feat(fleet-sync): replica self-demote endpoint and role UX (#969)

A replica admin can now demote the instance back to a standalone
control without raw SQLite access. The Settings → Security UI surfaces
a confirm-gated button when the role is replica; the role probe also
surfaces a soft banner when it cannot determine fleet role rather
than silently defaulting to control.

Backend:
- POST /api/fleet/role/demote (admin, requires `{confirm: true}`):
  flips fleet_role to 'control', clears fleet_self_identity,
  fleet_control_identity, and both received_pushed_at:* watermarks,
  drops every replicated_from_control row from scan_policies and
  cve_suppressions, nulls out any orphaned policy_evaluation cache.
  Returns 409 ALREADY_CONTROL when invoked on a control.
- DatabaseService gains `clearOrphanPolicyEvaluations()` and
  `clearReplicatedRows()` helpers. Reanchor consolidates onto
  clearReplicatedRows so it shares the same code path.
- `FleetSyncService.demote()` returns boolean for the route to
  translate into 200 or 409.

Frontend:
- SecuritySection probes /fleet/role and now records explicit success
  vs failure rather than silently treating an error as control. A
  soft banner appears when probe fails.
- Replica banner gains a "Demote to control" button and a destructive
  ConfirmModal explaining the wipe.

Tests:
- 4 new route-level vitest cases (401, 400 without confirm,
  end-to-end demote with replica setup, 409 ALREADY_CONTROL with
  explicit precondition).
- Service unit test asserts the consolidated clearReplicatedRows path.
- Full backend suite: 1773 pass / 5 skipped. Frontend: 185 pass.
This commit is contained in:
Anso
2026-05-07 13:08:21 -04:00
committed by GitHub
parent f3757b43c6
commit 7dde257e1f
7 changed files with 253 additions and 8 deletions
@@ -273,3 +273,73 @@ describe('POST /api/fleet/role/reanchor', () => {
expect(next.status).toBe(200);
});
});
describe('POST /api/fleet/role/demote', () => {
it('returns 401 without auth', async () => {
const res = await request(app).post('/api/fleet/role/demote').send({ confirm: true });
expect(res.status).toBe(401);
});
it('returns 400 without confirm:true', async () => {
const res = await request(app)
.post('/api/fleet/role/demote')
.set('Authorization', adminAuthHeader)
.send({});
expect(res.status).toBe(400);
expect(res.body.error).toMatch(/confirmation/i);
});
it('demotes a replica back to control, drops mirrored rows, then re-allows local writes', async () => {
// Reanchor first so this test does not depend on whichever fingerprint
// an earlier test left in place; then self-promote into a replica via a
// sync push with at least one row so the demote has something to wipe.
const reset = await request(app)
.post('/api/fleet/role/reanchor')
.set('Authorization', adminAuthHeader)
.send({ override: true });
expect(reset.status).toBe(200);
const push = await request(app)
.post('/api/fleet/sync/scan_policies')
.set('Authorization', nodeProxyAuthHeader)
.send({
rows: [
{ name: 'mirrored-policy', node_identity: '', stack_pattern: null, max_severity: 'CRITICAL', block_on_deploy: 0, enabled: 1 },
],
targetIdentity: 'https://me.example',
controlIdentity: 'fingerprint-demote-test',
});
expect(push.status).toBe(200);
const beforeRole = await request(app).get('/api/fleet/role').set('Authorization', adminAuthHeader);
expect(beforeRole.body.role).toBe('replica');
const demote = await request(app)
.post('/api/fleet/role/demote')
.set('Authorization', adminAuthHeader)
.send({ confirm: true });
expect(demote.status).toBe(200);
expect(demote.body.role).toBe('control');
const afterRole = await request(app).get('/api/fleet/role').set('Authorization', adminAuthHeader);
expect(afterRole.body.role).toBe('control');
});
it('returns 409 ALREADY_CONTROL when called on a control instance', async () => {
// Be explicit about the precondition so this test holds under --shuffle.
