fix(recovery): let preservation owners reopen retained sources (#17053)

## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Recovery must save retained workspace files before ordinary execution can resume. > - The environment runtime blocks access to retained sources to prevent unsafe reuse or deletion. > - The recovery worker also used that ordinary resume path. > - The guard returned no lease, so recovery reported an invalid source binding before it could save files. > - This pull request gives the physical preservation owner a separate, source-bound resume path. > - Ordinary execution and destructive cleanup keep their existing guards. ## Linked Issues or Issue Description Refs #16965. **What happened?** A retained workspace with a workspaceRestoreRecovery marker could not recover through the real environment runtime. The runtime returned null before calling the provider. Recovery classified this as source_binding_invalid. Earlier recovery fixtures replaced the runtime facade, so they did not exercise this guard. **Expected behavior** The verified preservation owner can reopen its exact stopped allocation to save files. Other callers remain blocked. Recovery preserves newer local files and does not start another agent or repeat a completed external action. **Steps to reproduce** 1. Seed a terminal source run with a retained sandbox, a recovery action, and the workspaceRestoreRecovery lease marker. 2. Give the source a verified stop receipt. 3. Call recoverRetainedWorkspace through environmentRuntimeService. 4. Before this change, the recovery action returns to blocked with source_binding_invalid. No files are copied. **Paperclip version or commit** Reproduced on master at 65f59e339a. The same failure was observed in a served build containing #16965. ## What Changed - Add an internal preservation resume method. It validates the exact company, task, run, action, lease, environment, current physical owner, and stop-only cleanup claim before provider access. - Keep ordinary resume and destructive cleanup blocked for protected sources. - Run all retained-workspace fixtures through the assembled environment runtime. Add a failing-before/passing-after protected-source regression and four rejection cases. - Use the same runtime path and protection marker in the opt-in Daytona interruption probe. Allow a verified immutable snapshot UUID as its input. - Document the authority boundary and verification recipes. ## Verification - Red: the protected-source regression failed with expected recovered, received blocked. It recorded source_binding_invalid before provider resume. - Green: 235 tests passed across workspace recovery, durable task Retry, and environment runtime. These include saved files, newer local files, bounded retries, competing owners, and no extra run or wake. - pnpm -r typecheck and pnpm build passed. The initial sandboxed typecheck could not open a tsx IPC pipe; the same command passed with local process access. - node scripts/check-maps.mjs and git diff --check passed. - Live Daytona qualification passed using the same immutable snapshot ID as the affected deployment. Export was interrupted after a real download; a second controller restored the files and verified one original agent start, one simulated external action, one saved snapshot, and a stopped retained source. The disposable sandbox was deleted after verification. - Greptile reviewed 804dfe24ef at 5/5 with no actionable findings or unresolved review threads. - Full current-commit CI passed: https://github.com/paperclipai/paperclip/actions/runs/39540771747. Required ci / verify and ci / e2e passed. Superagent, Socket, and Snyk checks passed. - Full local pnpm test:run did not pass: 22,992 tests passed, one managed-run identity test timed out at 60 seconds, and 100 tests skipped across the first general-server phase. The command stopped before its later phases. The unchanged agent-identity.test.ts suite then passed all 11 tests in isolation in 15.53 seconds. The original timeout remains recorded; it is not reclassified as a full local pass. The same commit's complete CI test matrix passed. - The optional Paperclip Review completed with a neutral advisory 5/5 on the same head and no actionable