require.NoError(t, err)
 	}
 	putEIP(ec2eip.EIPRecord{AllocationId: "eipalloc-other", AssociationId: "eipassoc-other", VpcId: "vpc-other", State: "associated"})
@@ -614,7 +614,7 @@ func TestDetachInternetGateway_MappedPublicAddress(t *testing.T) {
 	// Another account's address naming the same VPC ID is not this account's dependency.
 	foreign, err := json.Marshal(ec2eip.EIPRecord{AllocationId: "eipalloc-foreign", AssociationId: "eipassoc-foreign", VpcId: "vpc-test123", State: "associated"})
 	require.NoError(t, err)
-	_, err = svc.eipKV.Put(context.Background(), kvutil.AccountKey("999999999999", "eipalloc-foreign"), foreign)
+	_, err = svc.eipKV.Put(context.Background(), kvstore.AccountKey("999999999999", "eipalloc-foreign"), foreign)
 	require.NoError(t, err)
 
 	detach := func() error {
@@ -828,7 +828,7 @@ func TestAttachInternetGateway_CrossAccountVPCRejected(t *testing.T) {
 	require.NoError(t, err)
 
 	vpcID := "vpc-alpha123"
-	_, err = vpcKV.Put(t.Context(), kvutil.AccountKey(testAccountID, vpcID), []byte(`{"vpc_id":"vpc-alpha123","state":"available"}`))
+	_, err = vpcKV.Put(t.Context(), kvstore.AccountKey(testAccountID, vpcID), []byte(`{"vpc_id":"vpc-alpha123","state":"available"}`))
 	require.NoError(t, err)
 
 	// Refresh service to pick up the VPC KV bucket

spinifex/domains/ec2/instance/provider_efi_volume_test.go

@@ -1,126 +0,0 @@
-package instance
-
-import (
-	"bytes"
-	"context"
-	"testing"
-
-	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/providers/ebs"
-	"github.com/mulgadc/spinifex/spinifex/providers/objectstore"
-	"github.com/mulgadc/spinifex/spinifex/runtime/compute/vm"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-// varsTemplateBytes stands in for a firmware VARS template. The real one is
-// 540,672 bytes, but nothing here depends on the size beyond it being exact.
-func varsTemplateBytes() []byte { return bytes.Repeat([]byte{0xAB}, 4096) }
-
-func providerEFIService(t *testing.T, provider ebsprovider.EBSProvider) *InstanceServiceImpl {
-	t.Helper()
-	return providerRootVolumeService(t, provider, rootVolumeAMILoader(), objectstore.NewMemoryObjectStore())
-}
-
-// TestCreateEFIVolume_Provider_SeedsFromTemplate is the core of the EFI
-// decouple: the variable store is allocated through the provider with the
-// template as seed bytes, so no control-plane engine is built for it.
-func TestCreateEFIVolume_Provider_SeedsFromTemplate(t *testing.T) {
-	provider := ebsprovider.NewMemoryProvider(ebsprovider.Capabilities{VolumeSeeding: true})
-	svc := providerEFIService(t, provider)
-	template := varsTemplateBytes()
-
-	require.NoError(t, svc.createEFIVolumeViaProvider(context.Background(), "vol-efiseed001-efi", template))
-
-	volume, err := provider.GetVolume(context.Background(), ebsprovider.GetVolumeRequest{
-		Versioned: ebsprovider.NewVersioned(), VolumeID: "vol-efiseed001-efi",
-	})
-	require.NoError(t, err)
-	assert.Equal(t, int64(len(template)), volume.CapacityBytes,
-		"the store must be sized byte-exactly to the VARS template")
-
-	seed, ok := provider.SeedData("vol-efiseed001-efi")
-	require.True(t, ok)
-	assert.Equal(t, template, seed, "the template must reach the provider as seed bytes")
-}
-
-// A provider that rounds capacity to a GiB boundary produces a VARS region
-// pflash refuses, which would surface as an unbootable guest rather than a
-// failed launch. The mismatch has to fail the launch instead.
-func TestCreateEFIVolume_Provider_CapacityMismatchFails(t *testing.T) {
-	provider := &roundingProvider{
-		EBSProvider: ebsprovider.NewMemoryProvider(ebsprovider.Capabilities{VolumeSeeding: true}),
-	}
-	svc := providerEFIService(t, provider)
-
-	err := svc.createEFIVolumeViaProvider(context.Background(), "vol-efiround01-efi", varsTemplateBytes())
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
-}
-
-// A relaunch reissues the same create. Reseeding would overwrite the BootOrder
-// the guest wrote on its first boot, so the second call must be a no-op.
-func TestCreateEFIVolume_Provider_RepeatDoesNotReseed(t *testing.T) {
-	provider := ebsprovider.NewMemoryProvider(ebsprovider.Capabilities{VolumeSeeding: true})
-	svc := providerEFIService(t, provider)
-	ctx := context.Background()
-	template := varsTemplateBytes()
-
-	require.NoError(t, svc.createEFIVolumeViaProvider(ctx, "vol-efirepeat1-efi", template))
-	require.NoError(t, svc.createEFIVolumeViaProvider(ctx, "vol-efirepeat1-efi", template))
-
-	seed, ok := provider.SeedData("vol-efirepeat1-efi")
-	require.True(t, ok)
-	assert.Equal(t, template, seed)
-}
-
-// A seed above the wire bound must fail the launch with an error naming the
-// limit, not reach NATS and fail as an oversized publish.
-func TestCreateEFIVolume_Provider_OversizedTemplateFails(t *testing.T) {
-	provider := ebsprovider.NewMemoryProvider(ebsprovider.Capabilities{VolumeSeeding: true})
-	svc := providerEFIService(t, provider)
-
-	err := svc.createEFIVolumeViaProvider(context.Background(), "vol-efitoobig1-efi", make([]byte, ebsprovider.MaxSeedBytes+1))
-	require.Error(t, err)
-	assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
-}
-
-// TestPrepareEFIVolume_Provider_RegistersEBSRequest covers the whole function
-// against this host's real firmware, including the name the VM launches with.
-func TestPrepareEFIVolume_Provider_RegistersEBSRequest(t *testing.T) {
-	if _, _, _, err := vm.FirmwarePaths("x86_64"); err != nil {
-		t.Skipf("no UEFI firmware on this host: %v", err)
-	}
-
-	provider := ebsprovider.NewMemoryProvider(ebsprovider.Capabilities{VolumeSeeding: true})
-	svc := providerEFIService(t, provider)
-	instance := &vm.VM{}
-
-	require.NoError(t, svc.prepareEFIVolume(context.Background(), "vol-efiprepare001", instance, "x86_64"))
-
-	instance.EBSRequests.Mu.Lock()
-	defer instance.EBSRequests.Mu.Unlock()
-	require.Len(t, instance.EBSRequests.Requests, 1)
-	assert.Equal(t, "vol-efiprepare001-efi", instance.EBSRequests.Requests[0].Name)
-	assert.True(t, instance.EBSRequests.Requests[0].EFI)
-	assert.False(t, instance.EBSRequests.Requests[0].Boot)
-
-	seed, ok := provider.SeedData("vol-efiprepare001-efi")
-	require.True(t, ok)
-	assert.NotEmpty(t, seed, "the store must be seeded from this host's VARS template")
-}
-
-// roundingProvider reports a GiB-rounded capacity the way a provider with no
-// sub-GiB volume granularity would.
-type roundingProvider struct {
-	ebsprovider.EBSProvider
-}
-
-func (p *roundingProvider) CreateVolume(ctx context.Context, req ebsprovider.CreateVolumeRequest) (*ebsprovider.Volume, error) {
-	volume, err := p.EBSProvider.CreateVolume(ctx, req)
-	if err != nil {
-		return nil, err
-	}
-	volume.CapacityBytes = 1 << 30
-	return volume, nil
-}

spinifex/domains/ec2/instance/service_impl.go

@@ -10,7 +10,6 @@ import (
 	"github.com/mulgadc/spinifex/spinifex/domains/ec2/tagmirror"
 	"log/slog"
 	"maps"
-	"os"
 	"slices"
 	"sort"
 	"strconv"
@@ -1265,6 +1264,10 @@ func (s *InstanceServiceImpl) StartInstance(ctx context.Context, instance *vm.VM
 				slog.WarnContext(ctx, "StartInstance: vmMgr.Start failed", "instanceId", command.ID, "err", err)
 				return refusal
 			}
+			if errors.Is(err, vm.ErrUEFIClaimHeld) {
+				slog.WarnContext(ctx, "StartInstance: vmMgr.Start failed", "instanceId", command.ID, "err", err)
+				return errInsufficientCapacity
+			}
 			slog.ErrorContext(ctx, "StartInstance: vmMgr.Start failed", "instanceId", command.ID, "err", err)
 			return errors.New(awserrors.ErrorServerInternal)
 		}
@@ -1612,18 +1615,14 @@ func (s *InstanceServiceImpl) GenerateVolumes(ctx context.Context, input *ec2.Ru
 		return nil, err
 	}
 
-	// Step 2: Create EFI variable store (only when the AMI is UEFI; BIOS
-	// guests must not allocate an orphan VARS volume).
-	if instance.BootMode == "uefi" || instance.BootMode == "uefi-preferred" {
-		arch := instanceArchitecture(s.instanceTypes[*input.InstanceType])
-		err = s.prepareEFIVolume(ctx, imageId, instance, arch)
-		if err != nil {
-			return nil, err
-		}
+	// Step 2: Seed the UEFI variable store, which is not a volume. A failure
+	// terminates the instance, and terminate deletes whatever store it left.
+	if err := s.vmMgr.SeedUEFIVarStore(ctx, instance); err != nil {
+		slog.ErrorContext(ctx, "Failed to seed the UEFI variable store", "instanceId", instance.ID, "err", err)
+		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
 
-	// The EFI store is internal, so it is not reported; every volume the
-	// request named is.
+	// Every volume the request named is reported.
 	volumeInfos := []VolumeInfo{
 		{
 			VolumeId:            imageId,
@@ -1885,81 +1884,6 @@ func rootVolumeIOPS(requested int) int {
 	return requested
 }
 
-// prepareEFIVolume creates the per-VM EFI variable store, sized exactly to the
-// firmware VARS template (pflash requires byte-exact size) and seeded from it.
-// arch is "x86_64" | "arm64".
-func (s *InstanceServiceImpl) prepareEFIVolume(ctx context.Context, volumeID string, instance *vm.VM, arch string) error {
-	codePath, varsTemplate, varsSize, err := vm.FirmwarePaths(arch)
-	if err != nil {
-		slog.ErrorContext(ctx, "UEFI firmware not installed on this host", "arch", arch, "err", err)
-		return errors.New(awserrors.ErrorServerInternal)
-	}
-	template, err := os.ReadFile(varsTemplate)
-	if err != nil {
-		slog.ErrorContext(ctx, "Failed to read VARS template", "path", varsTemplate, "err", err)
-		return errors.New(awserrors.ErrorServerInternal)
-	}
-	if int64(len(template)) != varsSize {
-		slog.ErrorContext(ctx, "VARS template size mismatch between stat and read", "path", varsTemplate, "statSize", varsSize, "readSize", len(template))
-		return errors.New(awserrors.ErrorServerInternal)
-	}
-	slog.InfoContext(ctx, "Preparing EFI variable store", "arch", arch, "firmwarePath", codePath, "varsTemplate", varsTemplate, "size", varsSize)
-
-	efiVolumeName := fmt.Sprintf("%s-efi", volumeID)
-	if s.ebsProvider == nil {
-		slog.ErrorContext(ctx, "no EBS provider configured", "op", "prepareEFIVolume")
-		return errors.New(awserrors.ErrorServerInternal)
-	}
-	if err := s.createEFIVolumeViaProvider(ctx, efiVolumeName, template); err != nil {
-		return err
-	}
-
-	instance.EBSRequests.Mu.Lock()
-	instance.EBSRequests.Requests = append(instance.EBSRequests.Requests, viperblocklegacyv1.EBSRequest{
-		Name: efiVolumeName,
-		Boot: false,
-		EFI:  true,
-	})
-	instance.EBSRequests.Mu.Unlock()
-
-	return nil
-}
-
-// createEFIVolumeViaProvider allocates the EFI variable store through the
-// provider, seeding it with this node's firmware VARS template. The bytes
-// travel with the request so the region matches the firmware that will run.
-func (s *InstanceServiceImpl) createEFIVolumeViaProvider(ctx context.Context, efiVolumeName string, template []byte) error {
-	slog.InfoContext(ctx, "Creating EFI variable store via EBS provider", "name", efiVolumeName, "seedBytes", len(template))
-
-	// The store is internal to the launch, never attachable and filtered out of
-	// DescribeVolumes, so it gets no ebsmetadata document the way a root volume
-	// does. Its lifecycle is the root volume's DeleteVolume.
-	created, err := s.ebsProvider.CreateVolume(ctx, ebsprovider.CreateVolumeRequest{
-		Versioned:        ebsprovider.NewVersioned(),
-		VolumeID:         efiVolumeName,
-		CapacityRange:    ebsprovider.CapacityRange{RequiredBytes: int64(len(template))},
-		AvailabilityZone: s.config.AZ,
-		SeedData:         template,
-	})
-	if err != nil {
-		slog.ErrorContext(ctx, "Provider EFI volume creation failed", "name", efiVolumeName, "err", err)
-		return errors.New(awserrors.ErrorServerInternal)
-	}
-	if created == nil {
-		slog.ErrorContext(ctx, "Provider returned no volume for the EFI variable store", "name", efiVolumeName)
-		return errors.New(awserrors.ErrorServerInternal)
-	}
-
-	// pflash rejects a VARS region that is not byte-exact, so a provider that
-	// rounded the capacity up to a GiB boundary yields a guest that cannot boot.
-	if created.CapacityBytes != int64(len(template)) {
-		slog.ErrorContext(ctx, "Provider EFI volume capacity does not match the VARS template",
-			"name", efiVolumeName, "wantBytes", len(template), "gotBytes", created.CapacityBytes)
-		return errors.New(awserrors.ErrorServerInternal)
-	}
-	return nil
-}
-
 // checkImageArchitecture refuses an AMI whose architecture the instance type
 // does not support, as AWS does. An unrecorded architecture on either side is
 // not a mismatch.
@@ -2760,7 +2684,7 @@ func (s *InstanceServiceImpl) capacityRefusal(ctx context.Context, instance *vm.
 func (s *InstanceServiceImpl) launchClaimedInstance(ctx context.Context, instance *vm.VM, undo func()) error {
 	// Before anything is stamped or claimed, so the record handed back is the
 	// one that was claimed.
-	if err := vm.CheckUEFIVarStore(instance); err != nil {
+	if err := s.vmMgr.CheckUEFIVarStore(ctx, instance); err != nil {
 		slog.ErrorContext(ctx, "launchClaimedInstance: refusing to launch a UEFI instance without exactly one variable store",
 			"instanceId", instance.ID, "err", err)
 		undo()
@@ -2893,6 +2817,10 @@ func (s *InstanceServiceImpl) launchClaimedInstance(ctx context.Context, instanc
 		if refusal := volumeHeldElsewhereError(err); refusal != nil {
 			return refusal
 		}
+		// A live node holds the variable store, so another node may take it.
+		if errors.Is(err, vm.ErrUEFIClaimHeld) {
+			return s.capacityRefusal(ctx, instance)
+		}
 		return errors.New(awserrors.ErrorServerInternal)
 	}
 

spinifex/domains/ec2/instance/teardown_outstanding.go

@@ -1,33 +1,12 @@
 package instance
 
-import "context"
-
 // OutstandingInstance is an instance of an account that is not yet fully torn
 // down, with the state its record carries.
 type OutstandingInstance struct {
-	InstanceID string `json:"instance_id"`
+	InstanceID string
 	// State is the record's lifecycle state, such as running or terminated.
-	State string `json:"state"`
+	State string
 	// Teardown is the per-dependent progress (done, pending or failed) the
 	// record carries. It is empty when no dependent has been recorded.
-	Teardown map[string]string `json:"teardown,omitempty"`
-}
-
-// TeardownOutstandingInput is the empty request of the outstanding-teardown
-// read; the account comes from the message header.
-type TeardownOutstandingInput struct{}
-
-// TeardownOutstandingOutput carries every instance not yet fully torn down.
-type TeardownOutstandingOutput struct {
-	Instances []OutstandingInstance `json:"instances"`
-}
-
-// AccountTeardownOutstandingRPC serves AccountTeardownOutstanding over the
-// message bus, so a process without the instance record store can read it.
-func (s *InstanceServiceImpl) AccountTeardownOutstandingRPC(_ context.Context, _ *TeardownOutstandingInput, accountID string) (*TeardownOutstandingOutput, error) {
-	outstanding, err := s.AccountTeardownOutstanding(accountID)
-	if err != nil {
-		return nil, err
-	}
-	return &TeardownOutstandingOutput{Instances: outstanding}, nil
+	Teardown map[string]string
 }

spinifex/domains/ec2/instance/uefi_seed_test.go

@@ -0,0 +1,122 @@
+package instance
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/aws/aws-sdk-go/aws"
+	"github.com/aws/aws-sdk-go/service/ec2"
+	testutil "github.com/mulgadc/spinifex/internal/testkit"
+	"github.com/mulgadc/spinifex/spinifex/domains/ec2/instance/uefistore"
+	awserrors "github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
+	ebsprovider "github.com/mulgadc/spinifex/spinifex/providers/ebs"
+	"github.com/mulgadc/spinifex/spinifex/providers/objectstore"
+	"github.com/mulgadc/spinifex/spinifex/runtime/compute/vm"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+const testSeedInstance = "i-0a1b2c3d4e5f60718"
+
+type seedInstanceTypes map[string]vm.InstanceTypeSpec
+
+func (r seedInstanceTypes) Resolve(name string) (vm.InstanceTypeSpec, bool) {
+	spec, ok := r[name]
+	return spec, ok
+}
+
+// uefiSeedingService is an instance service whose VM manager seeds stores into
+// a real JetStream's store records and claims and an in-memory object store,
+// under a stand-in x86_64 firmware.
+func uefiSeedingService(t *testing.T) (*InstanceServiceImpl, *ebsprovider.MemoryProvider, *uefistore.Store, *objectstore.MemoryObjectStore) {
+	t.Helper()
+	fwDir := t.TempDir()
+	code := filepath.Join(fwDir, "OVMF_CODE.fd")
+	vars := filepath.Join(fwDir, "OVMF_VARS.fd")
+	require.NoError(t, os.WriteFile(code, []byte("code"), 0o600))
+	require.NoError(t, os.WriteFile(vars, make([]byte, 4096), 0o600))
+	orig, had := vm.FirmwarePathCandidates["x86_64"]
+	vm.FirmwarePathCandidates["x86_64"] = []vm.FirmwareCandidate{{Code: code, VarsTemplate: vars}}
+	t.Cleanup(func() {
+		if had {
+			vm.FirmwarePathCandidates["x86_64"] = orig
+		} else {
+			delete(vm.FirmwarePathCandidates, "x86_64")
+		}
+	})
+
+	_, _, js := testutil.StartTestJetStream(t)
+	records := uefistore.NewStore(js)
+	require.NoError(t, records.Open(t.Context()))
+
+	provider := ebsprovider.NewMemoryProvider(ebsprovider.Capabilities{CrashConsistentSnapshot: true, VolumeEnumeration: true})
+	seedProviderSnapshot(t, provider, "vol-origin", "snap-source")
+	objects := objectstore.NewMemoryObjectStore()
+	svc := providerRootVolumeService(t, provider, rootVolumeAMILoader(), objects)
+	svc.vmMgr = vm.NewManagerWithDeps(vm.Deps{
+		NodeID:        "node-a",
+		InstanceTypes: seedInstanceTypes{"t3.nano": {Architecture: "x86_64", VCPUs: 1, MemoryMiB: 512}},
+		UEFIVarStore: vm.UEFIVarStoreDeps{
+			Objects: objects, Bucket: testRootBucket, DataDir: t.TempDir(), Records: records,
+			Claims: uefistore.NewClaims(kvstore.OpenSyncBucket(js), "node-a"),
+		},
+	})
+	return svc, provider, records, objects
+}
+
+func uefiLaunch() (*ec2.RunInstancesInput, *vm.VM) {
+	input := &ec2.RunInstancesInput{ImageId: aws.String("ami-1"), InstanceType: aws.String("t3.nano")}
+	return input, &vm.VM{ID: testSeedInstance, AccountID: testRootAccount, InstanceType: "t3.nano", BootMode: "uefi"}
+}
+
+// A UEFI launch seeds its store as an object and a store record, and creates
+// no <root>-efi volume and no EFI request.
+func TestGenerateVolumes_UEFILaunchSeedsAStoreAndCreatesNoEFIVolume(t *testing.T) {
+	svc, provider, records, objects := uefiSeedingService(t)
+	input, instance := uefiLaunch()
+
+	infos, err := svc.GenerateVolumes(t.Context(), input, instance)
+	require.NoError(t, err)
+
+	require.Len(t, infos, 1, "only the root volume is reported")
+	instance.EBSRequests.Mu.Lock()
+	for _, req := range instance.EBSRequests.Requests {
+		assert.False(t, req.EFI, "no EFI request is appended")
+		assert.False(t, strings.HasSuffix(req.Name, "-efi"), req.Name)
+	}
+	instance.EBSRequests.Mu.Unlock()
+	listed, err := provider.ListVolumes(t.Context(), ebsprovider.ListVolumesRequest{Versioned: ebsprovider.NewVersioned()})
+	require.NoError(t, err)
+	for _, v := range listed.Volumes {
+		assert.False(t, strings.HasSuffix(v.ID, "-efi"), "no EFI volume is created: %s", v.ID)
+	}
+
+	rec, _, err := records.Get(t.Context(), testSeedInstance)
+	require.NoError(t, err)
+	assert.Equal(t, uefistore.StateLive, rec.State)
+	require.NotNil(t, rec.Pointer)
+	keys, err := uefistore.ListObjects(t.Context(), objects, testRootBucket, testSeedInstance)
+	require.NoError(t, err)
+	assert.Equal(t, []string{rec.Pointer.Key}, keys, "the pointer names the one template object")
+}
+
+// A seed that cannot be written fails the launch as InternalError, and no EFI
+// volume is created in its place.
+func TestGenerateVolumes_FailedSeedIsInternalError(t *testing.T) {
+	svc, provider, _, _ := uefiSeedingService(t)
+	svc.vmMgr = vm.NewManager()
+	input, instance := uefiLaunch()
+
+	_, err := svc.GenerateVolumes(t.Context(), input, instance)
+
+	require.Error(t, err)
+	assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
+	listed, err := provider.ListVolumes(t.Context(), ebsprovider.ListVolumesRequest{Versioned: ebsprovider.NewVersioned()})
+	require.NoError(t, err)
+	for _, v := range listed.Volumes {
+		assert.False(t, strings.HasSuffix(v.ID, "-efi"), v.ID)
+	}
+}

spinifex/domains/ec2/instance/uefistore/cleanup.go

@@ -10,7 +10,6 @@ import (
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/s3"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
 	"github.com/mulgadc/spinifex/spinifex/providers/objectstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
@@ -26,9 +25,9 @@ func (s *Store) Exists(ctx context.Context, instanceID string) (bool, error) {
 	return s.kv.Exists(ctx, instanceID)
 }
 