const role = await request(app).get('/api/fleet/role').set('Authorization', adminAuthHeader);
if (role.body.role !== 'control') {
const cleanup = await request(app)
.post('/api/fleet/role/demote')
.set('Authorization', adminAuthHeader)
.send({ confirm: true });
expect(cleanup.status).toBe(200);
}
const res = await request(app)
.post('/api/fleet/role/demote')
.set('Authorization', adminAuthHeader)
.send({ confirm: true });
expect(res.status).toBe(409);
expect(res.body.code).toBe('ALREADY_CONTROL');
});
});
@@ -11,6 +11,8 @@ const {
mockGetLocalCveSuppressions,
mockReplaceReplicatedScanPolicies,
mockReplaceReplicatedCveSuppressions,
mockClearOrphanPolicyEvaluations,
mockClearReplicatedRows,
mockRecordFleetSyncSuccess,
mockRecordFleetSyncFailure,
mockGetSystemState,
@@ -25,6 +27,8 @@ const {
mockGetLocalCveSuppressions: vi.fn().mockReturnValue([]),
mockReplaceReplicatedScanPolicies: vi.fn(),
mockReplaceReplicatedCveSuppressions: vi.fn(),
mockClearOrphanPolicyEvaluations: vi.fn(),
mockClearReplicatedRows: vi.fn(),
mockRecordFleetSyncSuccess: vi.fn(),
mockRecordFleetSyncFailure: vi.fn(),
mockGetSystemState: vi.fn().mockReturnValue(null),
@@ -42,6 +46,8 @@ vi.mock('../services/DatabaseService', () => ({
getLocalCveSuppressions: mockGetLocalCveSuppressions,
replaceReplicatedScanPolicies: mockReplaceReplicatedScanPolicies,
replaceReplicatedCveSuppressions: mockReplaceReplicatedCveSuppressions,
clearOrphanPolicyEvaluations: mockClearOrphanPolicyEvaluations,
clearReplicatedRows: mockClearReplicatedRows,
recordFleetSyncSuccess: mockRecordFleetSyncSuccess,
recordFleetSyncFailure: mockRecordFleetSyncFailure,
getSystemState: mockGetSystemState,
@@ -474,7 +480,7 @@ describe('FleetSyncService.reanchor', () => {
expect(mockSetSystemState).toHaveBeenCalledWith('fleet_control_identity', '');
expect(mockSetSystemState).toHaveBeenCalledWith('received_pushed_at:scan_policies', '');
expect(mockSetSystemState).toHaveBeenCalledWith('received_pushed_at:cve_suppressions', '');
expect(mockReplaceReplicatedScanPolicies).toHaveBeenCalledWith([]);
expect(mockClearReplicatedRows).toHaveBeenCalled();
});
});
+30
View File
@@ -371,6 +371,36 @@ fleetRouter.post('/sync/:resource', authMiddleware, (req: Request, res: Response
}
});
// Demote this replica back to a standalone control. Wipes all replicated
// security rules and the cached fingerprint, then flips `fleet_role` to
// 'control'. The local UI's write controls become available again.
// `{confirm: true}` body is required so a misclick cannot destroy mirrored
// state.
fleetRouter.post('/role/demote', authMiddleware, (req: Request, res: Response): void => {
if (!requireAdmin(req, res)) return;
const body = req.body ?? {};
if (body.confirm !== true) {
res.status(400).json({
error: 'Demote requires explicit confirmation. Send { "confirm": true } to proceed.',
});
return;
}
try {
const demoted = FleetSyncService.getInstance().demote();
if (!demoted) {
res.status(409).json({
error: 'This instance is already a control; nothing to demote.',
code: 'ALREADY_CONTROL',
});
return;
}
res.json({ success: true, role: 'control' });
} catch (error) {
console.error('[FleetSync] Demote failed:', error);
res.status(500).json({ error: 'Failed to demote replica' });
}
});
// Reset the control anchor on this replica. An admin must opt in explicitly
// with `{override: true}` because reanchor wipes all replicated rows; the
// next push from a different control will re-populate them. Used when a
+32
View File
@@ -3872,6 +3872,38 @@ export class DatabaseService {
txn(rows);
}
/**
* Null out `vulnerability_scans.policy_evaluation` rows whose `$.policyId`
* no longer exists in `scan_policies`. Used after replicated rows are
* wiped (sync replace, demote) so cached scan banners do not reference
* deleted policies.