findings. It inspected all eight changed files and connected code. It did not independently run tests or the provider probe. - Hosted ordinary Retry recorded one durable intent and no successor run while file preservation remained blocked. The existing published review stayed unchanged. Hosted recovery remains unverified until this follow-up is deployed. A Retry acknowledgement alone does not prove recovery. ## Risks - Recovery can reopen a retained sandbox. The dedicated path requires the current physical ownership claim and the exact source's stop-only intent. It does not clear the ordinary protection marker. - Provider failures retain the source and use the existing bounded recovery budget. The provider cannot substitute another allocation. - The live probe uses a disposable sandbox and simulated agent/external-effect counters. It does not prove a live LLM or GitHub review journey. - No schema, public API, permission, budget, pause, or approval changes. ## Model Used OpenAI GPT-6 through Codex, with repository inspection, code execution, browser verification, and test tools. The exact served model variant and context window are not exposed by this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass (focused suites; full suite still running) - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-10-24 01:41:42 +02:00 · 2026-10-21 07:27:45 -05:00
1 parent 30b753f614
commit eaacadeba3
@@ -15,6 +15,7 @@ import { NOOP_PLUGIN_TRACER } from "@paperclipai/plugin-sdk";
 import { startEmbeddedPostgresTestDatabase } from "../src/__tests__/helpers/embedded-postgres.js";
 import { readSystemRecovery } from "../src/services/system-recovery-state.js";
 import { recoverRetainedWorkspace } from "../src/services/workspace-recovery.js";
+import { classifyEnvironmentCapabilities, environmentRuntimeService } from "../src/services/environment-runtime.js";
 import { attachPreservedWorkspaceSnapshots } from "../src/services/workspace-recovery-snapshot.js";
 import { performSyncOut } from "../../packages/plugins/sandbox-providers/daytona/src/file-sync.js";

@@ -68,7 +69,11 @@ it.skipIf(!live || !worker)("provider controller", async () => {
        },
        syncOut: operations => performSyncOut({ sandbox, operations, remoteDir: f.remoteCwd, timeoutSeconds: 120 }),
      } };
-    const environmentRuntime = {
+    const driver = {
+      driver: "sandbox",
+      acquireRunLease: async () => { throw new Error("Recovery must not acquire another sandbox"); },
+      releaseRunLease: async () => { throw new Error("Recovery must not destroy its retained source"); },
+      resolveCapabilities: async () => classifyEnvironmentCapabilities({ verifiedMethods: [] }),
      resumeRunLease: async () => {
        await sandbox.start(120); await sandbox.refreshData();
        if (sandbox.state !== "started") throw new Error("Provider did not confirm start");
@@ -80,6 +85,7 @@ it.skipIf(!live || !worker)("provider controller", async () => {
        return { providerLeaseId: sandbox.id, state: "stopped" };
      },
    };
+    const environmentRuntime = environmentRuntimeService(db, { drivers: [driver] });
    await recoverRetainedWorkspace({ db, ...f, environmentRuntime });
    const [action] = await db.select().from(issueRecoveryActions).where(eq(issueRecoveryActions.id, f.actionId));
    const state = readSystemRecovery(action.evidence);
@@ -90,7 +96,9 @@ it.skipIf(!live || !worker)("provider controller", async () => {

 it.skipIf(!live || Boolean(worker))("preserves provider files across an interrupted export and controller restart without restarting the old agent", async () => {
  const image = process.env.PAPERCLIP_RECOVERY_IMAGE;
-  if (!image?.includes("@sha256:")) throw new Error("An immutable qualification image is required");
+  const snapshotId = process.env.PAPERCLIP_RECOVERY_SNAPSHOT_ID;
+  if (!image?.includes("@sha256:") && !/^[0-9a-f-]{36}$/i.test(snapshotId ?? ""))
+    throw new Error("An immutable qualification image or snapshot ID is required");
  const directory = await mkdtemp(path.join(os.tmpdir(), "paperclip-provider-recovery-"));
  const priorHome = process.env.PAPERCLIP_HOME;
  process.env.PAPERCLIP_HOME = path.join(directory, "home");