-// MarkDeleting records that instanceID's store is to be deleted, and drops any
-// write-back still owed. It reports false with no error when the instance has
-// no store record. Only a terminate or an operator's forget may call it.
+// MarkDeleting records that instanceID's store is to be deleted. It reports
+// false with no error when the instance has no store record. Only a terminate
+// or an operator's forget may call it.
 func (s *Store) MarkDeleting(ctx context.Context, instanceID string) (bool, error) {
 	for range markAttempts {
 		rec, rev, err := s.Get(ctx, instanceID)
@@ -42,7 +41,6 @@ func (s *Store) MarkDeleting(ctx context.Context, instanceID string) (bool, erro
 			return true, nil
 		}
 		rec.State = StateDeleting
-		rec.WriteBack = nil
 		_, err = s.CompareAndSet(ctx, rec, rev)
 		if err == nil {
 			return true, nil
@@ -72,7 +70,7 @@ func (s *Store) InstanceIDs(ctx context.Context) ([]string, error) {
 
 	var ids []string
 	for key := range lister.Keys() {
-		if key != kvutil.VersionKey {
+		if key != kvstore.VersionKey {
 			ids = append(ids, key)
 		}
 	}

spinifex/domains/ec2/instance/uefistore/cleanup_test.go

@@ -37,7 +37,7 @@ func keysUnder(t *testing.T, objects objectstore.ObjectStore, prefix string) []s
 }
 
 // Deletion is final: a writer holding the revision of a deleting record, such
-// as a write-back retry, must not bring the store back to live.
+// as a late pointer move, must not bring the store back to live.
 func TestCompareAndSet_RefusesDeletingToLive(t *testing.T) {
 	s, _ := openStore(t)
 	rev, err := s.Create(t.Context(), liveRecord())
@@ -47,8 +47,7 @@ func TestCompareAndSet_RefusesDeletingToLive(t *testing.T) {
 	rev, err = s.CompareAndSet(t.Context(), deleting, rev)
 	require.NoError(t, err)
 
-	revived := liveRecord()
-	revived.WriteBack = writeBack()
+	revived := pointedRecord()
 	_, err = s.CompareAndSet(t.Context(), revived, rev)
 	require.ErrorIs(t, err, uefistore.ErrDeleting)
 
@@ -58,24 +57,9 @@ func TestCompareAndSet_RefusesDeletingToLive(t *testing.T) {
 	assert.Equal(t, uefistore.StateDeleting, got.State)
 }
 
-// A write-back owed on a deleting store would put an object behind the delete.
-func TestCompareAndSet_RefusesDeletingWithAWriteBack(t *testing.T) {
+func TestMarkDeleting_MovesLiveToDeletingAndKeepsWhatItNames(t *testing.T) {
 	s, _ := openStore(t)
-	rev, err := s.Create(t.Context(), liveRecord())
-	require.NoError(t, err)
-	rec := liveRecord()
-	rec.State = uefistore.StateDeleting
-	rec.WriteBack = writeBack()
-
-	_, err = s.CompareAndSet(t.Context(), rec, rev)
-	require.ErrorIs(t, err, uefistore.ErrInvalidRecord)
-}
-
-func TestMarkDeleting_MovesLiveToDeletingAndDropsTheWriteBack(t *testing.T) {
-	s, _ := openStore(t)
-	rec := liveRecord()
-	rec.WriteBack = writeBack()
-	_, err := s.Create(t.Context(), rec)
+	_, err := s.Create(t.Context(), pointedRecord())
 	require.NoError(t, err)
 
 	has, err := s.MarkDeleting(t.Context(), instanceID)
@@ -85,8 +69,8 @@ func TestMarkDeleting_MovesLiveToDeletingAndDropsTheWriteBack(t *testing.T) {
 	got, rev, err := s.Get(t.Context(), instanceID)
 	require.NoError(t, err)
 	assert.Equal(t, uefistore.StateDeleting, got.State)
-	assert.Nil(t, got.WriteBack)
 	assert.Equal(t, "111122223333", got.AccountID)
+	assert.Equal(t, pointer("b"), got.Pointer, "the pointer stays readable until the prefix is gone")
 
 	has, err = s.MarkDeleting(t.Context(), instanceID)
 	require.NoError(t, err)

spinifex/domains/ec2/instance/uefistore/fence.go

@@ -0,0 +1,93 @@
+package uefistore
+
+import (
+	"context"
+	"errors"
+	"fmt"
+
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvsync"
+)
+
+// ClaimPrefix is the kvsync prefix of every store's claim. An instance's claim
+// is ClaimPrefix+instanceID, and names the node that may hold variable changes
+// no object has yet.
+const ClaimPrefix = "uefi."
+
+// NewClaims returns node's group of store claims over store. The daemon builds
+// it once per process, as kvsync requires of every prefix.
+func NewClaims(store kvsync.KVStore, node string) *kvsync.ClaimGroup {
+	return kvsync.NewClaimGroup(store, ClaimPrefix, node)
+}
+
+// ErrFenced is a stamp or pointer move refused because the record carries a
+// claim generation other than the caller's: a later acquisition has stamped
+// it, so the caller no longer owns the store and must change nothing.
+var ErrFenced = errors.New("uefistore: store record is fenced to another claim generation")
+
+// fenceAttempts bounds the read-then-write loops of Stamp and MovePointer. A
+// lost race is re-read, so a concurrent stamp is seen and refused, not retried.
+const fenceAttempts = 5
+
+// Stamp raises instanceID's record to generation, the claim token the caller
+// has just acquired or adopted, and returns the record it stamped, whose
+// pointer the caller boots from. One above generation is ErrFenced, and one
+// deleting is ErrDeleting. A record already at generation is still written by
+// CAS, because a follower can answer the read with an older pointer.
+func (s *Store) Stamp(ctx context.Context, instanceID string, generation uint64) (*UEFIStoreRecord, error) {
+	if generation == 0 {
+		return nil, fmt.Errorf("uefistore: stamp %s: generation 0 is no claim", instanceID)
+	}
+	for range fenceAttempts {
+		rec, rev, err := s.Get(ctx, instanceID)
+		if err != nil {
+			return nil, err
+		}
+		switch {
+		case rec.State == StateDeleting:
+			return nil, fmt.Errorf("%w: %s", ErrDeleting, instanceID)
+		case rec.Generation > generation:
+			return nil, fmt.Errorf("%w: %s is at generation %d, above the claim's %d", ErrFenced, instanceID, rec.Generation, generation)
+		}
+		rec.Generation = generation
+		_, err = s.CompareAndSet(ctx, rec, rev)
+		if err == nil {
+			return rec, nil
+		}
+		if !errors.Is(err, kvstore.ErrConflict) {
+			return nil, err
+		}
+	}
+	return nil, fmt.Errorf("uefistore: stamp %s: %w after %d attempts", instanceID, kvstore.ErrConflict, fenceAttempts)
+}
+
+// MovePointer points instanceID's record at ref only while the record is at
+// generation. A record at any other generation is ErrFenced and is left
+// unchanged. The leader answers the read, because a caller drops its copy on
+// ErrFenced or ErrDeleting, and the CAS keeps a lost race from succeeding.
+func (s *Store) MovePointer(ctx context.Context, instanceID string, generation uint64, ref UEFIVarStoreRef) error {
+	if err := ref.validate(instanceID); err != nil {
+		return fmt.Errorf("%w: move %s: %w", ErrInvalidRecord, instanceID, err)
+	}
+	for range fenceAttempts {
+		rec, rev, err := s.LeaderGet(ctx, instanceID)
+		if err != nil {
+			return err
+		}
+		switch {
+		case rec.State == StateDeleting:
+			return fmt.Errorf("%w: %s", ErrDeleting, instanceID)
+		case rec.Generation != generation:
+			return fmt.Errorf("%w: %s is at generation %d, and the writer holds %d", ErrFenced, instanceID, rec.Generation, generation)
+		}
+		rec.Pointer = &ref
+		_, err = s.CompareAndSet(ctx, rec, rev)
+		if err == nil {
+			return nil
+		}
+		if !errors.Is(err, kvstore.ErrConflict) {
+			return err
+		}
+	}
+	return fmt.Errorf("uefistore: move %s: %w after %d attempts", instanceID, kvstore.ErrConflict, fenceAttempts)
+}

spinifex/domains/ec2/instance/uefistore/fence_test.go

@@ -0,0 +1,213 @@
+package uefistore_test
+
+import (
+	"testing"
+
+	testutil "github.com/mulgadc/spinifex/internal/testkit"
+	"github.com/mulgadc/spinifex/spinifex/domains/ec2/instance/uefistore"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvsync"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func revisionOf(t *testing.T, s *uefistore.Store) uint64 {
+	t.Helper()
+	_, rev, err := s.Get(t.Context(), instanceID)
+	require.NoError(t, err)
+	return rev
+}
+
+func TestStamp_RaisesTheGenerationAndReturnsThePointerToBootFrom(t *testing.T) {
+	s, _ := openStore(t)
+	_, err := s.Create(t.Context(), pointedRecord())
+	require.NoError(t, err)
+
+	rec, err := s.Stamp(t.Context(), instanceID, 9)
+	require.NoError(t, err)
+	assert.Equal(t, uint64(9), rec.Generation)
+	assert.Equal(t, pointer("b"), rec.Pointer)
+
+	got, _, err := s.Get(t.Context(), instanceID)
+	require.NoError(t, err)
+	assert.Equal(t, rec, got, "the returned record is the one stored")
+}
+
+// A daemon restart adopts its claim at the same generation. The stamp still
+// writes by CAS, so a pointer read from a lagging follower fails the write and
+// is read again rather than booted from.
+func TestStamp_AtTheRecordsGenerationConfirmsTheReadByCAS(t *testing.T) {
+	s, _ := openStore(t)
+	_, err := s.Create(t.Context(), pointedRecord())
+	require.NoError(t, err)
+	before := revisionOf(t, s)
+
+	rec, err := s.Stamp(t.Context(), instanceID, 7)
+	require.NoError(t, err)
+	assert.Equal(t, pointer("b"), rec.Pointer)
+	assert.Equal(t, uint64(7), rec.Generation)
+	assert.Greater(t, revisionOf(t, s), before, "the read must be confirmed by a revision-checked write")
+}
+
+func TestStamp_RefusesAndChangesNothing(t *testing.T) {
+	tests := []struct {
+		name       string
+		deleting   bool
+		generation uint64
+		want       error
+	}{
+		{name: "a later acquisition stamped it", generation: 6, want: uefistore.ErrFenced},
+		{name: "the store is being deleted", deleting: true, generation: 8, want: uefistore.ErrDeleting},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			s, _ := openStore(t)
+			rev, err := s.Create(t.Context(), pointedRecord())
+			require.NoError(t, err)
+			if tt.deleting {
+				_, err = s.MarkDeleting(t.Context(), instanceID)
+				require.NoError(t, err)
+				rev = revisionOf(t, s)
+			}
+
+			_, err = s.Stamp(t.Context(), instanceID, tt.generation)
+			require.ErrorIs(t, err, tt.want)
+			assert.Equal(t, rev, revisionOf(t, s))
+		})
+	}
+}
+
+func TestStamp_MissingRecordIsNotFound(t *testing.T) {
+	s, _ := openStore(t)
+	_, err := s.Stamp(t.Context(), instanceID, 1)
+	require.ErrorIs(t, err, kvstore.ErrNotFound)
+}
+
+// The record keeps only the current version, so each move replaces the
+// pointer and names nothing it replaced.
+func TestMovePointer_ReplacesThePointer(t *testing.T) {
+	s, _ := openStore(t)
+	_, err := s.Create(t.Context(), pointedRecord())
+	require.NoError(t, err)
+
+	for _, name := range []string{"c", "d"} {
+		require.NoError(t, s.MovePointer(t.Context(), instanceID, 7, *pointer(name)))
+	}
+
+	got, _, err := s.Get(t.Context(), instanceID)
+	require.NoError(t, err)
+	assert.Equal(t, pointer("d"), got.Pointer)
+	assert.Equal(t, uint64(7), got.Generation)
+}
+
+// A move to the key the record already names is still a CAS at the revision
+// read, so a fenced writer that read a lagging follower cannot see success.
+func TestMovePointer_ToTheCurrentKeyConfirmsTheReadByCAS(t *testing.T) {
+	s, _ := openStore(t)
+	_, err := s.Create(t.Context(), pointedRecord())
+	require.NoError(t, err)
+	before := revisionOf(t, s)
+
+	require.NoError(t, s.MovePointer(t.Context(), instanceID, 7, *pointer("b")))
+	assert.Greater(t, revisionOf(t, s), before)
+}
+
+func TestMovePointer_RefusesAndChangesNothing(t *testing.T) {
+	tests := []struct {
+		name       string
+		deleting   bool
+		generation uint64
+		ref        *uefistore.UEFIVarStoreRef
+		want       error
+	}{
+		{name: "an earlier generation", generation: 6, ref: pointer("c"), want: uefistore.ErrFenced},
+		{name: "a later generation than the stamp", generation: 8, ref: pointer("c"), want: uefistore.ErrFenced},
+		{name: "a deleting store", deleting: true, generation: 7, ref: pointer("c"), want: uefistore.ErrDeleting},
+		{name: "a key under another instance", generation: 7,
+			ref:  &uefistore.UEFIVarStoreRef{Key: uefistore.KeyPrefix("i-0fedcba9876543210") + "c", Firmware: pointer("c").Firmware, WriterNode: "n"},
+			want: uefistore.ErrInvalidRecord},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			s, _ := openStore(t)
+			_, err := s.Create(t.Context(), pointedRecord())
+			require.NoError(t, err)
+			if tt.deleting {
+				_, err = s.MarkDeleting(t.Context(), instanceID)
+				require.NoError(t, err)
+			}
+			rev := revisionOf(t, s)
+
+			err = s.MovePointer(t.Context(), instanceID, tt.generation, *tt.ref)
+			require.ErrorIs(t, err, tt.want)
+			got, gotRev, err := s.Get(t.Context(), instanceID)
+			require.NoError(t, err)
+			assert.Equal(t, rev, gotRev)
+			assert.Equal(t, pointer("b"), got.Pointer)
+		})
+	}
+}
+
+// claimedStore is a store record and a claim group per node over one
+// JetStream, as two daemons see them.
+type claimedStore struct {
+	records *uefistore.Store
+	a, b    *kvsync.Claim
+}
+
+func newClaimedStore(t *testing.T) *claimedStore {
+	t.Helper()
+	_, _, js := testutil.StartTestJetStream(t)
+	records := uefistore.NewStore(js)
+	require.NoError(t, records.Open(t.Context()))
+	_, err := records.Create(t.Context(), liveRecord())
+	require.NoError(t, err)
+	return &claimedStore{
+		records: records,
+		a:       uefistore.NewClaims(kvstore.OpenSyncBucket(js), "node-a").Get(instanceID),
+		b:       uefistore.NewClaims(kvstore.OpenSyncBucket(js), "node-b").Get(instanceID),
+	}
+}
+
+// A node partitioned long enough to be taken over is fenced by the new
+// owner's stamp: its pointer move fails and the record keeps what it had.
+func TestFence_TheOldHoldersMoveFailsAfterTheNewOwnersStamp(t *testing.T) {
+	c := newClaimedStore(t)
+	first, err := c.a.TryLock(t.Context())
+	require.NoError(t, err)
+	_, err = c.records.Stamp(t.Context(), instanceID, first.Generation)
+	require.NoError(t, err)
+	require.NoError(t, c.records.MovePointer(t.Context(), instanceID, first.Generation, *pointer("b")))
+
+	second, err := c.b.TakeOver(t.Context(), first)
+	require.NoError(t, err)
+	_, err = c.records.Stamp(t.Context(), instanceID, second.Generation)
+	require.NoError(t, err)
+	rev := revisionOf(t, c.records)
+
+	err = c.records.MovePointer(t.Context(), instanceID, first.Generation, *pointer("late"))
+	require.ErrorIs(t, err, uefistore.ErrFenced)
+	got, gotRev, err := c.records.Get(t.Context(), instanceID)
+	require.NoError(t, err)
+	assert.Equal(t, rev, gotRev)
+	assert.Equal(t, pointer("b"), got.Pointer)
+	assert.Equal(t, second.Generation, got.Generation)
+}
+
+// The new owner reads the pointer after it stamps, so a move the old holder
+// landed between the takeover and the stamp is the one it boots from.
+func TestFence_AMoveBetweenTakeoverAndStampIsThePointerTheNewOwnerBoots(t *testing.T) {
+	c := newClaimedStore(t)
+	first, err := c.a.TryLock(t.Context())
+	require.NoError(t, err)
+	_, err = c.records.Stamp(t.Context(), instanceID, first.Generation)
+	require.NoError(t, err)
+
+	second, err := c.b.TakeOver(t.Context(), first)
+	require.NoError(t, err)
+	require.NoError(t, c.records.MovePointer(t.Context(), instanceID, first.Generation, *pointer("landed")))
+
+	rec, err := c.records.Stamp(t.Context(), instanceID, second.Generation)
+	require.NoError(t, err)
+	assert.Equal(t, keyOf("landed"), rec.Pointer.Key)
+}