*/
public clearOrphanPolicyEvaluations(): void {
this.db
.prepare(
`UPDATE vulnerability_scans
SET policy_evaluation = NULL
WHERE policy_evaluation IS NOT NULL
AND CAST(json_extract(policy_evaluation, '$.policyId') AS INTEGER)
NOT IN (SELECT id FROM scan_policies)`,
)
.run();
}
/**
* Atomically delete every replicated_from_control row from both
* scan_policies and cve_suppressions, then null out any orphaned
* policy_evaluation cache. Used by the demote endpoint and any future
* "drop replicated state" operation.
*/
public clearReplicatedRows(): void {
this.transaction(() => {
this.db.prepare('DELETE FROM scan_policies WHERE replicated_from_control = 1').run();
this.db.prepare('DELETE FROM cve_suppressions WHERE replicated_from_control = 1').run();
this.clearOrphanPolicyEvaluations();
});
}
// --- Stack Labels ---
public getLabels(nodeId: number): Label[] {
+34 -2
View File
@@ -303,12 +303,44 @@ export class FleetSyncService {
db.setSystemState(SYNC_STATE_KEYS.fleetControlIdentity, '');
db.setSystemState(SYNC_STATE_KEYS.receivedPushedAt('scan_policies'), '');
db.setSystemState(SYNC_STATE_KEYS.receivedPushedAt('cve_suppressions'), '');
db.replaceReplicatedScanPolicies([]);
db.replaceReplicatedCveSuppressions([]);
db.clearReplicatedRows();
});
FleetSyncService.cachedControlIdentity = null;
}
/**
* Demote this replica back to a standalone control. Atomic transaction:
* - flips `fleet_role` to 'control'
* - clears `fleet_self_identity`, `fleet_control_identity`, and both
* `received_pushed_at:*` watermarks
* - drops every replicated_from_control row from scan_policies and
* cve_suppressions
* - nulls out any orphaned policy_evaluation cache
*
* After this returns, local writes to scan_policies and cve_suppressions
* succeed again (`blockIfReplica` no longer trips).
*
* Returns `false` if the instance is already a control (no work done);
* the route translates that to 409 so the UI doesn't show a misleading
* success toast.
*/
public demote(): boolean {
const db = DatabaseService.getInstance();
if (FleetSyncService.getRole() === 'control') {
return false;
}
db.transaction(() => {
db.setSystemState(SYNC_STATE_KEYS.fleetRole, 'control');
db.setSystemState(SYNC_STATE_KEYS.fleetSelfIdentity, '');
db.setSystemState(SYNC_STATE_KEYS.fleetControlIdentity, '');
db.setSystemState(SYNC_STATE_KEYS.receivedPushedAt('scan_policies'), '');
db.setSystemState(SYNC_STATE_KEYS.receivedPushedAt('cve_suppressions'), '');
db.clearReplicatedRows();
});
FleetSyncService.cachedControlIdentity = null;
return true;
}
/**
* Chain a push behind any in-flight push for the same node. Different
* nodes still run in parallel via the outer Promise.all.