@@ -100,11 +108,16 @@ it.skipIf(!live || Boolean(worker))("preserves provider files across an interrup
  const db = createDb(database.connectionString);
  const daytona = await client();
  let sandbox;
-  const evidence = { image, interrupted: false, restored: false, oldAgentStarts: null, externalOperations: null,
+  const evidence = { image, snapshotId, interrupted: false, restored: false, oldAgentStarts: null, externalOperations: null,
    prematureCleanup: false, controllerProcesses: [], snapshotCount: 0 };
  try {
-    sandbox = await daytona.create({ image, autoStopInterval: 10, autoArchiveInterval: 60,
-      resources: { cpu: 2, memory: 4, disk: 10 }, labels: { purpose: "paperclip-recovery-qualification" } }, { timeout: 300 });
+    if (snapshotId) {
+      const snapshot = await daytona.snapshot.get(snapshotId);
+      expect(snapshot.id).toBe(snapshotId);
+      expect(snapshot.state).toBe("active");
+    }
+    sandbox = await daytona.create({ ...(snapshotId ? { snapshot: snapshotId } : { image, resources: { cpu: 2, memory: 4, disk: 10 } }),
+      autoStopInterval: 10, autoArchiveInterval: 60, labels: { purpose: "paperclip-recovery-qualification" } }, { timeout: 300 });
    const remoteCwd = "/tmp/paperclip-recovery-fixture";
    await sandbox.process.executeCommand(`mkdir -p ${remoteCwd}`, undefined, undefined, 30);
    await sandbox.process.createSession("old-agent");
@@ -124,7 +137,8 @@ it.skipIf(!live || Boolean(worker))("preserves provider files across an interrup
    const [lease] = await db.insert(environmentLeases).values({ companyId, issueId, heartbeatRunId: runId,
      environmentId: environment.id, provider: "daytona", providerLeaseId: sandbox.id,
      leasePolicy: "retain_on_failure", status: "released", releasedAt: new Date(),
-      metadata: { remoteCwd, driver: "sandbox", pluginId: randomUUID(), sandboxProviderPlugin: true } }).returning();
+      metadata: { remoteCwd, driver: "sandbox", pluginId: randomUUID(), sandboxProviderPlugin: true,
+        workspaceRestoreRecovery: { schema: "paperclip.workspace-restore-recovery.v1", runId } } }).returning();
    const recovery = { schema: "paperclip.workspace-restore-recovery.v1", leaseIds: [lease.id] };
    await db.update(heartbeatRuns).set({ resultJson: { workspaceRestoreRecovery: recovery, workspaceRestoreFailure: "restore_failed" } })
      .where(eq(heartbeatRuns.id, runId));
@@ -4,6 +4,8 @@ import { allowLegacyShutdownWorkspaceCleanup, legacyShutdownWorkspaceResourcePro
 import { nativeWorkspaceExportPendingForLease } from "./native-runtime/native-workspace-finalization-state.js";
 import { beginIdleTrackedWork } from "./task-admission.js";
 import { historicalProviderCredentials } from "./historical-provider-credentials.js";
+import { ownedWorkspaceRecoveryLease, type WorkspaceRecoverySourceIdentity } from "./workspace-recovery-source.js";
+import type { NativeWorkspaceFinalizationOwnership } from "./native-runtime/native-workspace-finalization-ownership.js";
 import { hasStopOnlyCleanup, prepareSandboxStopAndRetain, readStopOnlyCleanup, settleStopOnlyCleanup, stopOnlyCleanupKey } from "./sandbox-stop-and-retain.js";
 import { JsonRpcCallError, readEnvironmentAcquisitionDiagnostic, readEnvironmentCreationCleanupError } from "@paperclipai/plugin-sdk";
 import { preserveEnvironmentSyncOutErrorDiagnostic } from "./environment-sync-out-error.js";
@@ -4569,6 +4571,19 @@ export function environmentRuntimeService(
      return await driver.resumeRunLease(input);
    },

+    async resumeWorkspaceRecoveryLease(input: {
+      source: WorkspaceRecoverySourceIdentity; leaseId: string; environmentId: string;
+      ownership: NativeWorkspaceFinalizationOwnership;
+    }): Promise<PluginEnvironmentLease | EnvironmentLease | null> {
+      const { lease, environment } = await ownedWorkspaceRecoveryLease(db, { ...input.source,
+        leaseId: input.leaseId, environmentId: input.environmentId, ownership: input.ownership });
+      const driver = requireDriverKey(getLeaseDriverKey(lease, environment as Environment));
+      if (!driver.resumeRunLease) throw new Error(`Environment driver "${driver.driver}" does not support lease resume.`);
+      // Maintenance reopens the protected source only to preserve its files.