spinifex/domains/ec2/instance/uefistore/object.go

@@ -0,0 +1,121 @@
+package uefistore
+
+import (
+	"bytes"
+	"compress/gzip"
+	"context"
+	"crypto/rand"
+	"errors"
+	"fmt"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/aws/aws-sdk-go/aws"
+	"github.com/aws/aws-sdk-go/service/s3"
+	"github.com/mulgadc/spinifex/spinifex/providers/objectstore"
+)
+
+const (
+	// SourceTemplate is a store seeded from the launching node's VARS template.
+	SourceTemplate = "template"
+	// SourceWriteBack is a store written back from a node's live copy.
+	SourceWriteBack = "write-back"
+
+	// PutTimeout bounds one put of a store object. The object is a fixed-size
+	// region, so a put that takes longer is a wedged object store.
+	PutTimeout = 10 * time.Second
+)
+
+// ErrInvalidObject is a store object whose key, region or metadata cannot be
+// written as the owner's store.
+var ErrInvalidObject = errors.New("uefistore: invalid store object")
+
+// ObjectMeta is what a store object records about itself, as x-amz-meta-
+// entries, so its owner, writer and firmware are known from the object alone.
+type ObjectMeta struct {
+	OwnerID    string
+	WriterNode string
+	Firmware   FirmwareIdentity
+	CodePath   string
+	Source     string
+}
+
+// Metadata returns m as object metadata, keyed by the owner's kind.
+func (m ObjectMeta) Metadata() (map[string]*string, error) {
+	var ownerKey string
+	switch {
+	case strings.HasPrefix(m.OwnerID, "i-"):
+		ownerKey = "instance-id"
+	case strings.HasPrefix(m.OwnerID, "ami-"):
+		ownerKey = "image-id"
+	default:
+		return nil, fmt.Errorf("%w: owner %q is neither an instance nor an image", ErrInvalidObject, m.OwnerID)
+	}
+	switch {
+	case m.WriterNode == "":
+		return nil, fmt.Errorf("%w: %s names no writer node", ErrInvalidObject, m.OwnerID)
+	case m.Source == "":
+		return nil, fmt.Errorf("%w: %s names no source", ErrInvalidObject, m.OwnerID)
+	case m.Firmware.Arch == "" || m.Firmware.VarsSize <= 0 || m.Firmware.CodeSHA256 == "":
+		return nil, fmt.Errorf("%w: %s has an incomplete firmware identity", ErrInvalidObject, m.OwnerID)
+	}
+	return map[string]*string{
+		ownerKey:               aws.String(m.OwnerID),
+		"writer-node":          aws.String(m.WriterNode),
+		"firmware-arch":        aws.String(m.Firmware.Arch),
+		"firmware-vars-size":   aws.String(strconv.FormatInt(m.Firmware.VarsSize, 10)),
+		"firmware-code-sha256": aws.String(m.Firmware.CodeSHA256),
+		"firmware-code-path":   aws.String(m.CodePath),
+		"source":               aws.String(m.Source),
+	}, nil
+}
+
+// NewObjectKey returns a fresh key under ownerID's prefix, named by a random
+// version 4 UUID, so no write ever replaces an object a record points at.
+func NewObjectKey(ownerID string) (string, error) {
+	var b [16]byte
+	if _, err := rand.Read(b[:]); err != nil {
+		return "", fmt.Errorf("generate UEFI store key for %s: %w", ownerID, err)
+	}
+	b[6] = b[6]&0x0f | 0x40
+	b[8] = b[8]&0x3f | 0x80
+	return fmt.Sprintf("%s%x-%x-%x-%x-%x", KeyPrefix(ownerID), b[0:4], b[4:6], b[6:8], b[8:10], b[10:16]), nil
+}
+
+// PutObject writes region gzip-compressed to key with meta, within
+// PutTimeout. The region must be exactly the firmware's VARS size, because
+// pflash refuses any other.
+func PutObject(ctx context.Context, objects objectstore.ObjectStore, bucket, key string, region []byte, meta ObjectMeta) error {
+	if !ownsKey(meta.OwnerID, key) {
+		return fmt.Errorf("%w: key %q is not under %s", ErrInvalidObject, key, KeyPrefix(meta.OwnerID))
+	}
+	if int64(len(region)) != meta.Firmware.VarsSize {
+		return fmt.Errorf("%w: %s is a %d-byte region under firmware with a %d-byte VARS region",
+			ErrInvalidObject, key, len(region), meta.Firmware.VarsSize)
+	}
+	md, err := meta.Metadata()
+	if err != nil {
+		return err
+	}
+	var body bytes.Buffer
+	zw := gzip.NewWriter(&body)
+	if _, err := zw.Write(region); err != nil {
+		return fmt.Errorf("compress UEFI store %s: %w", key, err)
+	}
+	if err := zw.Close(); err != nil {
+		return fmt.Errorf("compress UEFI store %s: %w", key, err)
+	}
+
+	putCtx, cancel := context.WithTimeout(ctx, PutTimeout)
+	defer cancel()
+	if _, err := objects.PutObject(putCtx, &s3.PutObjectInput{
+		Bucket:   aws.String(bucket),
+		Key:      aws.String(key),
+		Body:     bytes.NewReader(body.Bytes()),
+		Metadata: md,
+	}); err != nil {
+		return fmt.Errorf("put UEFI store %s: %w", key, err)
+	}
+	return nil
+}

spinifex/domains/ec2/instance/uefistore/object_test.go

@@ -0,0 +1,164 @@
+package uefistore_test
+
+import (
+	"bytes"
+	"compress/gzip"
+	"context"
+	"errors"
+	"io"
+	"regexp"
+	"testing"
+
+	"github.com/aws/aws-sdk-go/aws"
+	"github.com/aws/aws-sdk-go/service/s3"
+	"github.com/mulgadc/spinifex/spinifex/domains/ec2/instance/uefistore"
+	"github.com/mulgadc/spinifex/spinifex/providers/objectstore"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+const testBucket = "instances"
+
+func testMeta(owner string) uefistore.ObjectMeta {
+	return uefistore.ObjectMeta{
+		OwnerID:    owner,
+		WriterNode: "node-1",
+		Firmware:   uefistore.FirmwareIdentity{Arch: "arm64", VarsSize: 1 << 16, CodeSHA256: "abc123"},
+		CodePath:   "/usr/share/AAVMF/AAVMF_CODE.fd",
+		Source:     uefistore.SourceTemplate,
+	}
+}
+
+var uuidV4 = regexp.MustCompile(`^[0-9a-f]{8}-[0-9a-f]{4}-4[0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$`)
+
+// Every key is a fresh version 4 UUID directly under the owner's prefix, so
+// a write never lands on an object a record already points at.
+func TestNewObjectKey_IsAFreshUUIDUnderTheOwnersPrefix(t *testing.T) {
+	seen := map[string]bool{}
+	for range 64 {
+		key, err := uefistore.NewObjectKey(instanceID)
+		require.NoError(t, err)
+		name, ok := bytes.CutPrefix([]byte(key), []byte(uefistore.KeyPrefix(instanceID)))
+		require.True(t, ok, key)
+		assert.Regexp(t, uuidV4, string(name))
+		assert.False(t, seen[key], "key %s repeated", key)
+		seen[key] = true
+	}
+}
+
+func TestObjectMeta_Metadata_NamesTheOwnerByKind(t *testing.T) {
+	md, err := testMeta(instanceID).Metadata()
+	require.NoError(t, err)
+	assert.Equal(t, map[string]string{
+		"instance-id":          instanceID,
+		"writer-node":          "node-1",
+		"firmware-arch":        "arm64",
+		"firmware-vars-size":   "65536",
+		"firmware-code-sha256": "abc123",
+		"firmware-code-path":   "/usr/share/AAVMF/AAVMF_CODE.fd",
+		"source":               "template",
+	}, aws.StringValueMap(md))
+
+	img := testMeta("ami-0123456789abcdef0")
+	img.Source = "image:ami-0fedcba9876543210"
+	md, err = img.Metadata()
+	require.NoError(t, err)
+	assert.Equal(t, "ami-0123456789abcdef0", aws.StringValue(md["image-id"]))
+	assert.NotContains(t, md, "instance-id")
+	assert.Equal(t, "image:ami-0fedcba9876543210", aws.StringValue(md["source"]))
+}
+
+func TestObjectMeta_Metadata_RefusesAnIncompleteObject(t *testing.T) {
+	tests := map[string]func(*uefistore.ObjectMeta){
+		"unknown owner":  func(m *uefistore.ObjectMeta) { m.OwnerID = "vol-1" },
+		"no writer":      func(m *uefistore.ObjectMeta) { m.WriterNode = "" },
+		"no source":      func(m *uefistore.ObjectMeta) { m.Source = "" },
+		"no arch":        func(m *uefistore.ObjectMeta) { m.Firmware.Arch = "" },
+		"no vars size":   func(m *uefistore.ObjectMeta) { m.Firmware.VarsSize = 0 },
+		"no code sha256": func(m *uefistore.ObjectMeta) { m.Firmware.CodeSHA256 = "" },
+	}
+	for name, mutate := range tests {
+		t.Run(name, func(t *testing.T) {
+			m := testMeta(instanceID)
+			mutate(&m)
+			_, err := m.Metadata()
+			require.ErrorIs(t, err, uefistore.ErrInvalidObject)
+		})
+	}
+}
+
+// The 64 MiB arm64 region is mostly erased flash, so it is stored compressed
+// and decompresses to the exact region.
+func TestPutObject_StoresTheRegionCompressed(t *testing.T) {
+	objects := objectstore.NewMemoryObjectStore()
+	meta := testMeta(instanceID)
+	region := bytes.Repeat([]byte{0xff}, int(meta.Firmware.VarsSize))
+	copy(region, "BootOrder")
+	key, err := uefistore.NewObjectKey(instanceID)
+	require.NoError(t, err)
+
+	require.NoError(t, uefistore.PutObject(t.Context(), objects, testBucket, key, region, meta))
+
+	out, err := objects.GetObject(t.Context(), &s3.GetObjectInput{Bucket: aws.String(testBucket), Key: aws.String(key)})
+	require.NoError(t, err)
+	body, err := io.ReadAll(out.Body)
+	require.NoError(t, err)
+	assert.Less(t, len(body), len(region)/10, "the erased region compresses")
+	gz, err := gzip.NewReader(bytes.NewReader(body))
+	require.NoError(t, err)
+	got, err := io.ReadAll(gz)
+	require.NoError(t, err)
+	assert.Equal(t, region, got)
+}
+
+// pflash needs a byte-exact region, and a key outside the owner's prefix is
+// one no store record owns; either is refused before anything is put.
+func TestPutObject_RefusesWhatTheOwnerCannotBoot(t *testing.T) {
+	meta := testMeta(instanceID)
+	good, err := uefistore.NewObjectKey(instanceID)
+	require.NoError(t, err)
+	other, err := uefistore.NewObjectKey("i-0fedcba9876543210")
+	require.NoError(t, err)
+	tests := []struct {
+		name   string
+		key    string
+		region []byte
+	}{
+		{name: "short region", key: good, region: make([]byte, meta.Firmware.VarsSize-1)},
+		{name: "long region", key: good, region: make([]byte, meta.Firmware.VarsSize+1)},
+		{name: "another owner's key", key: other, region: make([]byte, meta.Firmware.VarsSize)},
+		{name: "nested key", key: uefistore.KeyPrefix(instanceID) + "a/b", region: make([]byte, meta.Firmware.VarsSize)},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			objects := objectstore.NewMemoryObjectStore()
+			err := uefistore.PutObject(t.Context(), objects, testBucket, tt.key, tt.region, meta)
+			require.ErrorIs(t, err, uefistore.ErrInvalidObject)
+			out, listErr := objects.ListObjectsV2(t.Context(), &s3.ListObjectsV2Input{Bucket: aws.String(testBucket)})
+			require.NoError(t, listErr)
+			assert.Empty(t, out.Contents)
+		})
+	}
+}
+
+type failingPut struct {
+	*objectstore.MemoryObjectStore
+}
+
+func (failingPut) PutObject(ctx context.Context, _ *s3.PutObjectInput) (*s3.PutObjectOutput, error) {
+	if _, ok := ctx.Deadline(); !ok {
+		return nil, errors.New("put had no deadline")
+	}
+	return nil, errors.New("predastore unreachable")
+}
+
+// A put runs under a deadline, so a wedged object store fails the write
+// rather than holding the caller, and its error reaches the caller.
+func TestPutObject_IsBoundedAndReportsAFailedPut(t *testing.T) {
+	key, err := uefistore.NewObjectKey(instanceID)
+	require.NoError(t, err)
+	meta := testMeta(instanceID)
+	err = uefistore.PutObject(t.Context(), failingPut{objectstore.NewMemoryObjectStore()}, testBucket, key, make([]byte, meta.Firmware.VarsSize), meta)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "predastore unreachable")
+}

spinifex/domains/ec2/instance/uefistore/prune.go

@@ -0,0 +1,73 @@
+package uefistore
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"time"
+
+	"github.com/aws/aws-sdk-go/aws"
+	"github.com/aws/aws-sdk-go/service/s3"
+	"github.com/mulgadc/spinifex/spinifex/providers/objectstore"
+)
+
+// PruneMinAge is how old an object must be before a prune may delete it, so a
+// write-back whose pointer move has not landed yet never loses its object.
+const PruneMinAge = 10 * time.Minute
+
+// PruneResult is what one prune of a prefix did.
+type PruneResult struct {
+	Deleted int
+	// Young counts objects other than keep that were too new to delete, so the
+	// caller prunes the prefix again later.
+	Young int
+}
+
+// PruneObjects deletes every object under ownerID's prefix except keep that was
+// last modified before cutoff. keep must be the key the store record points at;
+// an empty or foreign keep deletes nothing. An object with no modification time
+// counts as young.
+func PruneObjects(ctx context.Context, objects objectstore.ObjectStore, bucket, ownerID, keep string, cutoff time.Time) (PruneResult, error) {
+	var res PruneResult
+	if err := checkOwnedKey(ownerID, keep); err != nil {
+		return res, fmt.Errorf("uefistore: refusing to prune %s: %w", ownerID, err)
+	}
+	contents, _, err := objectstore.ListAll(ctx, objects, &s3.ListObjectsV2Input{
+		Bucket: aws.String(bucket),
+		Prefix: aws.String(KeyPrefix(ownerID)),
+	})
+	if err != nil {
+		return res, fmt.Errorf("uefistore: list %s: %w", KeyPrefix(ownerID), err)
+	}
+	for _, obj := range contents {
+		key := aws.StringValue(obj.Key)
+		if key == keep {
+			continue
+		}
+		if obj.LastModified == nil || !obj.LastModified.Before(cutoff) {
+			res.Young++
+			continue
+		}
+		_, err := objects.DeleteObject(ctx, &s3.DeleteObjectInput{Bucket: aws.String(bucket), Key: aws.String(key)})
+		if err != nil && !objectstore.IsNoSuchKeyError(err) {
+			return res, fmt.Errorf("uefistore: delete %s: %w", key, err)
+		}
+		res.Deleted++
+	}
+	return res, nil
+}
+
+// checkOwnedKey fails a key that is empty or not one object under ownerID's
+// prefix.
+func checkOwnedKey(ownerID, key string) error {
+	if !validOwner(ownerID) {
+		return fmt.Errorf("%q is not an instance or image ID", ownerID)
+	}
+	if key == "" {
+		return errors.New("no key to keep")
+	}
+	if !ownsKey(ownerID, key) {
+		return fmt.Errorf("%s is not under %s", key, KeyPrefix(ownerID))
+	}
+	return nil
+}

spinifex/domains/ec2/instance/uefistore/prune_test.go

@@ -0,0 +1,84 @@
+package uefistore_test
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/aws/aws-sdk-go/aws"
+	"github.com/aws/aws-sdk-go/service/s3"
+	"github.com/mulgadc/spinifex/spinifex/domains/ec2/instance/uefistore"
+	"github.com/mulgadc/spinifex/spinifex/providers/objectstore"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// datedStore lists each object with the modification time the test set; a
+// key it has none for is listed with none.
+type datedStore struct {
+	*objectstore.MemoryObjectStore
+
+	modified map[string]time.Time
+}
+
+func (s *datedStore) ListObjectsV2(ctx context.Context, in *s3.ListObjectsV2Input) (*s3.ListObjectsV2Output, error) {
+	out, err := s.MemoryObjectStore.ListObjectsV2(ctx, in)
+	if err != nil {
+		return nil, err
+	}
+	for _, obj := range out.Contents {
+		obj.LastModified = nil
+		if at, ok := s.modified[aws.StringValue(obj.Key)]; ok {
+			obj.LastModified = aws.Time(at)
+		}
+	}
+	return out, nil
+}
+
+// A prune keeps the key it is told to and every object not older than the
+// cutoff, counting those it kept for age, and an undated object counts as
+// young (ADR-0003:S7).
+func TestS7_PruneObjectsKeepsThePointerAndEveryYoungObject(t *testing.T) {
+	const owner = "i-0123456789abcdef0"
+	prefix := uefistore.KeyPrefix(owner)
+	cutoff := time.Unix(1700000000, 0)
+	keep, stale, young, undated, other := prefix+"keep", prefix+"stale", prefix+"young", prefix+"undated", uefistore.KeyPrefix("i-0fedcba9876543210")+"stale"
+	store := &datedStore{MemoryObjectStore: objectstore.NewMemoryObjectStore(), modified: map[string]time.Time{
+		keep:  cutoff.Add(-time.Hour),
+		stale: cutoff.Add(-time.Second),
+		young: cutoff,
+		other: cutoff.Add(-time.Hour),
+	}}
+	for _, k := range []string{keep, stale, young, undated, other} {
+		putObject(t, store, k)
+	}
+
+	res, err := uefistore.PruneObjects(t.Context(), store, bucket, owner, keep, cutoff)
+
+	require.NoError(t, err)
+	assert.Equal(t, uefistore.PruneResult{Deleted: 1, Young: 2}, res)
+	assert.ElementsMatch(t, []string{keep, young, undated}, keysUnder(t, store, prefix),
+		"ADR-0003:S7: a prune must keep the pointed object and every object younger than the cutoff")
+	assert.Equal(t, []string{other}, keysUnder(t, store, uefistore.KeyPrefix("i-0fedcba9876543210")), "another owner's prefix is untouched")
+}
+
+// With no key to keep, or one outside the owner's prefix, a prune cannot tell
+// the pointer from a leftover, so it deletes nothing (ADR-0003:S7).
+func TestS7_PruneObjectsWithoutAKeyToKeepDeletesNothing(t *testing.T) {
+	const owner = "i-0123456789abcdef0"
+	prefix := uefistore.KeyPrefix(owner)
+	store := &datedStore{MemoryObjectStore: objectstore.NewMemoryObjectStore(), modified: map[string]time.Time{
+		prefix + "stale": time.Unix(1, 0),
+	}}
+	putObject(t, store, prefix+"stale")
+
+	for _, keep := range []string{"", uefistore.KeyPrefix("i-0fedcba9876543210") + "stale", prefix + "nested/key"} {
+		_, err := uefistore.PruneObjects(t.Context(), store, bucket, owner, keep, time.Now())
+		require.Error(t, err, keep)
+	}
+	_, err := uefistore.PruneObjects(t.Context(), store, bucket, "i-*", prefix+"stale", time.Now())
+	require.Error(t, err)
+
+	assert.Equal(t, []string{prefix + "stale"}, keysUnder(t, store, prefix),
+		"ADR-0003:S7: a prune with no valid pointer to keep must delete nothing")
+}

spinifex/domains/ec2/instance/uefistore/uefistore.go

@@ -1,6 +1,6 @@
 // Package uefistore holds the types that describe an instance's UEFI variable
-// store and the bucket of store records that carry its deletion and write-back
-// obligations past the life of the instance record.
+// store, the bucket of store records that carry its pointer and deletion
+// obligation past the life of the instance record, and the claim that fences it.
 package uefistore
 
 import (
@@ -53,22 +53,27 @@ type FirmwareIdentity struct {
 // UEFIStoreRecord is the durable owner of one instance's variable store
 // objects. It outlives the instance record, so an obligation recorded here is
 // not lost when a terminated record expires.
+//
+// Pointer names the only version of the store the record keeps, the object the
+// instance boots from, and only a writer whose claim token has generation
+// Generation may move it.
 type UEFIStoreRecord struct {
-	SchemaVersion uint16         `json:"schema_version"`
-	InstanceID    string         `json:"instance_id"`
-	AccountID     string         `json:"account_id"`
-	State         string         `json:"state"`
-	WriteBack     *UEFIWriteBack `json:"write_back,omitempty"`
+	SchemaVersion uint16           `json:"schema_version"`
+	InstanceID    string           `json:"instance_id"`
+	AccountID     string           `json:"account_id"`
+	State         string           `json:"state"`
+	Generation    uint64           `json:"generation"`
+	Pointer       *UEFIVarStoreRef `json:"pointer,omitempty"`
 }
 
-// UEFIWriteBack is a write-back that failed and is still owed. Key is the
-// object every retry puts to, so retries leave no extra objects; SeededFrom is
-// the pointer Node seeded from, which a retry must still find on the record.
-type UEFIWriteBack struct {
-	Node       string    `json:"node"`
-	Key        string    `json:"key"`
-	SeededFrom string    `json:"seeded_from"`
-	Since      time.Time `json:"since"`
+// UEFIVarStoreRef names the store object an instance launches from, with the
+// firmware it was written under, so a launch can check it without reading the
+// object's metadata.
+type UEFIVarStoreRef struct {
+	Key        string           `json:"key"`
+	Firmware   FirmwareIdentity `json:"firmware"`
+	WriterNode string           `json:"writer_node"`
+	WrittenAt  time.Time        `json:"written_at"`
 }
 
 // KeyPrefix is the object-store prefix every store object of ownerID (an
@@ -83,8 +88,8 @@ func ownsKey(ownerID, key string) bool {
 	return ok && name != "" && !strings.Contains(name, "/")
 }
 