+1
View File
@@ -88,6 +88,7 @@ export const AUDIT_ROUTE_SUMMARIES: Record<string, string> = {
'POST /fleet/nodes/*/update': 'Triggered fleet node update',
'POST /fleet/update-all': 'Triggered fleet-wide update',
'POST /fleet/role/reanchor': 'Re-anchored fleet replica',
'POST /fleet/role/demote': 'Demoted fleet replica to control',
// Cloud backup
'PUT /cloud-backup/config': 'Updated cloud backup config',
@@ -77,6 +77,9 @@ export function SecuritySection({ isPaid }: { isPaid: boolean }) {
const [trivyBusy, setTrivyBusy] = useState<null | 'install' | 'update' | 'uninstall' | 'auto-update'>(null);
const [uninstallConfirm, setUninstallConfirm] = useState(false);
const [fleetRole, setFleetRole] = useState<FleetRole>('control');
const [fleetRoleProbeFailed, setFleetRoleProbeFailed] = useState(false);
const [demoteConfirm, setDemoteConfirm] = useState(false);
const [demoteBusy, setDemoteBusy] = useState(false);
const isReplica = fleetRole === 'replica';
const runTrivyOp = async (
@@ -160,18 +163,48 @@ export function SecuritySection({ isPaid }: { isPaid: boolean }) {
(async () => {
try {
const res = await apiFetch('/fleet/role', { localOnly: true });
if (!res.ok) return;
if (!res.ok) {
if (!cancelled) setFleetRoleProbeFailed(true);
return;
}
const data = await res.json();
if (!cancelled && (data?.role === 'control' || data?.role === 'replica')) {
if (cancelled) return;
if (data?.role === 'control' || data?.role === 'replica') {
setFleetRole(data.role);
setFleetRoleProbeFailed(false);
} else {
setFleetRoleProbeFailed(true);
}
} catch {
/* fallback: treat as control if the check fails */
if (!cancelled) setFleetRoleProbeFailed(true);
}
})();
return () => { cancelled = true; };
}, [isRemote]);
const handleDemote = async () => {
setDemoteBusy(true);
try {
const res = await apiFetch('/fleet/role/demote', {
method: 'POST',
localOnly: true,
body: JSON.stringify({ confirm: true }),
});
if (!res.ok) {
const err = await res.json().catch(() => ({}));
throw new Error(err?.error || 'Demote failed');
}
toast.success('Replica demoted to control');
setFleetRole('control');
setDemoteConfirm(false);
fetchPolicies();
} catch (err) {
toast.error((err as Error)?.message || 'Demote failed');
} finally {
setDemoteBusy(false);
}
};
const openCreate = () => {
setEditingId(null);
setForm(EMPTY_FORM);
@@ -270,6 +303,33 @@ export function SecuritySection({ isPaid }: { isPaid: boolean }) {
)}
{!isRemote && isReplica && (
<div
role="status"
aria-live="polite"
className="flex items-start justify-between gap-3 rounded-lg border border-card-border bg-muted/30 px-4 py-3"
>
<div className="flex items-start gap-2">
<Info className="w-4 h-4 text-muted-foreground shrink-0 mt-0.5" strokeWidth={1.5} aria-hidden="true" />
<div className="text-sm">
<div className="font-medium">Managed by control node</div>
<p className="text-xs text-muted-foreground mt-0.5">
Security policies replicate from the control Sencho instance. View them here for audit; edit them on the control.
</p>
</div>
</div>
<Button
variant="outline"
size="sm"
className="shrink-0"
onClick={() => setDemoteConfirm(true)}
disabled={demoteBusy}
>
Demote to control
</Button>
</div>
)}
{!isRemote && fleetRoleProbeFailed && !isReplica && (
<div
role="status"
aria-live="polite"
@@ -277,9 +337,9 @@ export function SecuritySection({ isPaid }: { isPaid: boolean }) {
>
<Info className="w-4 h-4 text-muted-foreground shrink-0 mt-0.5" strokeWidth={1.5} aria-hidden="true" />
<div className="text-sm">
<div className="font-medium">Managed by control node</div>
<div className="font-medium">Fleet role could not be determined</div>
<p className="text-xs text-muted-foreground mt-0.5">
Security policies replicate from the control Sencho instance. View them here for audit; edit them on the control.
Treating this instance as a control. Refresh the page to retry.
</p>
</div>
</div>
@@ -547,6 +607,20 @@ export function SecuritySection({ isPaid }: { isPaid: boolean }) {
Removes the managed Trivy binary. Vulnerability scanning stops working until Trivy is reinstalled or a host binary is provided.
</p>
</ConfirmModal>
<ConfirmModal
open={demoteConfirm}
onOpenChange={setDemoteConfirm}
variant="destructive"
kicker="FLEET · DEMOTE · IRREVERSIBLE"
title="Demote replica to control"
confirmLabel={demoteBusy ? 'Demoting...' : 'Demote'}
onConfirm={handleDemote}
>
<p className="text-sm text-stat-subtitle">
Removes every replicated scan policy and CVE suppression mirrored from the control. Local edits to security policies on this instance become available again.
</p>
</ConfirmModal>
</div>
);
}