-// validate checks rec against key. A write-back key outside the instance's own
-// prefix is refused because a superseded obligation deletes that key.
+// validate checks rec against key. The pointer must name one object directly
+// under the instance's own prefix, because a deletion acts on it.
 func (rec *UEFIStoreRecord) validate(key string) error {
 	switch {
 	case rec.SchemaVersion > RecordSchemaVersion:
@@ -98,29 +103,32 @@ func (rec *UEFIStoreRecord) validate(key string) error {
 		return fmt.Errorf("%w: instance_id %s stored at key %s", ErrInvalidRecord, rec.InstanceID, key)
 	case rec.State != StateLive && rec.State != StateDeleting:
 		return fmt.Errorf("%w: %s has state %q", ErrInvalidRecord, key, rec.State)
-	case rec.State == StateDeleting && rec.WriteBack != nil:
-		// A write-back retry puts an object, which would land behind the delete.
-		return fmt.Errorf("%w: %s is deleting with a write-back still owed", ErrInvalidRecord, key)
 	}
-	if wb := rec.WriteBack; wb != nil {
-		switch {
-		case wb.Node == "":
-			return fmt.Errorf("%w: %s write-back names no node", ErrInvalidRecord, key)
-		case wb.Since.IsZero():
-			return fmt.Errorf("%w: %s write-back has no since", ErrInvalidRecord, key)
-		case !ownsKey(rec.InstanceID, wb.Key):
-			return fmt.Errorf("%w: %s write-back key %q is not under %s", ErrInvalidRecord, key, wb.Key, KeyPrefix(rec.InstanceID))
-		case !ownsKey(rec.InstanceID, wb.SeededFrom):
-			return fmt.Errorf("%w: %s write-back seeded_from %q is not under %s", ErrInvalidRecord, key, wb.SeededFrom, KeyPrefix(rec.InstanceID))
+	if p := rec.Pointer; p != nil {
+		if err := p.validate(rec.InstanceID); err != nil {
+			return fmt.Errorf("%w: %s pointer: %w", ErrInvalidRecord, key, err)
 		}
 	}
 	return nil
 }
 
+// validate checks ref as a pointer of instanceID's store.
+func (ref *UEFIVarStoreRef) validate(instanceID string) error {
+	switch {
+	case !ownsKey(instanceID, ref.Key):
+		return fmt.Errorf("key %q is not under %s", ref.Key, KeyPrefix(instanceID))
+	case ref.Firmware.Arch == "" || ref.Firmware.VarsSize <= 0:
+		return fmt.Errorf("key %q names no firmware arch and VARS size", ref.Key)
+	case ref.WriterNode == "":
+		return fmt.Errorf("key %q names no writer node", ref.Key)
+	}
+	return nil
+}
+
 func bucketConfig() kvstore.Config {
 	return kvstore.Config{
 		Name:        BucketName,
-		Description: "Spinifex UEFI variable store records (deletion and write-back obligations)",
+		Description: "Spinifex UEFI variable store records (pointers and deletion obligations)",
 		History:     1,
 		// Left false: a lost bucket has lost obligations, and an empty
 		// recreated one would report every store as having none.
@@ -172,6 +180,19 @@ func (s *Store) Get(ctx context.Context, instanceID string) (*UEFIStoreRecord, u
 	return rec, rev, nil
 }
 
+// LeaderGet is Get answered by the stream leader, for a caller that deletes
+// objects by what the record does not point at.
+func (s *Store) LeaderGet(ctx context.Context, instanceID string) (*UEFIStoreRecord, uint64, error) {
+	rec, rev, err := s.kv.LeaderGet(ctx, instanceID)
+	if err != nil {
+		return nil, 0, err
+	}
+	if err := rec.validate(instanceID); err != nil {
+		return nil, 0, err
+	}
+	return rec, rev, nil
+}
+
 // Create writes a new record, which must be live: a store is recorded before
 // its first object is written. An existing record is kvstore.ErrExists. Every
 // write stamps rec with RecordSchemaVersion.

spinifex/domains/ec2/instance/uefistore/uefistore_test.go

@@ -11,7 +11,6 @@ import (
 	testutil "github.com/mulgadc/spinifex/internal/testkit"
 	"github.com/mulgadc/spinifex/spinifex/domains/ec2/instance/uefistore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
 	"github.com/nats-io/nats.go/jetstream"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -31,15 +30,25 @@ func liveRecord() *uefistore.UEFIStoreRecord {
 	return &uefistore.UEFIStoreRecord{InstanceID: instanceID, AccountID: "111122223333", State: uefistore.StateLive}
 }
 
-func writeBack() *uefistore.UEFIWriteBack {
-	return &uefistore.UEFIWriteBack{
-		Node:       "node-2",
-		Key:        uefistore.KeyPrefix(instanceID) + "b",
-		SeededFrom: uefistore.KeyPrefix(instanceID) + "a",
-		Since:      time.Date(2026, 11, 3, 1, 2, 3, 0, time.UTC),
+func pointer(name string) *uefistore.UEFIVarStoreRef {
+	return &uefistore.UEFIVarStoreRef{
+		Key:        keyOf(name),
+		Firmware:   uefistore.FirmwareIdentity{Arch: "x86_64", VarsSize: 540672, CodeSHA256: "c0de"},
+		WriterNode: "node-2",
+		WrittenAt:  time.Date(2026, 11, 3, 1, 2, 3, 0, time.UTC),
 	}
 }
 
+func keyOf(name string) string { return uefistore.KeyPrefix(instanceID) + name }
+
+// pointedRecord is a live record stamped at generation 7 that points at "b".
+func pointedRecord() *uefistore.UEFIStoreRecord {
+	rec := liveRecord()
+	rec.Generation = 7
+	rec.Pointer = pointer("b")
+	return rec
+}
+
 func rawPut(t *testing.T, js jetstream.JetStream, key, body string) {
 	t.Helper()
 	kv, err := js.KeyValue(t.Context(), uefistore.BucketName)
@@ -60,15 +69,14 @@ func TestOpen_CreatesBucketWithNoTTLAndStampsVersion(t *testing.T) {
 	assert.Equal(t, int64(1), status.History())
 	assert.Zero(t, status.TTL(), "a TTL would expire deletion obligations before they are settled")
 
-	v, err := kvutil.ReadVersion(t.Context(), kv)
+	v, err := kvstore.ReadVersion(t.Context(), kv)
 	require.NoError(t, err)
 	assert.Equal(t, uefistore.BucketVersion, v)
 }
 
 func TestCreate_RoundTripsAndStampsSchemaVersion(t *testing.T) {
 	s, _ := openStore(t)
-	rec := liveRecord()
-	rec.WriteBack = writeBack()
+	rec := pointedRecord()
 
 	rev, err := s.Create(t.Context(), rec)
 	require.NoError(t, err)
@@ -83,9 +91,7 @@ func TestCreate_RoundTripsAndStampsSchemaVersion(t *testing.T) {
 // The wire names are what every release reads, so they are pinned.
 func TestRecord_WireShape(t *testing.T) {
 	s, js := openStore(t)
-	rec := liveRecord()
-	rec.WriteBack = writeBack()
-	_, err := s.Create(t.Context(), rec)
+	_, err := s.Create(t.Context(), pointedRecord())
 	require.NoError(t, err)
 
 	kv, err := js.KeyValue(t.Context(), uefistore.BucketName)
@@ -97,11 +103,12 @@ func TestRecord_WireShape(t *testing.T) {
 		"instance_id": "i-0123456789abcdef0",
 		"account_id": "111122223333",
 		"state": "live",
-		"write_back": {
-			"node": "node-2",
+		"generation": 7,
+		"pointer": {
 			"key": "uefi/i-0123456789abcdef0/b",
-			"seeded_from": "uefi/i-0123456789abcdef0/a",
-			"since": "2026-11-03T01:02:03Z"
+			"firmware": {"arch": "x86_64", "vars_size": 540672, "code_sha256": "c0de"},
+			"writer_node": "node-2",
+			"written_at": "2026-11-03T01:02:03Z"
 		}
 	}`, string(entry.Value()))
 }
@@ -145,8 +152,7 @@ func TestCompareAndSet_StaleRevisionConflictsAndWritesNothing(t *testing.T) {
 	first, err := s.Create(t.Context(), liveRecord())
 	require.NoError(t, err)
 
-	moved := liveRecord()
-	moved.WriteBack = writeBack()
+	moved := pointedRecord()
 	second, err := s.CompareAndSet(t.Context(), moved, first)
 	require.NoError(t, err)
 	assert.Greater(t, second, first)
@@ -160,7 +166,7 @@ func TestCompareAndSet_StaleRevisionConflictsAndWritesNothing(t *testing.T) {
 	require.NoError(t, err)
 	assert.Equal(t, second, rev)
 	assert.Equal(t, uefistore.StateLive, got.State)
-	assert.Equal(t, writeBack(), got.WriteBack)
+	assert.Equal(t, pointer("b"), got.Pointer)
 }
 
 // Two actors that read the same revision race to change it; exactly one may
@@ -180,12 +186,11 @@ func TestCompareAndSet_ConcurrentWritersFromOneReadHaveOneWinner(t *testing.T) {
 	)
 	for i := range writers {
 		wg.Go(func() {
-			rec := liveRecord()
-			rec.WriteBack = writeBack()
-			rec.WriteBack.Node = fmt.Sprintf("node-%d", i)
+			rec := pointedRecord()
+			rec.Pointer.WriterNode = fmt.Sprintf("node-%d", i)
 			if _, err := s.CompareAndSet(context.Background(), rec, rev); err == nil {
 				mu.Lock()
-				wins = append(wins, rec.WriteBack.Node)
+				wins = append(wins, rec.Pointer.WriterNode)
 				mu.Unlock()
 			} else {
 				assert.ErrorIs(t, err, kvstore.ErrConflict)
@@ -197,7 +202,7 @@ func TestCompareAndSet_ConcurrentWritersFromOneReadHaveOneWinner(t *testing.T) {
 	require.Len(t, wins, 1)
 	got, _, err := s.Get(t.Context(), instanceID)
 	require.NoError(t, err)
-	assert.Equal(t, wins[0], got.WriteBack.Node)
+	assert.Equal(t, wins[0], got.Pointer.WriterNode)
 }
 
 func TestCompareAndDelete_StaleRevisionKeepsRecord(t *testing.T) {
@@ -218,29 +223,20 @@ func TestCompareAndDelete_StaleRevisionKeepsRecord(t *testing.T) {
 	require.ErrorIs(t, err, kvstore.ErrNotFound)
 }
 
-// A superseded obligation deletes its recorded key, so a key outside the
+// A deletion acts on the key the record points at, so a key outside the
 // instance's own prefix must never reach the bucket.
 func TestCompareAndSet_RefusesInvalidRecords(t *testing.T) {
+	other := uefistore.KeyPrefix("i-0fedcba9876543210") + "b"
 	cases := map[string]func(*uefistore.UEFIStoreRecord){
-		"unknown state":            func(r *uefistore.UEFIStoreRecord) { r.State = "gone" },
-		"not an instance ID":       func(r *uefistore.UEFIStoreRecord) { r.InstanceID = "ami-0123456789abcdef0" },
-		"write-back with no node":  func(r *uefistore.UEFIStoreRecord) { r.WriteBack.Node = "" },
-		"write-back with no since": func(r *uefistore.UEFIStoreRecord) { r.WriteBack.Since = time.Time{} },
-		"write-back key under another instance": func(r *uefistore.UEFIStoreRecord) {
-			r.WriteBack.Key = uefistore.KeyPrefix("i-0fedcba9876543210") + "b"
-		},
-		"write-back key under an image": func(r *uefistore.UEFIStoreRecord) {
-			r.WriteBack.Key = uefistore.KeyPrefix("ami-0123456789abcdef0") + "b"
-		},
-		"write-back key is the prefix itself": func(r *uefistore.UEFIStoreRecord) {
-			r.WriteBack.Key = uefistore.KeyPrefix(instanceID)
-		},
-		"write-back key nested below the prefix": func(r *uefistore.UEFIStoreRecord) {
-			r.WriteBack.Key = uefistore.KeyPrefix(instanceID) + "x/y"
-		},
-		"write-back seeded from another instance": func(r *uefistore.UEFIStoreRecord) {
-			r.WriteBack.SeededFrom = uefistore.KeyPrefix("i-0fedcba9876543210") + "a"
-		},
+		"unknown state":                   func(r *uefistore.UEFIStoreRecord) { r.State = "gone" },
+		"not an instance ID":              func(r *uefistore.UEFIStoreRecord) { r.InstanceID = "ami-0123456789abcdef0" },
+		"pointer under another instance":  func(r *uefistore.UEFIStoreRecord) { r.Pointer.Key = other },
+		"pointer under an image":          func(r *uefistore.UEFIStoreRecord) { r.Pointer.Key = uefistore.KeyPrefix("ami-0123456789abcdef0") + "b" },
+		"pointer is the prefix itself":    func(r *uefistore.UEFIStoreRecord) { r.Pointer.Key = uefistore.KeyPrefix(instanceID) },
+		"pointer nested below the prefix": func(r *uefistore.UEFIStoreRecord) { r.Pointer.Key = keyOf("x/y") },
+		"pointer with no firmware arch":   func(r *uefistore.UEFIStoreRecord) { r.Pointer.Firmware.Arch = "" },
+		"pointer with no VARS size":       func(r *uefistore.UEFIStoreRecord) { r.Pointer.Firmware.VarsSize = 0 },
+		"pointer with no writer node":     func(r *uefistore.UEFIStoreRecord) { r.Pointer.WriterNode = "" },
 	}
 	for name, mutate := range cases {
 		t.Run(name, func(t *testing.T) {
@@ -248,8 +244,7 @@ func TestCompareAndSet_RefusesInvalidRecords(t *testing.T) {
 			rev, err := s.Create(t.Context(), liveRecord())
 			require.NoError(t, err)
 
-			rec := liveRecord()
-			rec.WriteBack = writeBack()
+			rec := pointedRecord()
 			mutate(rec)
 			_, err = s.CompareAndSet(t.Context(), rec, rev)
 			require.ErrorIs(t, err, uefistore.ErrInvalidRecord)
@@ -257,7 +252,7 @@ func TestCompareAndSet_RefusesInvalidRecords(t *testing.T) {
 			got, gotRev, err := s.Get(t.Context(), instanceID)
 			require.NoError(t, err)
 			assert.Equal(t, rev, gotRev, "an invalid record must not be written")
-			assert.Nil(t, got.WriteBack)
+			assert.Nil(t, got.Pointer)
 		})
 	}
 }
@@ -266,11 +261,11 @@ func TestCompareAndSet_RefusesInvalidRecords(t *testing.T) {
 // is reported rather than decoded into something a reaper would act on.
 func TestGet_RefusesRecordsItCannotTrust(t *testing.T) {
 	cases := map[string]string{
-		"newer schema":       `{"schema_version":2,"instance_id":"i-0123456789abcdef0","state":"live"}`,
-		"no schema":          `{"instance_id":"i-0123456789abcdef0","state":"live"}`,
-		"another instance":   `{"schema_version":1,"instance_id":"i-0fedcba9876543210","state":"deleting"}`,
-		"unknown state":      `{"schema_version":1,"instance_id":"i-0123456789abcdef0","state":"purged"}`,
-		"foreign write-back": `{"schema_version":1,"instance_id":"i-0123456789abcdef0","state":"live","write_back":{"node":"n","key":"uefi/i-0fedcba9876543210/b","seeded_from":"uefi/i-0123456789abcdef0/a","since":"2026-11-03T00:00:00Z"}}`,
+		"newer schema":     `{"schema_version":2,"instance_id":"i-0123456789abcdef0","state":"live"}`,
+		"no schema":        `{"instance_id":"i-0123456789abcdef0","state":"live"}`,
+		"another instance": `{"schema_version":1,"instance_id":"i-0fedcba9876543210","state":"deleting"}`,
+		"unknown state":    `{"schema_version":1,"instance_id":"i-0123456789abcdef0","state":"purged"}`,
+		"foreign pointer":  `{"schema_version":1,"instance_id":"i-0123456789abcdef0","state":"live","generation":3,"pointer":{"key":"uefi/i-0fedcba9876543210/b","firmware":{"arch":"x86_64","vars_size":540672},"writer_node":"n"}}`,
 	}
 	for name, body := range cases {
 		t.Run(name, func(t *testing.T) {

spinifex/domains/ec2/instance/volume_held_elsewhere_test.go

@@ -102,3 +102,38 @@ func TestStartPaths_ReportAVolumeHeldElsewhereByCode(t *testing.T) {
 		assert.Equal(t, awserrors.ErrorIncorrectState, awserrors.ValidErrorCodeFromError(err))
 	})
 }
+
+// uefiClaimHeldMounter fails the launch the way a start does while a live
+// node holds the instance's UEFI variable store.
+type uefiClaimHeldMounter struct{ raceVolumeMounter }
+
+func (uefiClaimHeldMounter) Mount(context.Context, *vm.VM) error {
+	return fmt.Errorf("%w: node-a holds i-uefi: %w", vm.ErrUEFIClaimHeld, vm.ErrMountRetryable)
+}
+
+// A held UEFI store is a capacity refusal on both start paths, so the caller
+// tries another node rather than reporting a fault.
+func TestStartPaths_ReportAHeldUEFIStoreAsCapacity(t *testing.T) {
+	t.Setenv("XDG_RUNTIME_DIR", t.TempDir())
+	types, _ := defaultPrepareInstanceTypes()
+
+	t.Run("claimed relaunch", func(t *testing.T) {
+		mgr := vm.NewManager()
+		mgr.SetDeps(vm.Deps{NodeID: "node-b", VolumeMounter: uefiClaimHeldMounter{}})
+		svc := &InstanceServiceImpl{resourceMgr: &fakeResourceCapacityProvider{instanceTypes: types}, vmMgr: mgr}
+
+		err := svc.launchClaimedInstance(t.Context(), &vm.VM{ID: "i-uefi-claim", AccountID: "acc", InstanceType: "t3.micro", Status: vm.StatePending}, func() {})
+		assert.Equal(t, awserrors.ErrorInsufficientInstanceCapacity, awserrors.ValidErrorCodeFromError(err))
+	})
+
+	t.Run("live owner", func(t *testing.T) {
+		const id = "i-uefi-live"
+		mgr := mgrWith(map[string]*vm.VM{id: {ID: id, Status: vm.StateStopped, InstanceType: "t3.micro"}})
+		mgr.SetDeps(vm.Deps{NodeID: "node-b", VolumeMounter: uefiClaimHeldMounter{}})
+		v, _ := mgr.Get(id)
+		svc := &InstanceServiceImpl{resourceMgr: &fakeResourceCapacityProvider{instanceTypes: types}, vmMgr: mgr}
+
+		err := svc.StartInstance(t.Context(), v, ec2v1.EC2InstanceCommand{ID: id})
+		assert.Equal(t, awserrors.ErrorInsufficientInstanceCapacity, awserrors.ValidErrorCodeFromError(err))
+	})
+}

spinifex/domains/ec2/inventory/inventory.go

@@ -1,67 +0,0 @@
-// Package inventory is the EC2-owned read capability for an account's complete
-// instance list, for callers outside the EC2 domain.
-package inventory
-
-import (
-	"context"
-	"encoding/json"
-	"errors"
-	"fmt"
-
-	"github.com/aws/aws-sdk-go/service/ec2"
-	ec2instanceapi "github.com/mulgadc/spinifex/spinifex/domains/ec2/awsapi/instance"
-	ec2vpc "github.com/mulgadc/spinifex/spinifex/domains/ec2/vpc"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
-	"github.com/nats-io/nats.go"
-	"github.com/nats-io/nats.go/jetstream"
-)
-
-// Instances lists every non-terminated instance an account owns, or returns an
-// error. It never returns a shorter list than the truth.
-type Instances interface {
-	ListAccountInstances(ctx context.Context, accountID string) (*ec2.DescribeInstancesOutput, error)
-}
-
-// NATSInstances reads the complete list across nodeIDs over NATS.
-type NATSInstances struct {
-	nc      *nats.Conn
-	nodeIDs []string
-}
-
-var _ Instances = (*NATSInstances)(nil)
-
-// NewNATSInstances returns a reader over the configured node set.
-func NewNATSInstances(nc *nats.Conn, nodeIDs []string) *NATSInstances {
-	return &NATSInstances{nc: nc, nodeIDs: nodeIDs}
-}
-
-// ListAccountInstances passes through the strict listing unchanged: any node
-// that did not answer, or any record that could not be read, is an error.
-func (n *NATSInstances) ListAccountInstances(ctx context.Context, accountID string) (*ec2.DescribeInstancesOutput, error) {
-	return ec2instanceapi.DescribeAccountInstancesStrict(ctx, n.nc, n.nodeIDs, accountID)
-}
-
-// NewENITagReader returns a reader of one ENI's tags straight from its stored
-// record, so an ENI that is absent and one that cannot be read are told apart:
-// a missing key is found=false, a failed read or an undecodable record is an
-// error. accountID is the account the ENI is stored under.
-func NewENITagReader(js jetstream.JetStream, accountID string) func(ctx context.Context, eniID string) (map[string]string, bool, error) {
-	return func(ctx context.Context, eniID string) (map[string]string, bool, error) {
-		kv, err := js.KeyValue(ctx, ec2vpc.KVBucketENIs)
-		if err != nil {
-			return nil, false, fmt.Errorf("open ENI bucket: %w", err)
-		}
-		entry, err := kv.Get(ctx, kvutil.AccountKey(accountID, eniID))
-		if errors.Is(err, jetstream.ErrKeyNotFound) {
-			return nil, false, nil
-		}
-		if err != nil {
-			return nil, false, fmt.Errorf("read ENI record %s: %w", eniID, err)
-		}
-		var record ec2vpc.ENIRecord
-		if err := json.Unmarshal(entry.Value(), &record); err != nil {
-			return nil, false, fmt.Errorf("decode ENI record %s: %w", eniID, err)
-		}
-		return record.Tags, true, nil
-	}
-}

spinifex/domains/ec2/inventory/inventory_test.go

@@ -1,49 +0,0 @@
-package inventory
-
-import (
-	"encoding/json"
-	"testing"
-
-	testutil "github.com/mulgadc/spinifex/internal/testkit"
-	ec2vpc "github.com/mulgadc/spinifex/spinifex/domains/ec2/vpc"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestListAccountInstancesRefusesAnUnknownNodeSet(t *testing.T) {
-	out, err := NewNATSInstances(nil, nil).ListAccountInstances(t.Context(), "000000000000")
-	require.Error(t, err)
-	assert.Nil(t, out, "an unknown node set must not yield a list")
-}
-
-func TestENITagReaderSeparatesGoneFromUnreadable(t *testing.T) {
-	_, nc, _ := testutil.StartTestJetStream(t)
-	js := testutil.NewJetStream(t, nc)
-	read := NewENITagReader(js, "000000000000")
-
-	_, _, err := read(t.Context(), "eni-1")
-	require.Error(t, err, "a missing ENI bucket is a read failure, not an absent ENI")
-
-	enis, err := kvutil.GetOrCreateBucket(t.Context(), js, ec2vpc.KVBucketENIs, 10)
-	require.NoError(t, err)
-	good, err := json.Marshal(ec2vpc.ENIRecord{NetworkInterfaceId: "eni-good", Tags: map[string]string{"k": "v"}})
-	require.NoError(t, err)
-	_, err = enis.Put(t.Context(), kvutil.AccountKey("000000000000", "eni-good"), good)
-	require.NoError(t, err)
-	_, err = enis.Put(t.Context(), kvutil.AccountKey("000000000000", "eni-corrupt"), []byte("{not json"))
-	require.NoError(t, err)
-
-	tags, found, err := read(t.Context(), "eni-good")
-	require.NoError(t, err)
-	assert.True(t, found)
-	assert.Equal(t, map[string]string{"k": "v"}, tags)
-
-	_, found, err = read(t.Context(), "eni-absent")
-	require.NoError(t, err)
-	assert.False(t, found, "a missing key is the ENI gone")
-
-	_, found, err = read(t.Context(), "eni-corrupt")
-	require.Error(t, err, "an undecodable record is a read failure")
-	assert.False(t, found)
-}

spinifex/domains/ec2/launchtemplate/account_cleanup.go

@@ -9,7 +9,7 @@ import (
 	"strings"
 	"time"
 
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -33,7 +33,7 @@ type AccountCleanup struct {
 // Unreadable headers, their bodies and unrecognised entries are left for the
 // owner scan to report. A listing failure is returned and nothing is removed.
 func (s *LaunchTemplateServiceImpl) CleanupAccountResidue(ctx context.Context, accountID string, settledBefore time.Time) (*AccountCleanup, error) {
-	keys, err := kvutil.Keys(ctx, s.kv)
+	keys, err := kvstore.Keys(ctx, s.kv)
 	if err != nil && !errors.Is(err, jetstream.ErrNoKeysFound) {
 		return nil, fmt.Errorf("list launch-template keys: %w", err)
 	}
@@ -145,7 +145,7 @@ type AccountTemplateDeletes struct {
 // it cannot read or decode is left for the owner scan to report. A listing
 // failure is returned and nothing is deleted.
 func (s *LaunchTemplateServiceImpl) DeleteLiveTemplates(ctx context.Context, accountID string, referenced func(id, name string) string) (*AccountTemplateDeletes, error) {
-	keys, err := kvutil.Keys(ctx, s.kv)
+	keys, err := kvstore.Keys(ctx, s.kv)
 	if err != nil && !errors.Is(err, jetstream.ErrNoKeysFound) {
 		return nil, fmt.Errorf("list launch-template keys: %w", err)
 	}

spinifex/domains/ec2/launchtemplate/account_scan.go

@@ -9,7 +9,7 @@ import (
 	"strconv"
 	"strings"
 
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -37,7 +37,7 @@ type AccountScan struct {
 // Headers and name-index entries are read; version bodies are classified from
 // their keys alone, so template data and user data are never loaded.
 func (s *LaunchTemplateServiceImpl) ScanAccount(ctx context.Context, accountID string) (*AccountScan, error) {
-	keys, err := kvutil.Keys(ctx, s.kv)
+	keys, err := kvstore.Keys(ctx, s.kv)
 	if err != nil && !errors.Is(err, jetstream.ErrNoKeysFound) {
 		return nil, fmt.Errorf("list launch-template keys: %w", err)
 	}

spinifex/domains/ec2/launchtemplate/account_scan_test.go

@@ -11,7 +11,7 @@ import (
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
 	"github.com/mulgadc/spinifex/internal/testkit/readonlystore"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -45,7 +45,7 @@ func TestScanAccountClassifiesLaunchTemplateRecords(t *testing.T) {
 	// A header removed alone, as a delete interrupted after it leaves the rest.
 	deleted := createWithUserData(t, svc, "deleted")
 	inner := svc.kv
-	require.NoError(t, inner.Delete(t.Context(), kvutil.AccountKey(testAccountID, deleted)))
+	require.NoError(t, inner.Delete(t.Context(), kvstore.AccountKey(testAccountID, deleted)))
 
 	// A create whose header write and clean-up both fail.
 	svc.kv = &faultyKV{KeyValue: inner,
@@ -59,16 +59,16 @@ func TestScanAccountClassifiesLaunchTemplateRecords(t *testing.T) {
 	require.Error(t, err)
 	svc.kv = inner
 
-	putRaw(t, inner, kvutil.AccountKey(testAccountID, "lt-0badbadbadbadbad0"), "{not json")
-	putRaw(t, inner, kvutil.AccountKey(testAccountID, "lt-0badbadbadbadbad0.v1"), `{"version_number":1}`)
-	putRaw(t, inner, kvutil.AccountKey(testAccountID, "lt-0mismatch00000000"), `{"launch_template_id":"lt-0other"}`)
-	putRaw(t, inner, kvutil.AccountKey(testAccountID, "lt-0unreadable000000"), `{"launch_template_id":"lt-0unreadable000000"}`)
+	putRaw(t, inner, kvstore.AccountKey(testAccountID, "lt-0badbadbadbadbad0"), "{not json")
+	putRaw(t, inner, kvstore.AccountKey(testAccountID, "lt-0badbadbadbadbad0.v1"), `{"version_number":1}`)
+	putRaw(t, inner, kvstore.AccountKey(testAccountID, "lt-0mismatch00000000"), `{"launch_template_id":"lt-0other"}`)
+	putRaw(t, inner, kvstore.AccountKey(testAccountID, "lt-0unreadable000000"), `{"launch_template_id":"lt-0unreadable000000"}`)
 	putRaw(t, inner, nameKey(testAccountID, "unreadable"), "lt-0unreadable000000")
-	putRaw(t, inner, kvutil.AccountKey(testAccountID, "lt-0bad.vx"), "{}")
+	putRaw(t, inner, kvstore.AccountKey(testAccountID, "lt-0bad.vx"), "{}")
 
 	recorder := &readRecordingKV{KeyValue: readonlystore.NewKV(t, inner)}
 	svc.kv = &faultyKV{KeyValue: recorder, failGet: func(key string) bool {
-		return key == kvutil.AccountKey(testAccountID, "lt-0unreadable000000")
+		return key == kvstore.AccountKey(testAccountID, "lt-0unreadable000000")
 	}}
 	scan, err := svc.ScanAccount(context.Background(), testAccountID)
 	require.NoError(t, err)

spinifex/domains/ec2/launchtemplate/account_termination_current_behaviour_test.go

@@ -16,7 +16,7 @@ import (
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -159,7 +159,7 @@ func TestCurrentBehaviour_AccountTermination_LaunchTemplateStorageLayout(t *test
 
 	header := testAccountID + "." + id
 	name := testAccountID + ".name.776562283129" // hex of "web(1)"
-	assert.Equal(t, sortedKeys(header, header+".v1", header+".v2", name, kvutil.VersionKey), storedKeys(t, svc.kv))
+	assert.Equal(t, sortedKeys(header, header+".v1", header+".v2", name, kvstore.VersionKey), storedKeys(t, svc.kv))
 
 	entry, err := svc.kv.Get(t.Context(), name)
 	require.NoError(t, err)
@@ -223,13 +223,13 @@ func TestAccountTermination_DeleteLaunchTemplateKeepsTheTemplateWhenAStepFails(t
 			assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
 
 			svc.kv = inner
-			assert.Equal(t, sortedKeys(append(tc.left(header, nameIdx), kvutil.VersionKey)...), storedKeys(t, inner))
+			assert.Equal(t, sortedKeys(append(tc.left(header, nameIdx), kvstore.VersionKey)...), storedKeys(t, inner))
 			assert.Equal(t, []string{id}, describedTemplateIDs(t, svc, testAccountID))
 
 			out, err := svc.DeleteLaunchTemplate(context.Background(), &ec2.DeleteLaunchTemplateInput{LaunchTemplateId: aws.String(id)}, testAccountID)
 			require.NoError(t, err)
 			assert.Equal(t, id, aws.StringValue(out.LaunchTemplate.LaunchTemplateId))
-			assert.Equal(t, []string{kvutil.VersionKey}, storedKeys(t, inner))
+			assert.Equal(t, []string{kvstore.VersionKey}, storedKeys(t, inner))
 			assert.Empty(t, describedTemplateIDs(t, svc, testAccountID))
 		})
 	}
@@ -260,7 +260,7 @@ func TestAccountTermination_DeleteLaunchTemplateRestartsWhenTheHeaderChanges(t *
 	_, err = svc.DeleteLaunchTemplate(context.Background(), &ec2.DeleteLaunchTemplateInput{LaunchTemplateId: aws.String(id)}, testAccountID)
 	require.NoError(t, err)
 	assert.True(t, raced)
-	assert.Equal(t, []string{kvutil.VersionKey}, storedKeys(t, inner))
+	assert.Equal(t, []string{kvstore.VersionKey}, storedKeys(t, inner))
 }
 
 // A create that fails after claiming the name removes what it wrote, so no
@@ -287,7 +287,7 @@ func TestAccountTermination_FailedCreateRemovesWhatItWrote(t *testing.T) {
 			assert.Equal(t, awserrors.ErrorServerInternal, err.Error())
 
 			svc.kv = inner
-			assert.Equal(t, []string{kvutil.VersionKey}, storedKeys(t, inner))
+			assert.Equal(t, []string{kvstore.VersionKey}, storedKeys(t, inner))
 			retried := createTemplate(t, svc, "web", "t3.micro")
 			assert.Equal(t, []string{aws.StringValue(retried.LaunchTemplateId)}, describedTemplateIDs(t, svc, testAccountID))
 		})
@@ -299,7 +299,7 @@ func TestAccountTermination_FailedCreateRemovesWhatItWrote(t *testing.T) {
 // failure is an internal error, never a missing template.
 func TestCurrentBehaviour_AccountTermination_DescribeLaunchTemplatesHidesUnreadableHeaders(t *testing.T) {
 	const brokenID = "lt-0badbadbadbadbad0"
-	brokenHeader := kvutil.AccountKey(testAccountID, brokenID)
+	brokenHeader := kvstore.AccountKey(testAccountID, brokenID)
 	tests := []struct {
 		name       string
 		seed       func(t *testing.T, svc *LaunchTemplateServiceImpl)
@@ -350,7 +350,7 @@ func TestAccountCleanup_KeepsTemplatesAndRemovesSettledResidue(t *testing.T) {
 	require.NoError(t, err)
 	orphanBody := versionKey(testAccountID, "lt-0orphan000000000", 1)
 	orphanName := nameKey(testAccountID, "gone")
-	brokenHeader := kvutil.AccountKey(testAccountID, "lt-0broken000000000")
+	brokenHeader := kvstore.AccountKey(testAccountID, "lt-0broken000000000")
 	brokenBody := versionKey(testAccountID, "lt-0broken000000000", 1)
 	brokenName := nameKey(testAccountID, "broken")
 	for key, value := range map[string]string{
@@ -384,7 +384,7 @@ func TestAccountCleanup_KeepsTemplatesAndRemovesSettledResidue(t *testing.T) {
 func TestAccountCleanup_InterruptedTemplateDeleteIsLeftToThePublicRetry(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	kv := svc.kv
 	svc.kv = &faultyKV{KeyValue: kv, failDelete: func(key string) bool { return key == header }}
 	_, err := svc.DeleteLaunchTemplate(context.Background(), &ec2.DeleteLaunchTemplateInput{LaunchTemplateId: aws.String(id)}, testAccountID)

spinifex/domains/ec2/launchtemplate/service_impl.go

@@ -19,7 +19,7 @@ import (
 	awsfilters "github.com/mulgadc/spinifex/spinifex/foundation/aws/filters"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
 	awstags "github.com/mulgadc/spinifex/spinifex/foundation/aws/tags"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -64,7 +64,7 @@ func NewLaunchTemplateServiceImplWithNATS(ctx context.Context, cfg *config.Confi
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	kv, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketLaunchTemplates, 10)
+	kv, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketLaunchTemplates, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketLaunchTemplates, err)
 	}
@@ -87,24 +87,24 @@ func NewLaunchTemplateServiceImplWithNATS(ctx context.Context, cfg *config.Confi
 // Launch template names allow characters such as parentheses that are not valid
 // in NATS KV keys, so the name must not be used in its raw form here.
 func nameKey(accountID, name string) string {
-	return kvutil.AccountKey(accountID, "name."+hex.EncodeToString([]byte(name)))
+	return kvstore.AccountKey(accountID, "name."+hex.EncodeToString([]byte(name)))
 }
 
 // versionKey returns a version-body key: account.lt-<id>.v<n>.
 func versionKey(accountID, ltID string, n int64) string {
-	return kvutil.AccountKey(accountID, ltID+".v"+strconv.FormatInt(n, 10))
+	return kvstore.AccountKey(accountID, ltID+".v"+strconv.FormatInt(n, 10))
 }
 
 // versionPrefix returns the scan prefix for a template's version bodies.
 func versionPrefix(accountID, ltID string) string {
-	return kvutil.AccountKey(accountID, ltID+".v")
+	return kvstore.AccountKey(accountID, ltID+".v")
 }
 
 // --- record load helpers ---
 
 // getHeaderByID reads a header and its KV entry (for CAS) by launch template id.
 func (s *LaunchTemplateServiceImpl) getHeaderByID(ctx context.Context, accountID, ltID string) (*LaunchTemplateHeader, jetstream.KeyValueEntry, error) {
-	entry, err := s.kv.Get(ctx, kvutil.AccountKey(accountID, ltID))
+	entry, err := s.kv.Get(ctx, kvstore.AccountKey(accountID, ltID))
 	if err != nil {
 		if errors.Is(err, jetstream.ErrKeyNotFound) {
 			return nil, nil, errors.New(awserrors.ErrorInvalidLaunchTemplateIdNotFound)
@@ -386,7 +386,7 @@ func (s *LaunchTemplateServiceImpl) claimName(ctx context.Context, accountID, na
 	// header means the name is taken; any other read error (transient fault, or a
 	// concurrent in-flight create whose header is not yet written) fails closed so
 	// the name is never stolen from a possibly-live template.
-	_, herr := s.kv.Get(ctx, kvutil.AccountKey(accountID, string(entry.Value())))
+	_, herr := s.kv.Get(ctx, kvstore.AccountKey(accountID, string(entry.Value())))
 	switch {
 	case herr == nil:
 		return errNameInUse()
@@ -404,7 +404,7 @@ func (s *LaunchTemplateServiceImpl) putHeader(ctx context.Context, accountID str
 	if err != nil {
 		return errors.New(awserrors.ErrorServerInternal)
 	}
-	if _, err := s.kv.Put(ctx, kvutil.AccountKey(accountID, h.LaunchTemplateId), data); err != nil {
+	if _, err := s.kv.Put(ctx, kvstore.AccountKey(accountID, h.LaunchTemplateId), data); err != nil {
 		return errors.New(awserrors.ErrorServerInternal)
 	}
 	return nil
@@ -510,13 +510,13 @@ func (s *LaunchTemplateServiceImpl) commitVersion(ctx context.Context, accountID
 // re-reading on a conflict. A non-nil check vets each re-read header and
 // stops the commit with its error.
 func (s *LaunchTemplateServiceImpl) commitHeader(ctx context.Context, op, accountID, ltID string, entry jetstream.KeyValueEntry, check func(*LaunchTemplateHeader) error) error {
-	key := kvutil.AccountKey(accountID, ltID)
+	key := kvstore.AccountKey(accountID, ltID)
 	for range maxVersionRetries {
 		_, err := s.kv.Update(ctx, key, entry.Value(), entry.Revision())
 		if err == nil {
 			return nil
 		}
-		if !kvutil.IsRevisionConflict(err) {
+		if !kvstore.IsRevisionConflict(err) {
 			slog.ErrorContext(ctx, op+": header commit failed", "launchTemplateId", ltID, "err", err)
 			return err
 		}
@@ -590,7 +590,7 @@ func (s *LaunchTemplateServiceImpl) ModifyLaunchTemplate(ctx context.Context, in
 // header deleted meanwhile is NotFound for the identifier the request used.
 func (s *LaunchTemplateServiceImpl) setDefaultVersion(ctx context.Context, accountID string, input *ec2.ModifyLaunchTemplateInput, header *LaunchTemplateHeader, entry jetstream.KeyValueEntry, sel string) (*LaunchTemplateHeader, error) {
 	notFound := fmt.Sprintf("The launch template version %s is not found for the specified launch template.", sel)
-	key := kvutil.AccountKey(accountID, header.LaunchTemplateId)
+	key := kvstore.AccountKey(accountID, header.LaunchTemplateId)
 	for range maxVersionRetries {
 		n, err := s.resolveVersionNumber(ctx, accountID, header, sel)
 		if err != nil {
@@ -611,7 +611,7 @@ func (s *LaunchTemplateServiceImpl) setDefaultVersion(ctx context.Context, accou
 		if err == nil {
 			return header, nil
 		}
-		if !kvutil.IsRevisionConflict(err) {
+		if !kvstore.IsRevisionConflict(err) {
 			slog.ErrorContext(ctx, "ModifyLaunchTemplate: header update failed", "launchTemplateId", header.LaunchTemplateId, "err", err)
 			return nil, errors.New(awserrors.ErrorServerInternal)
 		}
@@ -662,7 +662,7 @@ const maxHeaderDeleteAttempts = 3
 // through the public delete; a header changed in between restarts the sweep.
 func (s *LaunchTemplateServiceImpl) deleteTemplate(ctx context.Context, accountID string, header *LaunchTemplateHeader, entry jetstream.KeyValueEntry) error {
 	ltID := header.LaunchTemplateId
-	headerKey := kvutil.AccountKey(accountID, ltID)
+	headerKey := kvstore.AccountKey(accountID, ltID)
 	for attempt := 1; ; attempt++ {
 		if err := s.deleteVersionBodies(ctx, accountID, ltID); err != nil {
 			return err
@@ -824,7 +824,7 @@ func (s *LaunchTemplateServiceImpl) commitVersionDeletes(ctx context.Context, ac
 	}
 	// A template delete that removed the header before the restore has already
 	// swept its bodies, so the restored ones would be left headerless.
-	if _, gerr := s.kv.Get(ctx, kvutil.AccountKey(accountID, ltID)); errors.Is(gerr, jetstream.ErrKeyNotFound) {
+	if _, gerr := s.kv.Get(ctx, kvstore.AccountKey(accountID, ltID)); errors.Is(gerr, jetstream.ErrKeyNotFound) {
 		for n := range deleted {
 			if derr := s.kv.Delete(ctx, versionKey(accountID, ltID, n)); derr != nil {
 				slog.ErrorContext(ctx, "DeleteLaunchTemplateVersions: failed to remove a restored version", "launchTemplateId", ltID, "version", n, "err", derr)

spinifex/domains/ec2/launchtemplate/service_impl_test.go

@@ -9,7 +9,7 @@ import (
 	"github.com/aws/aws-sdk-go/service/ec2"
 	"github.com/mulgadc/spinifex/internal/testkit"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -116,7 +116,7 @@ func TestCreateLaunchTemplate_OrphanNameReclaim(t *testing.T) {
 	lt := createTemplate(t, svc, "orphan", "t3.micro")
 
 	// Simulate a crash after the name claim but with the header lost.
-	require.NoError(t, svc.kv.Delete(t.Context(), kvutil.AccountKey(testAccountID, aws.StringValue(lt.LaunchTemplateId))))
+	require.NoError(t, svc.kv.Delete(t.Context(), kvstore.AccountKey(testAccountID, aws.StringValue(lt.LaunchTemplateId))))
 
 	// The name index still points at the now-orphaned id; create must reclaim it.
 	out, err := svc.CreateLaunchTemplate(context.Background(), &ec2.CreateLaunchTemplateInput{

spinifex/domains/ec2/launchtemplate/template_lifecycle_test.go

@@ -10,7 +10,7 @@ import (
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -38,7 +38,7 @@ func TestAccountTermination_VersionCreateRacingADeleteLeavesNoBody(t *testing.T)
 		t.Run(pause, func(t *testing.T) {
 			svc := setupTestService(t)
 			id := createWithUserData(t, svc, "web")
-			header := kvutil.AccountKey(testAccountID, id)
+			header := kvstore.AccountKey(testAccountID, id)
 			raced := false
 			race := func(key string, at string) {
 				if raced || pause != at {
@@ -71,7 +71,7 @@ func TestAccountTermination_DeleteRacingACommittedVersionRemovesIt(t *testing.T)
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	inner := svc.kv
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	raced := false
 	svc.kv = &faultyKV{KeyValue: inner, beforeDelete: func(key string) {
 		if key != header || raced {
@@ -95,7 +95,7 @@ func TestAccountTermination_FailedVersionCommitRemovesTheBody(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	before := accountKeys(t, svc.kv, testAccountID)
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	svc.kv = &faultyKV{KeyValue: svc.kv, failUpdate: func(key string) bool { return key == header }}
 
 	_, err := svc.CreateLaunchTemplateVersion(context.Background(), newVersionInput(id), testAccountID)
@@ -110,7 +110,7 @@ func TestAccountTermination_VersionCommitKeepsAConcurrentHeaderChange(t *testing
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	inner := svc.kv
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	raced := false
 	svc.kv = &faultyKV{KeyValue: inner, beforeUpdate: func(key string) {
 		if key != header || raced {
@@ -215,10 +215,10 @@ func TestAccountTemplateReferences_FailsClosed(t *testing.T) {
 			require.NoError(t, err)
 		},
 		"header unreadable": func(t *testing.T, svc *LaunchTemplateServiceImpl, id string) {
-			svc.kv = &faultyKV{KeyValue: svc.kv, failGet: func(key string) bool { return key == kvutil.AccountKey(testAccountID, id) }}
+			svc.kv = &faultyKV{KeyValue: svc.kv, failGet: func(key string) bool { return key == kvstore.AccountKey(testAccountID, id) }}
 		},
 		"header undecodable": func(t *testing.T, svc *LaunchTemplateServiceImpl, _ string) {
-			_, err := svc.kv.Put(t.Context(), kvutil.AccountKey(testAccountID, "lt-0broken000000000"), []byte("{not json"))
+			_, err := svc.kv.Put(t.Context(), kvstore.AccountKey(testAccountID, "lt-0broken000000000"), []byte("{not json"))
 			require.NoError(t, err)
 		},
 	}
@@ -241,7 +241,7 @@ func TestModifyLaunchTemplate_RacingADeleteIsNotFound(t *testing.T) {
 		t.Run(map[bool]string{false: "by-id", true: "by-name"}[byName], func(t *testing.T) {
 			svc := setupTestService(t)
 			id := createWithUserData(t, svc, "web")
-			header := kvutil.AccountKey(testAccountID, id)
+			header := kvstore.AccountKey(testAccountID, id)
 			raced := false
 			svc.kv = &faultyKV{KeyValue: svc.kv, beforeUpdate: func(key string) {
 				if key == header && !raced {
@@ -275,7 +275,7 @@ func TestModifyLaunchTemplate_RetriesAConcurrentHeaderChange(t *testing.T) {
 	_, err := svc.CreateLaunchTemplateVersion(context.Background(), newVersionInput(id), testAccountID) // v3
 	require.NoError(t, err)
 	inner := svc.kv
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	raced := false
 	svc.kv = &faultyKV{KeyValue: inner, beforeUpdate: func(key string) {
 		if key != header || raced {
@@ -306,7 +306,7 @@ func TestModifyLaunchTemplate_ConflictRetriesAreBounded(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	inner := svc.kv
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	attempts := 0
 	svc.kv = &faultyKV{KeyValue: inner, beforeUpdate: func(key string) {
 		if key != header {
@@ -333,7 +333,7 @@ func TestModifyLaunchTemplate_ConflictReReadFailureIsInternal(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	inner := svc.kv
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	conflicted := false
 	svc.kv = &faultyKV{
 		KeyValue: inner,
@@ -410,7 +410,7 @@ func TestDefaultChangeRacingAVersionDeleteFailsNotFound(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	inner := svc.kv
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	raced := false
 	svc.kv = &faultyKV{KeyValue: inner, beforeUpdate: func(key string) {
 		if key != header || raced {
@@ -460,7 +460,7 @@ func TestFailedVersionDeleteCommitRestoresTheBodies(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	before := accountKeys(t, svc.kv, testAccountID)
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	svc.kv = &faultyKV{KeyValue: svc.kv, failUpdate: func(key string) bool { return key == header }}
 
 	_, err := deleteVersions(svc, id, "2")
@@ -474,7 +474,7 @@ func TestFailedVersionDeleteCommitRestoresTheBodies(t *testing.T) {
 func TestVersionDeleteRacingATemplateDeleteLeavesNothing(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	raced := false
 	svc.kv = &faultyKV{KeyValue: svc.kv, beforeUpdate: func(key string) {
 		if key != header || raced {
@@ -531,7 +531,7 @@ func TestDefaultChangeRetriesAfterAConcurrentVersionCreate(t *testing.T) {
 	svc := setupTestService(t)
 	id := createWithUserData(t, svc, "web")
 	inner := svc.kv
-	header := kvutil.AccountKey(testAccountID, id)
+	header := kvstore.AccountKey(testAccountID, id)
 	raced := false
 	svc.kv = &faultyKV{KeyValue: inner, beforeUpdate: func(key string) {
 		if key != header || raced {

spinifex/domains/ec2/launchtemplate/template_references.go

@@ -8,7 +8,7 @@ import (
 	"strings"
 
 	"github.com/aws/aws-sdk-go/aws"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -27,7 +27,7 @@ type TemplateReferences struct {
 // header or body it cannot read or decode, is an error, because the unread
 // record may name the resource. Headerless bodies resolve from no launch.
 func (s *LaunchTemplateServiceImpl) AccountTemplateReferences(ctx context.Context, accountID string) (*TemplateReferences, error) {
-	keys, err := kvutil.Keys(ctx, s.kv)
+	keys, err := kvstore.Keys(ctx, s.kv)
 	if err != nil && !errors.Is(err, jetstream.ErrNoKeysFound) {
 		return nil, fmt.Errorf("list launch-template keys: %w", err)
 	}

spinifex/domains/ec2/natgw/lifecycle_invariants_test.go

@@ -10,7 +10,7 @@ import (
 	"github.com/aws/aws-sdk-go/service/ec2"
 	ec2eip "github.com/mulgadc/spinifex/spinifex/domains/ec2/eip"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -42,7 +42,7 @@ func TestRLC3_NatGatewayDeletesWhileRouted(t *testing.T) {
 		natgwID := createTestNatGateway(t, svc)
 
 		testutil.SeedKV(t, js, kvBucketRouteTables, map[string][]byte{
-			kvutil.AccountKey(testAccountID, "rtb-routed"): fmt.Appendf(nil,
+			kvstore.AccountKey(testAccountID, "rtb-routed"): fmt.Appendf(nil,
 				`{"route_table_id":"rtb-routed","vpc_id":"vpc-test1","routes":[{"destination_cidr_block":"0.0.0.0/0","nat_gateway_id":%q}]}`, natgwID),
 		})
 
@@ -93,7 +93,7 @@ func createTestNatGateway(t *testing.T, svc *NatGatewayServiceImpl) string {
 
 func readEIP(t *testing.T, svc *NatGatewayServiceImpl, allocID string) ec2eip.EIPRecord {
 	t.Helper()
-	entry, err := svc.eipKV.Get(t.Context(), kvutil.AccountKey(testAccountID, allocID))
+	entry, err := svc.eipKV.Get(t.Context(), kvstore.AccountKey(testAccountID, allocID))
 	require.NoError(t, err)
 	var eip ec2eip.EIPRecord
 	require.NoError(t, json.Unmarshal(entry.Value(), &eip))

spinifex/domains/ec2/natgw/service_impl.go

@@ -20,7 +20,7 @@ import (
 	awsfilters "github.com/mulgadc/spinifex/spinifex/foundation/aws/filters"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
 	awstags "github.com/mulgadc/spinifex/spinifex/foundation/aws/tags"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -45,7 +45,7 @@ func NewNatGatewayServiceImplWithNATS(ctx context.Context, natsConn *nats.Conn)
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	natgwKV, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketNatGateways, 10)
+	natgwKV, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketNatGateways, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketNatGateways, err)
 	}
@@ -61,17 +61,17 @@ func NewNatGatewayServiceImplWithNATS(ctx context.Context, natsConn *nats.Conn)
 		return nil, fmt.Errorf("migrate %s: %w", KVBucketDeletedNatGateways, err)
 	}
 
-	eipKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2eip.KVBucketEIPs, 10)
+	eipKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2eip.KVBucketEIPs, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get EIP KV bucket: %w", err)
 	}
 
-	subnetKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketSubnets, 10)
+	subnetKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketSubnets, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get subnet KV bucket: %w", err)
 	}
 
-	vpcKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
+	vpcKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get VPC KV bucket: %w", err)
 	}
@@ -92,7 +92,7 @@ func NewNatGatewayServiceImplWithNATS(ctx context.Context, natsConn *nats.Conn)
 // for as long as a client may still ask about it.
 func getOrCreateDeletedBucket(ctx context.Context, js jetstream.KeyValueManager) (jetstream.KeyValue, error) {
 	// Terraform polls DescribeNatGateways after delete and expects state=deleted.
-	return kvutil.GetOrCreateBucketWithOptions(ctx, js, kvutil.BucketOptions{
+	return kvstore.GetOrCreateBucketWithOptions(ctx, js, kvstore.BucketOptions{
 		Name:        KVBucketDeletedNatGateways,
 		Description: "Deleted NAT Gateways (auto-expire after 1 hour)",
 		History:     1,
@@ -113,7 +113,7 @@ func (s *NatGatewayServiceImpl) CreateNatGateway(ctx context.Context, input *ec2
 	allocID := *input.AllocationId
 
 	// Validate subnet exists and get its VPC
-	subnetEntry, err := s.subnetKV.Get(ctx, kvutil.AccountKey(accountID, subnetID))
+	subnetEntry, err := s.subnetKV.Get(ctx, kvstore.AccountKey(accountID, subnetID))
 	if err != nil {
 		return nil, errors.New(awserrors.ErrorInvalidSubnetIDNotFound)
 	}
@@ -123,7 +123,7 @@ func (s *NatGatewayServiceImpl) CreateNatGateway(ctx context.Context, input *ec2
 	}
 
 	// Validate EIP exists and is not already associated
-	eipEntry, err := s.eipKV.Get(ctx, kvutil.AccountKey(accountID, allocID))
+	eipEntry, err := s.eipKV.Get(ctx, kvstore.AccountKey(accountID, allocID))
 	if err != nil {
 		return nil, errors.New(awserrors.ErrorInvalidAllocationIDNotFound)
 	}
@@ -152,7 +152,7 @@ func (s *NatGatewayServiceImpl) CreateNatGateway(ctx context.Context, input *ec2
 	if err != nil {
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
-	if _, err := s.natgwKV.Put(ctx, kvutil.AccountKey(accountID, natgwID), data); err != nil {
+	if _, err := s.natgwKV.Put(ctx, kvstore.AccountKey(accountID, natgwID), data); err != nil {
 		slog.ErrorContext(ctx, "Failed to store NAT Gateway", "natGatewayId", natgwID, "err", err)
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
@@ -165,7 +165,7 @@ func (s *NatGatewayServiceImpl) CreateNatGateway(ctx context.Context, input *ec2
 	eipRecord.VpcId = subnetRecord.VpcId
 	eipRecord.State = "associated"
 	if eipData, err := json.Marshal(eipRecord); err == nil {
-		if _, err := s.eipKV.Update(ctx, kvutil.AccountKey(accountID, allocID), eipData, eipEntry.Revision()); err != nil {
+		if _, err := s.eipKV.Update(ctx, kvstore.AccountKey(accountID, allocID), eipData, eipEntry.Revision()); err != nil {
 			slog.WarnContext(ctx, "CreateNatGateway: failed to mark EIP associated", "natGatewayId", natgwID, "allocationId", allocID, "err", err)
 		}
 	}
@@ -185,7 +185,7 @@ func (s *NatGatewayServiceImpl) DeleteNatGateway(ctx context.Context, input *ec2
 	}
 
 	natgwID := *input.NatGatewayId
-	key := kvutil.AccountKey(accountID, natgwID)
+	key := kvstore.AccountKey(accountID, natgwID)
 
 	entry, err := s.natgwKV.Get(ctx, key)
 	if err != nil {
@@ -255,7 +255,7 @@ func (s *NatGatewayServiceImpl) publishDeleteEventsForNatGateway(ctx context.Con
 		slog.WarnContext(ctx, "DeleteNatGateway: JetStream setup failed, SNAT teardown deferred to route delete", "natGatewayId", record.NatGatewayId, "err", err)
 		return
 	}
-	rtbKV, err := kvutil.GetOrCreateBucket(ctx, js, kvBucketRouteTables, 10)
+	rtbKV, err := kvstore.GetOrCreateBucket(ctx, js, kvBucketRouteTables, 10)
 	if err != nil {
 		slog.WarnContext(ctx, "DeleteNatGateway: route-table scan failed, SNAT teardown deferred to route delete", "natGatewayId", record.NatGatewayId, "err", err)
 		return
@@ -266,7 +266,7 @@ func (s *NatGatewayServiceImpl) publishDeleteEventsForNatGateway(ctx context.Con
 	}
 	prefix := accountID + "."
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := rtbKV.Get(ctx, key)
@@ -302,7 +302,7 @@ func (s *NatGatewayServiceImpl) publishDeleteEventsForNatGateway(ctx context.Con
 			if assoc.SubnetId == "" {
 				continue
 			}
-			subnetEntry, err := s.subnetKV.Get(ctx, kvutil.AccountKey(accountID, assoc.SubnetId))
+			subnetEntry, err := s.subnetKV.Get(ctx, kvstore.AccountKey(accountID, assoc.SubnetId))
 			if err != nil {
 				continue
 			}
@@ -322,7 +322,7 @@ func (s *NatGatewayServiceImpl) disassociateEIP(ctx context.Context, record *Nat
 	if record.AllocationId == "" {
 		return
 	}
-	key := kvutil.AccountKey(accountID, record.AllocationId)
+	key := kvstore.AccountKey(accountID, record.AllocationId)
 	entry, err := s.eipKV.Get(ctx, key)
 	if err != nil {
 		if !errors.Is(err, jetstream.ErrKeyNotFound) {
@@ -380,7 +380,7 @@ func (s *NatGatewayServiceImpl) DescribeNatGateways(ctx context.Context, input *
 	foundIDs := make(map[string]bool)
 
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 
@@ -415,7 +415,7 @@ func (s *NatGatewayServiceImpl) DescribeNatGateways(ctx context.Context, input *
 		if foundIDs[id] {
 			continue
 		}
-		key := kvutil.AccountKey(accountID, id)
+		key := kvstore.AccountKey(accountID, id)
 		entry, err := s.deletedNatgwKV.Get(ctx, key)
 		if err != nil {
 			if errors.Is(err, jetstream.ErrKeyNotFound) {

spinifex/domains/ec2/natgw/service_impl_test.go

@@ -11,7 +11,7 @@ import (
 	ec2vpc "github.com/mulgadc/spinifex/spinifex/domains/ec2/vpc"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	awsfilters "github.com/mulgadc/spinifex/spinifex/foundation/aws/filters"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -31,18 +31,18 @@ func setupTestServiceJS(t *testing.T) (*NatGatewayServiceImpl, jetstream.JetStre
 
 	// Seed VPC KV
 	testutil.SeedKV(t, js, ec2vpc.KVBucketVPCs, map[string][]byte{
-		kvutil.AccountKey(testAccountID, "vpc-test1"): []byte(`{"vpc_id":"vpc-test1","cidr_block":"10.0.0.0/16","state":"available"}`),
+		kvstore.AccountKey(testAccountID, "vpc-test1"): []byte(`{"vpc_id":"vpc-test1","cidr_block":"10.0.0.0/16","state":"available"}`),
 	})
 
 	// Seed subnet KV (public subnet)
 	testutil.SeedKV(t, js, ec2vpc.KVBucketSubnets, map[string][]byte{
-		kvutil.AccountKey(testAccountID, "subnet-pub1"): []byte(`{"subnet_id":"subnet-pub1","vpc_id":"vpc-test1","cidr_block":"10.0.1.0/24","state":"available","map_public_ip_on_launch":true}`),
+		kvstore.AccountKey(testAccountID, "subnet-pub1"): []byte(`{"subnet_id":"subnet-pub1","vpc_id":"vpc-test1","cidr_block":"10.0.1.0/24","state":"available","map_public_ip_on_launch":true}`),
 	})
 
 	// Seed EIP KV
 	testutil.SeedKV(t, js, ec2eip.KVBucketEIPs, map[string][]byte{
-		kvutil.AccountKey(testAccountID, "eipalloc-test1"): []byte(`{"allocation_id":"eipalloc-test1","public_ip":"203.0.113.50","state":"allocated"}`),
-		kvutil.AccountKey(testAccountID, "eipalloc-used"):  []byte(`{"allocation_id":"eipalloc-used","public_ip":"203.0.113.51","state":"associated","association_id":"eipassoc-xxx"}`),
+		kvstore.AccountKey(testAccountID, "eipalloc-test1"): []byte(`{"allocation_id":"eipalloc-test1","public_ip":"203.0.113.50","state":"allocated"}`),
+		kvstore.AccountKey(testAccountID, "eipalloc-used"):  []byte(`{"allocation_id":"eipalloc-used","public_ip":"203.0.113.51","state":"associated","association_id":"eipassoc-xxx"}`),
 	})
 
 	svc, err := NewNatGatewayServiceImplWithNATS(t.Context(), nc)
@@ -170,7 +170,7 @@ func TestDeleteNatGateway(t *testing.T) {
 	assert.Equal(t, natgwID, *deleteOut.NatGatewayId)
 
 	// Should be gone
-	_, err = svc.natgwKV.Get(t.Context(), kvutil.AccountKey(testAccountID, natgwID))
+	_, err = svc.natgwKV.Get(t.Context(), kvstore.AccountKey(testAccountID, natgwID))
 	assert.ErrorIs(t, err, jetstream.ErrKeyNotFound)
 }
 

spinifex/domains/ec2/placementgroup/account_cleanup.go

@@ -8,7 +8,7 @@ import (
 	"log/slog"
 	"strings"
 
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -39,7 +39,7 @@ type AccountCleanup struct {
 // referenced returns a reason is kept; unreadable records are left for the owner scan to report. A
 // listing failure is returned and nothing is deleted.
 func (s *PlacementGroupServiceImpl) CleanupAccount(ctx context.Context, accountID string, referenced func(groupID, groupName string) string) (*AccountCleanup, error) {
-	keys, err := kvutil.Keys(ctx, s.kv)
+	keys, err := kvstore.Keys(ctx, s.kv)
 	if err != nil && !errors.Is(err, jetstream.ErrNoKeysFound) {
 		return nil, fmt.Errorf("list placement-group keys: %w", err)
 	}
@@ -47,7 +47,7 @@ func (s *PlacementGroupServiceImpl) CleanupAccount(ctx context.Context, accountI
 	prefix := accountID + "."
 	out := &AccountCleanup{Referenced: map[string]string{}}
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := s.kv.Get(ctx, key)

spinifex/domains/ec2/placementgroup/account_scan.go

@@ -8,7 +8,7 @@ import (
 	"log/slog"
 	"strings"
 
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -34,7 +34,7 @@ type AccountScan struct {
 // A listing failure is returned as an error; per-record read failures are
 // counted. The bucket's schema-version key belongs to no account.
 func (s *PlacementGroupServiceImpl) ScanAccount(ctx context.Context, accountID string) (*AccountScan, error) {
-	keys, err := kvutil.Keys(ctx, s.kv)
+	keys, err := kvstore.Keys(ctx, s.kv)
 	if err != nil && !errors.Is(err, jetstream.ErrNoKeysFound) {
 		return nil, fmt.Errorf("list placement-group keys: %w", err)
 	}
@@ -42,7 +42,7 @@ func (s *PlacementGroupServiceImpl) ScanAccount(ctx context.Context, accountID s
 	prefix := accountID + "."
 	scan := &AccountScan{}
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := s.kv.Get(ctx, key)

spinifex/domains/ec2/placementgroup/account_scan_test.go

@@ -11,7 +11,7 @@ import (
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
 	"github.com/mulgadc/spinifex/internal/testkit/readonlystore"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -42,14 +42,14 @@ func TestScanAccountClassifiesPlacementGroupRecords(t *testing.T) {
 		GroupName: "held", NodeInstances: map[string][]string{"node-2": {"i-0bbb", "i-0aaa"}, "node-3": {"i-0aaa"}},
 	}, testAccountID)
 	require.NoError(t, err)
-	_, err = svc.kv.Put(t.Context(), kvutil.AccountKey(testAccountID, "undecodable"), []byte("{not json"))
+	_, err = svc.kv.Put(t.Context(), kvstore.AccountKey(testAccountID, "undecodable"), []byte("{not json"))
 	require.NoError(t, err)
-	_, err = svc.kv.Put(t.Context(), kvutil.AccountKey(testAccountID, "anonymous"), []byte(`{"group_name":"anonymous"}`))
+	_, err = svc.kv.Put(t.Context(), kvstore.AccountKey(testAccountID, "anonymous"), []byte(`{"group_name":"anonymous"}`))
 	require.NoError(t, err)
 	createTestGroup(t, svc, "unreadable", ec2.PlacementStrategySpread)
-	require.Contains(t, bucketKeys(t, svc.kv), kvutil.VersionKey)
+	require.Contains(t, bucketKeys(t, svc.kv), kvstore.VersionKey)
 
-	svc.kv = &hookedKV{KeyValue: readonlystore.NewKV(t, svc.kv), failGet: kvutil.AccountKey(testAccountID, "unreadable")}
+	svc.kv = &hookedKV{KeyValue: readonlystore.NewKV(t, svc.kv), failGet: kvstore.AccountKey(testAccountID, "unreadable")}
 	scan, err := svc.ScanAccount(context.Background(), testAccountID)
 	require.NoError(t, err)
 

spinifex/domains/ec2/placementgroup/account_termination_current_behaviour_test.go

@@ -14,7 +14,7 @@ import (
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -84,7 +84,7 @@ func TestCurrentBehaviour_AccountTermination_PlacementGroupStorageLayout(t *test
 	}, testAccountID)
 	require.NoError(t, err)
 
-	assert.Equal(t, []string{testAccountID + ".spread-a", kvutil.VersionKey}, bucketKeys(t, svc.kv))
+	assert.Equal(t, []string{testAccountID + ".spread-a", kvstore.VersionKey}, bucketKeys(t, svc.kv))
 
 	entry, err := svc.kv.Get(t.Context(), testAccountID+".spread-a")
 	require.NoError(t, err)
@@ -100,7 +100,7 @@ func TestCurrentBehaviour_AccountTermination_PlacementGroupStorageLayout(t *test
 // a record cannot be removed through DeletePlacementGroup either, and a read
 // failure is an internal error, never an unknown group.
 func TestCurrentBehaviour_AccountTermination_DescribePlacementGroupsHidesUnreadableRecords(t *testing.T) {
-	brokenKey := kvutil.AccountKey(testAccountID, "broken")
+	brokenKey := kvstore.AccountKey(testAccountID, "broken")
 	tests := []struct {
 		name       string
 		seed       func(t *testing.T, svc *PlacementGroupServiceImpl)
@@ -139,7 +139,7 @@ func TestCurrentBehaviour_AccountTermination_DescribePlacementGroupsHidesUnreada
 func TestAccountTermination_DeletePlacementGroupLosesToAConcurrentMember(t *testing.T) {
 	svc := setupTestService(t)
 	createTestGroup(t, svc, "cluster-a", ec2.PlacementStrategyCluster)
-	key := kvutil.AccountKey(testAccountID, "cluster-a")
+	key := kvstore.AccountKey(testAccountID, "cluster-a")
 
 	launcher := &PlacementGroupServiceImpl{kv: svc.kv, region: svc.region}
 	var reserveErr, finalizeErr error
@@ -222,7 +222,7 @@ func TestAccountCleanup_DeletesEmptyGroupsAndBlocksOnMembers(t *testing.T) {
 	empty := createTestGroup(t, svc, "empty", ec2.PlacementStrategySpread)
 	held := createTestGroup(t, svc, "held", ec2.PlacementStrategyCluster)
 	require.NoError(t, addMember(t, svc, "held", "i-0123456789abcdef0"))
-	_, err := svc.kv.Put(t.Context(), kvutil.AccountKey(testAccountID, "broken"), []byte("{not json"))
+	_, err := svc.kv.Put(t.Context(), kvstore.AccountKey(testAccountID, "broken"), []byte("{not json"))
 	require.NoError(t, err)
 	other := "999988887777"
 	_, err = svc.CreatePlacementGroup(context.Background(), &ec2.CreatePlacementGroupInput{
@@ -238,7 +238,7 @@ func TestAccountCleanup_DeletesEmptyGroupsAndBlocksOnMembers(t *testing.T) {
 
 	assert.Equal(t, []string{"held"}, describedGroupNames(t, svc, testAccountID))
 	assert.Equal(t, []string{"theirs"}, describedGroupNames(t, svc, other))
-	_, err = svc.kv.Get(t.Context(), kvutil.AccountKey(testAccountID, "broken"))
+	_, err = svc.kv.Get(t.Context(), kvstore.AccountKey(testAccountID, "broken"))
 	require.NoError(t, err, "an unreadable record is never deleted")
 }
 
@@ -247,7 +247,7 @@ func TestAccountCleanup_DeletesEmptyGroupsAndBlocksOnMembers(t *testing.T) {
 func TestAccountCleanup_ConcurrentMemberBlocksTheDelete(t *testing.T) {
 	svc := setupTestService(t)
 	group := createTestGroup(t, svc, "cluster-a", ec2.PlacementStrategyCluster)
-	key := kvutil.AccountKey(testAccountID, "cluster-a")
+	key := kvstore.AccountKey(testAccountID, "cluster-a")
 	launcher := &PlacementGroupServiceImpl{kv: svc.kv, region: svc.region}
 	var memberErr error
 	// Read 1 is the cleanup's listing read; read 2 is the delete's check.

spinifex/domains/ec2/placementgroup/service_impl.go

@@ -17,7 +17,7 @@ import (
 	awsfilters "github.com/mulgadc/spinifex/spinifex/foundation/aws/filters"
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
 	awstags "github.com/mulgadc/spinifex/spinifex/foundation/aws/tags"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -87,7 +87,7 @@ func NewPlacementGroupServiceImplWithNATS(ctx context.Context, cfg *config.Confi
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	kv, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketPlacementGroups, 10)
+	kv, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketPlacementGroups, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketPlacementGroups, err)
 	}
@@ -137,7 +137,7 @@ func (s *PlacementGroupServiceImpl) CreatePlacementGroup(ctx context.Context, in
 	}
 
 	groupName := *input.GroupName
-	key := kvutil.AccountKey(accountID, groupName)
+	key := kvstore.AccountKey(accountID, groupName)
 	groupID := awsidentifiers.GenerateResourceID("pg")
 
 	record := PlacementGroupRecord{
@@ -175,7 +175,7 @@ func (s *PlacementGroupServiceImpl) DeletePlacementGroup(ctx context.Context, in
 	}
 
 	groupName := *input.GroupName
-	if _, err := s.deleteGroup(ctx, accountID, kvutil.AccountKey(accountID, groupName)); err != nil {
+	if _, err := s.deleteGroup(ctx, accountID, kvstore.AccountKey(accountID, groupName)); err != nil {
 		if errors.Is(err, errPGAbsent) {
 			// Lower case, unlike DescribePlacementGroups' message for the same code.
 			return nil, awserrors.Errorf(awserrors.ErrorInvalidPlacementGroupUnknown, "The placement group '%s' is unknown.", groupName)
@@ -194,7 +194,7 @@ func (s *PlacementGroupServiceImpl) DeletePlacementGroup(ctx context.Context, in
 // The returned count is the members seen by the last check.
 func (s *PlacementGroupServiceImpl) deleteGroup(ctx context.Context, accountID, key string) (int, error) {
 	members := 0
-	record, err := kvutil.DeleteIf(ctx, s.kv, key, casConfig(), func(record *PlacementGroupRecord) error {
+	record, err := kvstore.DeleteIf(ctx, s.kv, key, casConfig(), func(record *PlacementGroupRecord) error {
 		members = memberCount(record.NodeInstances)
 		if members > 0 {
 			return errPGInUse
@@ -238,7 +238,7 @@ func (s *PlacementGroupServiceImpl) DescribePlacementGroups(ctx context.Context,
 
 	var groups []*ec2.PlacementGroup
 	for _, k := range keys {
-		if k == kvutil.VersionKey {
+		if k == kvstore.VersionKey {
 			continue
 		}
 		if !strings.HasPrefix(k, prefix) {
@@ -341,7 +341,7 @@ func pgMatchesFilters(record *PlacementGroupRecord, filters map[string][]string)
 
 const maxCASRetries = 5
 
-// errPGAbsent and errPGContended mark the two kvutil outcomes that map to
+// errPGAbsent and errPGContended mark the two kvstore outcomes that map to
 // specific AWS errors rather than a generic internal failure, and neither
 // escapes this file. errPGInUse is the AWS error itself.
 var (
@@ -350,8 +350,8 @@ var (
 	errPGInUse     = errors.New(awserrors.ErrorInvalidPlacementGroupInUse)
 )
 
-// mapCASError turns a kvutil failure into the AWS error the caller returns.
-// A non-kvutil error came from mutate and is already an AWS code.
+// mapCASError turns a kvstore failure into the AWS error the caller returns.
+// A non-kvstore error came from mutate and is already an AWS code.
 func mapCASError(ctx context.Context, op, groupName, accountID string, err error) error {
 	switch {
 	case errors.Is(err, errPGAbsent):
@@ -359,16 +359,16 @@ func mapCASError(ctx context.Context, op, groupName, accountID string, err error
 	case errors.Is(err, errPGContended):
 		slog.ErrorContext(ctx, "Placement group CAS retries exhausted under contention",
 			"op", op, "groupName", groupName, "accountID", accountID)
-	case errors.Is(err, kvutil.ErrRead):
+	case errors.Is(err, kvstore.ErrRead):
 		slog.ErrorContext(ctx, "Failed to read placement group from KV",
 			"op", op, "groupName", groupName, "accountID", accountID, "err", err)
-	case errors.Is(err, kvutil.ErrDecode):
+	case errors.Is(err, kvstore.ErrDecode):
 		slog.ErrorContext(ctx, "Corrupt placement group record in KV",
 			"op", op, "groupName", groupName, "accountID", accountID, "err", err)
-	case errors.Is(err, kvutil.ErrEncode):
+	case errors.Is(err, kvstore.ErrEncode):
 		slog.ErrorContext(ctx, "Failed to marshal placement group record",
 			"op", op, "groupName", groupName, "accountID", accountID, "err", err)
-	case errors.Is(err, kvutil.ErrWrite):
+	case errors.Is(err, kvstore.ErrWrite):
 		slog.ErrorContext(ctx, "Failed to write placement group to KV",
 			"op", op, "groupName", groupName, "accountID", accountID, "err", err)
 	default:
@@ -378,8 +378,8 @@ func mapCASError(ctx context.Context, op, groupName, accountID string, err error
 }
 
 // casConfig is the shared CAS policy for every placement group mutation.
-func casConfig() kvutil.CASConfig {
-	return kvutil.CASConfig{
+func casConfig() kvstore.CASConfig {
+	return kvstore.CASConfig{
 		Attempts:  maxCASRetries,
 		NotFound:  errPGAbsent,
 		Exhausted: func(string, int) error { return errPGContended },
@@ -394,7 +394,7 @@ func (s *PlacementGroupServiceImpl) ReserveSpreadNodes(ctx context.Context, inpu
 	}
 
 	var selected []string
-	_, err := kvutil.Update(ctx, s.kv, kvutil.AccountKey(accountID, input.GroupName), casConfig(),
+	_, err := kvstore.Update(ctx, s.kv, kvstore.AccountKey(accountID, input.GroupName), casConfig(),
 		func(record *PlacementGroupRecord) (bool, error) {
 			if record.State != ec2.PlacementGroupStateAvailable {
 				return false, errors.New(awserrors.ErrorInvalidPlacementGroupUnknown)
@@ -437,7 +437,7 @@ func (s *PlacementGroupServiceImpl) FinalizeSpreadInstances(ctx context.Context,
 		return nil, errors.New(awserrors.ErrorMissingParameter)
 	}
 
-	_, err := kvutil.Update(ctx, s.kv, kvutil.AccountKey(accountID, input.GroupName), casConfig(),
+	_, err := kvstore.Update(ctx, s.kv, kvstore.AccountKey(accountID, input.GroupName), casConfig(),
 		func(record *PlacementGroupRecord) (bool, error) {
 			maps.Copy(record.NodeInstances, input.NodeInstances)
 			return true, nil
@@ -462,7 +462,7 @@ func (s *PlacementGroupServiceImpl) ReleaseSpreadNodes(ctx context.Context, inpu
 		releaseSet[n] = true
 	}
 
-	_, err := kvutil.Update(ctx, s.kv, kvutil.AccountKey(accountID, input.GroupName), casConfig(),
+	_, err := kvstore.Update(ctx, s.kv, kvstore.AccountKey(accountID, input.GroupName), casConfig(),
 		func(record *PlacementGroupRecord) (bool, error) {
 			changed := false
 			for node := range releaseSet {
@@ -490,7 +490,7 @@ func (s *PlacementGroupServiceImpl) RemoveInstance(ctx context.Context, input *R
 		return nil, errors.New(awserrors.ErrorMissingParameter)
 	}
 
-	_, err := kvutil.Update(ctx, s.kv, kvutil.AccountKey(accountID, input.GroupName), casConfig(),
+	_, err := kvstore.Update(ctx, s.kv, kvstore.AccountKey(accountID, input.GroupName), casConfig(),
 		func(record *PlacementGroupRecord) (bool, error) {
 			instances, exists := record.NodeInstances[input.NodeName]
 			if !exists {
@@ -533,7 +533,7 @@ func (s *PlacementGroupServiceImpl) ReserveClusterNode(ctx context.Context, inpu
 	}
 
 	var targetNode string
-	_, err := kvutil.Update(ctx, s.kv, kvutil.AccountKey(accountID, input.GroupName), casConfig(),
+	_, err := kvstore.Update(ctx, s.kv, kvstore.AccountKey(accountID, input.GroupName), casConfig(),
 		func(record *PlacementGroupRecord) (bool, error) {
 			if record.State != ec2.PlacementGroupStateAvailable {
 				return false, errors.New(awserrors.ErrorInvalidPlacementGroupUnknown)
@@ -570,7 +570,7 @@ func (s *PlacementGroupServiceImpl) FinalizeClusterInstances(ctx context.Context
 		return nil, errors.New(awserrors.ErrorMissingParameter)
 	}
 
-	_, err := kvutil.Update(ctx, s.kv, kvutil.AccountKey(accountID, input.GroupName), casConfig(),
+	_, err := kvstore.Update(ctx, s.kv, kvstore.AccountKey(accountID, input.GroupName), casConfig(),
 		func(record *PlacementGroupRecord) (bool, error) {
 			// Append rather than replace: a cluster group can have several launches in flight at once.
 			for node, ids := range input.NodeInstances {

spinifex/domains/ec2/placementgroup/service_impl_test.go

@@ -10,7 +10,7 @@ import (
 	"github.com/mulgadc/spinifex/internal/testkit"
 	"github.com/mulgadc/spinifex/spinifex/bootstrap/config"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -39,7 +39,7 @@ func createTestGroup(t *testing.T, svc *PlacementGroupServiceImpl, name, strateg
 // readRecord returns the stored record for assertions.
 func readRecord(t *testing.T, svc *PlacementGroupServiceImpl, groupName string) *PlacementGroupRecord {
 	t.Helper()
-	entry, err := svc.kv.Get(t.Context(), kvutil.AccountKey(testAccountID, groupName))
+	entry, err := svc.kv.Get(t.Context(), kvstore.AccountKey(testAccountID, groupName))
 	require.NoError(t, err)
 	var record PlacementGroupRecord
 	require.NoError(t, json.Unmarshal(entry.Value(), &record))
@@ -50,7 +50,7 @@ func readRecord(t *testing.T, svc *PlacementGroupServiceImpl, groupName string)
 // starting state the public API cannot produce.
 func seedRecord(t *testing.T, svc *PlacementGroupServiceImpl, groupName string, mutate func(*PlacementGroupRecord)) {
 	t.Helper()
-	_, err := kvutil.Update(t.Context(), svc.kv, kvutil.AccountKey(testAccountID, groupName), kvutil.CASConfig{},
+	_, err := kvstore.Update(t.Context(), svc.kv, kvstore.AccountKey(testAccountID, groupName), kvstore.CASConfig{},
 		func(record *PlacementGroupRecord) (bool, error) {
 			mutate(record)
 			return true, nil

spinifex/domains/ec2/routetable/service_impl.go

@@ -25,7 +25,7 @@ import (
 	awsidentifiers "github.com/mulgadc/spinifex/spinifex/foundation/aws/identifiers"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/paging"
 	awstags "github.com/mulgadc/spinifex/spinifex/foundation/aws/tags"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -52,7 +52,7 @@ func NewRouteTableServiceImplWithNATS(ctx context.Context, cfg *config.Config, n
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	rtbKV, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketRouteTables, 10)
+	rtbKV, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketRouteTables, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketRouteTables, err)
 	}
@@ -60,22 +60,22 @@ func NewRouteTableServiceImplWithNATS(ctx context.Context, cfg *config.Config, n
 		return nil, fmt.Errorf("migrate %s: %w", KVBucketRouteTables, err)
 	}
 
-	vpcKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
+	vpcKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketVPCs, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get VPC KV bucket: %w", err)
 	}
 
-	igwKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2igw.KVBucketIGW, 10)
+	igwKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2igw.KVBucketIGW, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get IGW KV bucket: %w", err)
 	}
 
-	subnetKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketSubnets, 10)
+	subnetKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2vpc.KVBucketSubnets, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get subnet KV bucket: %w", err)
 	}
 
-	natgwKV, err := kvutil.GetOrCreateBucket(ctx, js, ec2natgw.KVBucketNatGateways, 10)
+	natgwKV, err := kvstore.GetOrCreateBucket(ctx, js, ec2natgw.KVBucketNatGateways, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to get NAT Gateway KV bucket: %w", err)
 	}
@@ -95,7 +95,7 @@ func NewRouteTableServiceImplWithNATS(ctx context.Context, cfg *config.Config, n
 
 // getRouteTable retrieves a route table record from KV.
 func (s *RouteTableServiceImpl) getRouteTable(ctx context.Context, accountID, rtbID string) (*RouteTableRecord, error) {
-	entry, err := s.rtbKV.Get(ctx, kvutil.AccountKey(accountID, rtbID))
+	entry, err := s.rtbKV.Get(ctx, kvstore.AccountKey(accountID, rtbID))
 	if err != nil {
 		if errors.Is(err, jetstream.ErrKeyNotFound) {
 			return nil, rtbNotFoundError(rtbID)
@@ -118,7 +118,7 @@ func (s *RouteTableServiceImpl) putRouteTable(ctx context.Context, accountID str
 		slog.ErrorContext(ctx, "Failed to marshal route table record", "routeTableId", record.RouteTableId, "err", err)
 		return errors.New(awserrors.ErrorServerInternal)
 	}
-	if _, err := s.rtbKV.Put(ctx, kvutil.AccountKey(accountID, record.RouteTableId), data); err != nil {
+	if _, err := s.rtbKV.Put(ctx, kvstore.AccountKey(accountID, record.RouteTableId), data); err != nil {
 		slog.ErrorContext(ctx, "Failed to write route table to KV", "routeTableId", record.RouteTableId, "err", err)
 		return errors.New(awserrors.ErrorServerInternal)
 	}
@@ -130,7 +130,7 @@ func (s *RouteTableServiceImpl) putRouteTable(ctx context.Context, accountID str
 // parallel AssociateRouteTable calls Terraform fans out per route table.
 const rtbCASMaxRetries = 16
 
-// errRTBAbsent and errRTBContended mark the two kvutil outcomes that map to
+// errRTBAbsent and errRTBContended mark the two kvstore outcomes that map to
 // specific AWS errors rather than a generic internal failure. Neither
 // escapes this file.
 var (
@@ -143,7 +143,7 @@ var (
 // several subnets with one route table at once. mutate reports whether
 // it changed the record; a false return commits nothing.
 func (s *RouteTableServiceImpl) mutateRouteTableCAS(ctx context.Context, accountID, rtbID string, mutate func(*RouteTableRecord) (bool, error)) error {
-	_, err := kvutil.Update(ctx, s.rtbKV, kvutil.AccountKey(accountID, rtbID), kvutil.CASConfig{
+	_, err := kvstore.Update(ctx, s.rtbKV, kvstore.AccountKey(accountID, rtbID), kvstore.CASConfig{
 		Attempts:  rtbCASMaxRetries,
 		NotFound:  errRTBAbsent,
 		Exhausted: func(string, int) error { return errRTBContended },
@@ -157,13 +157,13 @@ func (s *RouteTableServiceImpl) mutateRouteTableCAS(ctx context.Context, account
 		// Contended rather than broken.
 		slog.ErrorContext(ctx, "Route table CAS retries exhausted under contention",
 			"routeTableId", rtbID, "attempts", rtbCASMaxRetries)
-	case errors.Is(err, kvutil.ErrRead):
+	case errors.Is(err, kvstore.ErrRead):
 		slog.ErrorContext(ctx, "Failed to read route table from KV", "routeTableId", rtbID, "err", err)
-	case errors.Is(err, kvutil.ErrDecode):
+	case errors.Is(err, kvstore.ErrDecode):
 		slog.ErrorContext(ctx, "Corrupt route table record in KV", "routeTableId", rtbID, "err", err)
-	case errors.Is(err, kvutil.ErrEncode):
+	case errors.Is(err, kvstore.ErrEncode):
 		slog.ErrorContext(ctx, "Failed to marshal route table record", "routeTableId", rtbID, "err", err)
-	case errors.Is(err, kvutil.ErrWrite):
+	case errors.Is(err, kvstore.ErrWrite):
 		slog.ErrorContext(ctx, "Failed to write route table to KV", "routeTableId", rtbID, "err", err)
 	default:
 		return err
@@ -173,7 +173,7 @@ func (s *RouteTableServiceImpl) mutateRouteTableCAS(ctx context.Context, account
 
 // getVPCCidr looks up a VPC's CIDR block from the VPC KV bucket.
 func (s *RouteTableServiceImpl) getVPCCidr(ctx context.Context, accountID, vpcID string) (string, error) {
-	entry, err := s.vpcKV.Get(ctx, kvutil.AccountKey(accountID, vpcID))
+	entry, err := s.vpcKV.Get(ctx, kvstore.AccountKey(accountID, vpcID))
 	if err != nil {
 		return "", awserrors.IDNotFound(awserrors.ErrorInvalidVpcIDNotFound, "vpc", vpcID)
 	}
@@ -229,7 +229,7 @@ func (s *RouteTableServiceImpl) subnetsImplicitlyOnMainRT(ctx context.Context, a
 	}
 	var implicit []string
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := s.subnetKV.Get(ctx, key)
@@ -387,7 +387,7 @@ func (s *RouteTableServiceImpl) allSubnetsForVPC(ctx context.Context, accountID,
 	}
 	var subnets []string
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := s.subnetKV.Get(ctx, key)
@@ -455,7 +455,7 @@ func (s *RouteTableServiceImpl) allRouteTablesForVPC(ctx context.Context, accoun
 
 	var results []RouteTableRecord
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := s.rtbKV.Get(ctx, key)
@@ -577,7 +577,7 @@ func (s *RouteTableServiceImpl) DeleteRouteTable(ctx context.Context, input *ec2
 		}
 	}
 
-	if err := s.rtbKV.Delete(ctx, kvutil.AccountKey(accountID, rtbID)); err != nil {
+	if err := s.rtbKV.Delete(ctx, kvstore.AccountKey(accountID, rtbID)); err != nil {
 		return nil, errors.New(awserrors.ErrorServerInternal)
 	}
 
@@ -633,7 +633,7 @@ func (s *RouteTableServiceImpl) DescribeRouteTables(ctx context.Context, input *
 	foundIDs := make(map[string]bool)
 
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 
@@ -740,7 +740,7 @@ func (s *RouteTableServiceImpl) CreateRoute(ctx context.Context, input *ec2.Crea
 	case input.GatewayId != nil && *input.GatewayId != "":
 		igwID := *input.GatewayId
 		// Verify IGW exists and is attached to the same VPC
-		igwEntry, err := s.igwKV.Get(ctx, kvutil.AccountKey(accountID, igwID))
+		igwEntry, err := s.igwKV.Get(ctx, kvstore.AccountKey(accountID, igwID))
 		if err != nil {
 			return nil, gatewayNotFoundError(igwID)
 		}
@@ -770,7 +770,7 @@ func (s *RouteTableServiceImpl) CreateRoute(ctx context.Context, input *ec2.Crea
 	case input.NatGatewayId != nil && *input.NatGatewayId != "":
 		natgwID := *input.NatGatewayId
 		// Verify NAT GW exists and belongs to the same VPC
-		natgwEntry, err := s.natgwKV.Get(ctx, kvutil.AccountKey(accountID, natgwID))
+		natgwEntry, err := s.natgwKV.Get(ctx, kvstore.AccountKey(accountID, natgwID))
 		if err != nil {
 			return nil, errors.New(awserrors.ErrorInvalidNatGatewayIDNotFound)
 		}
@@ -924,7 +924,7 @@ func (s *RouteTableServiceImpl) ReplaceRoute(ctx context.Context, input *ec2.Rep
 	igwID := *input.GatewayId
 
 	// AWS reports a missing gateway ahead of a missing route.
-	igwEntry, err := s.igwKV.Get(ctx, kvutil.AccountKey(accountID, igwID))
+	igwEntry, err := s.igwKV.Get(ctx, kvstore.AccountKey(accountID, igwID))
 	if err != nil {
 		return nil, gatewayNotFoundError(igwID)
 	}
@@ -983,7 +983,7 @@ func (s *RouteTableServiceImpl) AssociateRouteTable(ctx context.Context, input *
 	}
 
 	// Verify subnet exists and belongs to the same VPC
-	subnetEntry, err := s.subnetKV.Get(ctx, kvutil.AccountKey(accountID, subnetID))
+	subnetEntry, err := s.subnetKV.Get(ctx, kvstore.AccountKey(accountID, subnetID))
 	if err != nil {
 		return nil, errors.New(awserrors.ErrorInvalidSubnetIDNotFound)
 	}
@@ -1077,7 +1077,7 @@ func (s *RouteTableServiceImpl) DisassociateRouteTable(ctx context.Context, inpu
 	}
 
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 
@@ -1168,7 +1168,7 @@ func (s *RouteTableServiceImpl) ReplaceRouteTableAssociation(ctx context.Context
 	}
 
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 
@@ -1369,7 +1369,7 @@ func (s *RouteTableServiceImpl) publishNatGatewayDeleteEvents(ctx context.Contex
 	if s.natsConn == nil {
 		return
 	}
-	natgwEntry, err := s.natgwKV.Get(ctx, kvutil.AccountKey(accountID, natgwID))
+	natgwEntry, err := s.natgwKV.Get(ctx, kvstore.AccountKey(accountID, natgwID))
 	if err != nil {
 		slog.WarnContext(ctx, "NAT GW event: natgw lookup failed", "topic", networkv1.NATGatewayDeleteSubject, "natGatewayId", natgwID, "err", err)
 		return
@@ -1417,7 +1417,7 @@ func (s *RouteTableServiceImpl) publishNatGatewayEventsForAssociation(ctx contex
 		if r.NatGatewayId == "" {
 			continue
 		}
-		natgwEntry, err := s.natgwKV.Get(ctx, kvutil.AccountKey(accountID, r.NatGatewayId))
+		natgwEntry, err := s.natgwKV.Get(ctx, kvstore.AccountKey(accountID, r.NatGatewayId))
 		if err != nil {
 			slog.WarnContext(ctx, "NAT GW event: natgw lookup failed", "topic", topic, "natGatewayId", r.NatGatewayId, "err", err)
 			continue
@@ -1442,7 +1442,7 @@ func (s *RouteTableServiceImpl) publishNatGatewayEventForSubnet(ctx context.Cont
 	if s.natsConn == nil {
 		return
 	}
-	subnetEntry, err := s.subnetKV.Get(ctx, kvutil.AccountKey(accountID, subnetID))
+	subnetEntry, err := s.subnetKV.Get(ctx, kvstore.AccountKey(accountID, subnetID))
 	if err != nil {
 		slog.WarnContext(ctx, "NAT GW event: subnet lookup failed", "topic", topic, "subnetId", subnetID, "err", err)
 		return

spinifex/domains/ec2/routetable/service_impl_test.go

@@ -13,7 +13,7 @@ import (
 	ec2natgw "github.com/mulgadc/spinifex/spinifex/domains/ec2/natgw"
 	ec2vpc "github.com/mulgadc/spinifex/spinifex/domains/ec2/vpc"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -28,23 +28,23 @@ func setupTestService(t *testing.T) *RouteTableServiceImpl {
 
 	// Seed VPC KV
 	testutil.SeedKV(t, js, ec2vpc.KVBucketVPCs, map[string][]byte{
-		kvutil.AccountKey(testAccountID, "vpc-test1"): []byte(`{"vpc_id":"vpc-test1","cidr_block":"10.0.0.0/16","state":"available"}`),
+		kvstore.AccountKey(testAccountID, "vpc-test1"): []byte(`{"vpc_id":"vpc-test1","cidr_block":"10.0.0.0/16","state":"available"}`),
 	})
 
 	// Seed IGW KV (attached to vpc-test1)
 	testutil.SeedKV(t, js, ec2igw.KVBucketIGW, map[string][]byte{
-		kvutil.AccountKey(testAccountID, "igw-test1"): []byte(`{"internet_gateway_id":"igw-test1","vpc_id":"vpc-test1","state":"available"}`),
+		kvstore.AccountKey(testAccountID, "igw-test1"): []byte(`{"internet_gateway_id":"igw-test1","vpc_id":"vpc-test1","state":"available"}`),
 	})
 
 	// Seed subnet KV
 	testutil.SeedKV(t, js, ec2vpc.KVBucketSubnets, map[string][]byte{
-		kvutil.AccountKey(testAccountID, "subnet-test1"): []byte(`{"subnet_id":"subnet-test1","vpc_id":"vpc-test1","cidr_block":"10.0.1.0/24","state":"available"}`),
-		kvutil.AccountKey(testAccountID, "subnet-priv1"): []byte(`{"subnet_id":"subnet-priv1","vpc_id":"vpc-test1","cidr_block":"10.0.11.0/24","state":"available"}`),
+		kvstore.AccountKey(testAccountID, "subnet-test1"): []byte(`{"subnet_id":"subnet-test1","vpc_id":"vpc-test1","cidr_block":"10.0.1.0/24","state":"available"}`),
+		kvstore.AccountKey(testAccountID, "subnet-priv1"): []byte(`{"subnet_id":"subnet-priv1","vpc_id":"vpc-test1","cidr_block":"10.0.11.0/24","state":"available"}`),
 	})
 
 	// Seed NAT Gateway KV
 	testutil.SeedKV(t, js, ec2natgw.KVBucketNatGateways, map[string][]byte{
-		kvutil.AccountKey(testAccountID, "nat-test1"): []byte(`{"nat_gateway_id":"nat-test1","vpc_id":"vpc-test1","subnet_id":"subnet-test1","public_ip":"192.168.1.100","state":"available"}`),
+		kvstore.AccountKey(testAccountID, "nat-test1"): []byte(`{"nat_gateway_id":"nat-test1","vpc_id":"vpc-test1","subnet_id":"subnet-test1","public_ip":"192.168.1.100","state":"available"}`),
 	})
 
 	svc, err := NewRouteTableServiceImplWithNATS(t.Context(), nil, nc)

spinifex/domains/ec2/spotinstance/account_scan.go

@@ -10,7 +10,7 @@ import (
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/nats-io/nats.go/jetstream"
 
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 )
 
 // AccountSpotRequest is one persisted request in an account scan.
@@ -54,7 +54,7 @@ func scanBucket(ctx context.Context, kv jetstream.KeyValue, accountID string) ([
 	prefix := accountID + "."
 	var out []AccountSpotRequest
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		entry, err := kv.Get(ctx, key)

spinifex/domains/ec2/spotinstance/account_scan_test.go

@@ -11,7 +11,7 @@ import (
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 )
 
 func TestScanAccountListsActiveAndTerminalForOneAccount(t *testing.T) {
@@ -35,7 +35,7 @@ func TestScanAccountListsActiveAndTerminalForOneAccount(t *testing.T) {
 func TestScanAccountFailsOnAnUndecodableRecord(t *testing.T) {
 	svc := setupTestService(t)
 	putSIRs(t, svc, makeSIR("sir-aaa", "i-aaa"))
-	_, err := svc.activeKV.Put(t.Context(), kvutil.AccountKey(testAccountID, "sir-bad"), []byte("{not json"))
+	_, err := svc.activeKV.Put(t.Context(), kvstore.AccountKey(testAccountID, "sir-bad"), []byte("{not json"))
 	require.NoError(t, err)
 
 	scan, err := svc.ScanAccount(t.Context(), testAccountID)
@@ -45,7 +45,7 @@ func TestScanAccountFailsOnAnUndecodableRecord(t *testing.T) {
 
 func TestScanAccountFailsOnARecordWithoutARequest(t *testing.T) {
 	svc := setupTestService(t)
-	_, err := svc.terminalKV.Put(t.Context(), kvutil.AccountKey(testAccountID, "sir-empty"), []byte(`{"account_id":"x"}`))
+	_, err := svc.terminalKV.Put(t.Context(), kvstore.AccountKey(testAccountID, "sir-empty"), []byte(`{"account_id":"x"}`))
 	require.NoError(t, err)
 
 	_, err = svc.ScanAccount(t.Context(), testAccountID)
@@ -76,7 +76,7 @@ func (k failingGetKV) Get(ctx context.Context, key string) (jetstream.KeyValueEn
 func TestScanAccountFailsWhenARecordCannotBeRead(t *testing.T) {
 	svc := setupTestService(t)
 	putSIRs(t, svc, makeSIR("sir-aaa", "i-aaa"), makeSIR("sir-bbb", "i-bbb"))
-	svc.activeKV = failingGetKV{KeyValue: svc.activeKV, key: kvutil.AccountKey(testAccountID, "sir-bbb")}
+	svc.activeKV = failingGetKV{KeyValue: svc.activeKV, key: kvstore.AccountKey(testAccountID, "sir-bbb")}
 
 	scan, err := svc.ScanAccount(t.Context(), testAccountID)
 	require.Error(t, err)

spinifex/domains/ec2/spotinstance/launch_references.go

@@ -8,7 +8,7 @@ import (
 
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/nats-io/nats.go/jetstream"
 )
 
@@ -24,14 +24,14 @@ type LaunchReferences struct {
 // account. It is complete or an error: an unlistable bucket or an unreadable
 // request is an error, because the unread request may name the resource.
 func (s *SpotInstanceServiceImpl) AccountLaunchReferences(ctx context.Context, accountID string) (*LaunchReferences, error) {
-	keys, err := kvutil.Keys(ctx, s.activeKV)
+	keys, err := kvstore.Keys(ctx, s.activeKV)
 	if err != nil && !errors.Is(err, jetstream.ErrNoKeysFound) {
 		return nil, fmt.Errorf("list spot requests: %w", err)
 	}
 	prefix := accountID + "."
 	out := &LaunchReferences{}
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		record, err := s.readRecord(ctx, s.activeKV, key)

spinifex/domains/ec2/spotinstance/launch_references_test.go

@@ -5,7 +5,7 @@ import (
 
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/ec2"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -41,7 +41,7 @@ func TestAccountLaunchReferences_ReadsOpenAndActiveRequests(t *testing.T) {
 func TestAccountLaunchReferences_UnreadableRequestFailsTheCheck(t *testing.T) {
 	svc := setupTestService(t)
 	putSIRs(t, svc, makeSIR("sir-active01", "i-0000000000000001"))
-	_, err := svc.activeKV.Put(t.Context(), kvutil.AccountKey(testAccountID, "sir-broken01"), []byte("{not json"))
+	_, err := svc.activeKV.Put(t.Context(), kvstore.AccountKey(testAccountID, "sir-broken01"), []byte("{not json"))
 	require.NoError(t, err)
 
 	_, err = svc.AccountLaunchReferences(t.Context(), testAccountID)

spinifex/domains/ec2/spotinstance/service_impl.go

@@ -15,7 +15,7 @@ import (
 	"github.com/mulgadc/spinifex/spinifex/bootstrap/config"
 	"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
 	awsfilters "github.com/mulgadc/spinifex/spinifex/foundation/aws/filters"
-	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvutil"
+	"github.com/mulgadc/spinifex/spinifex/foundation/state/kvstore"
 	"github.com/mulgadc/spinifex/spinifex/foundation/state/migrate"
 	"github.com/nats-io/nats.go"
 	"github.com/nats-io/nats.go/jetstream"
@@ -41,7 +41,7 @@ func NewSpotInstanceServiceImplWithNATS(ctx context.Context, cfg *config.Config,
 		return nil, fmt.Errorf("failed to get JetStream context: %w", err)
 	}
 
-	activeKV, err := kvutil.GetOrCreateBucket(ctx, js, KVBucketSpotRequests, 10)
+	activeKV, err := kvstore.GetOrCreateBucket(ctx, js, KVBucketSpotRequests, 10)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create KV bucket %s: %w", KVBucketSpotRequests, err)
 	}
@@ -71,7 +71,7 @@ func NewSpotInstanceServiceImplWithNATS(ctx context.Context, cfg *config.Config,
 // getOrCreateTerminalBucket opens the terminal bucket without changing its
 // existing replica or placement configuration, creating it only when absent.
 func getOrCreateTerminalBucket(ctx context.Context, js jetstream.KeyValueManager) (jetstream.KeyValue, error) {
-	return kvutil.GetOrCreateBucketWithOptions(ctx, js, kvutil.BucketOptions{
+	return kvstore.GetOrCreateBucketWithOptions(ctx, js, kvstore.BucketOptions{
 		Name:        KVBucketSpotRequestsTerminal,
 		Description: "Terminal Spot Instance Requests (auto-expire after 1 hour)",
 		History:     1,
@@ -91,7 +91,7 @@ func (s *SpotInstanceServiceImpl) PutSpotInstanceRequests(ctx context.Context, i
 		if err != nil {
 			return nil, errors.New(awserrors.ErrorServerInternal)
 		}
-		if _, err := s.activeKV.Put(ctx, kvutil.AccountKey(accountID, sirID), data); err != nil {
+		if _, err := s.activeKV.Put(ctx, kvstore.AccountKey(accountID, sirID), data); err != nil {
 			slog.ErrorContext(ctx, "PutSpotInstanceRequests: KV put failed", "sirId", sirID, "err", err)
 			return nil, errors.New(awserrors.ErrorServerInternal)
 		}
@@ -171,7 +171,7 @@ func (s *SpotInstanceServiceImpl) CancelSpotInstanceRequests(ctx context.Context
 	cancelled := make([]*ec2.CancelledSpotInstanceRequest, 0, len(input.SpotInstanceRequestIds))
 	for _, idPtr := range input.SpotInstanceRequestIds {
 		sirID := aws.StringValue(idPtr)
-		key := kvutil.AccountKey(accountID, sirID)
+		key := kvstore.AccountKey(accountID, sirID)
 
 		record, err := s.readRecord(ctx, s.activeKV, key)
 		if err == nil {
@@ -213,7 +213,7 @@ func (s *SpotInstanceServiceImpl) CloseForInstance(ctx context.Context, instance
 
 	prefix := accountID + "."
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		record, err := s.readRecord(ctx, s.activeKV, key)
@@ -251,7 +251,7 @@ func (s *SpotInstanceServiceImpl) listRequests(ctx context.Context, kv jetstream
 	prefix := accountID + "."
 	var requests []*ec2.SpotInstanceRequest
 	for _, key := range keys {
-		if key == kvutil.VersionKey || !strings.HasPrefix(key, prefix) {
+		if key == kvstore.VersionKey || !strings.HasPrefix(key, prefix) {
 			continue
 		}
 		record, err := s.readRecord(ctx, kv, key)

spinifex/domains/ec2/systeminstance/launcher.go

@@ -57,7 +57,7 @@ const (
 	BootAMI BootMode = "ami"
 )
 
-// SystemInstanceLauncher launches, reboots and terminates system-managed VMs.
+// SystemInstanceLauncher launches and terminates system-managed VMs.
 // Defined here to avoid a circular import between elbv2/eks and daemon.
 type SystemInstanceLauncher interface {
 	// LaunchSystemInstance creates and starts a system-managed VM, returning its ID and private IP.
@@ -65,10 +65,6 @@ type SystemInstanceLauncher interface {
 
 	// TerminateSystemInstance stops and cleans up a system-managed VM.
 	TerminateSystemInstance(instanceID string) error
-
-	// RebootSystemInstance restarts a system-managed load balancer VM into a
-	// fresh QEMU on its owning node, keeping its ID, ENIs and addresses.
-	RebootSystemInstance(instanceID string) error
 }
 
 // SystemInstanceInput describes a system VM to launch.