+      // Ordinary resume and destructive cleanup keep their existing fences.
+      return await driver.resumeRunLease({ environment: environment as Environment, lease: lease as EnvironmentLease });
+    },
+
    async destroyRunLease(input: EnvironmentDriverLeaseInput): Promise<EnvironmentLease | null> {
      if (await nativeWorkspaceExportPendingForLease(db, input.lease)) return null;
      if (await legacyShutdownWorkspaceResourceProtected(db, input.lease)) return null;
@@ -3,11 +3,31 @@ import { environmentLeases, environments, heartbeatRuns, issueRecoveryActions, n
 import { adapterExecutionControls } from "./adapter-execution-control.js";
 import { readNativeWorkspaceSyncReference } from "./native-runtime/native-workspace-sync.js";
 import { SystemRecoveryFailure } from "./system-recovery-state.js";
+import type { NativeWorkspaceFinalizationOwnership } from "./native-runtime/native-workspace-finalization-ownership.js";
+import { readStopOnlyCleanup } from "./sandbox-stop-and-retain.js";

 export const workspaceExportCauses = ["native_workspace_sync_out_unrecoverable", "native_workspace_sync_out_retry_exhausted"];
 export const terminalWorkspaceRunStates = ["failed", "interrupted", "timed_out", "cancelled", "succeeded"];
 export type WorkspaceRecoverySourceIdentity = { companyId: string; issueId: string; runId: string; actionId: string };

+/** The preservation owner may reopen only its exact stop-only allocation.
+ * This does not release the source's ordinary execution or destruction fence. */
+export async function ownedWorkspaceRecoveryLease(db: Db, input: WorkspaceRecoverySourceIdentity & {
+  leaseId: string; environmentId: string; ownership: NativeWorkspaceFinalizationOwnership;
+}) {
+  await input.ownership.assertHeld();
+  const inspected = await inspectWorkspaceRecoverySource(db, input);
+  const source = inspected.sources.find(item => item.lease.id === input.leaseId);
+  if (!source || source.environment.id !== input.environmentId || source.lease.status !== "pending_cleanup"
+    || source.lease.metadata?.pendingCleanupInFlight !== true
+    || source.lease.metadata?.pendingCleanupAttemptId !== input.ownership.token
+    || !readStopOnlyCleanup(source.lease)) {
+    throw new SystemRecoveryFailure("source_binding_invalid", "service_required");
+  }
+  await input.ownership.assertHeld();
+  return source;
+}
+
 /** Resolve only the immutable source recorded by the failed run. Never discover
 * a candidate by recency, mutable adapter configuration, or a provider search. */
 export async function inspectWorkspaceRecoverySource(db: Db, input: WorkspaceRecoverySourceIdentity) {
@@ -1,6 +1,7 @@
 import { prepareHeartbeatWorkspace } from "./heartbeat/workspace-preparation.js";
 import { createHeartbeatWorkspaceResolver, resolveExecutionWorkspaceReuseRequestForIssue } from "./heartbeat/workspaces.js";
 import { environmentService } from "./environments.js";
+import { classifyEnvironmentCapabilities, environmentRuntimeService } from "./environment-runtime.js";
 import { environmentRunOrchestrator } from "./environment-run-orchestrator.js";
 import { executionWorkspaceService } from "./execution-workspaces.js";
 import { workspaceOperationService } from "./workspace-operations.js";
@@ -111,7 +112,11 @@ describe("automatic retained workspace recovery", () => {
      fingerprint: runId, nextAction: "Manual repair required", evidence: { runId, workspaceRestoreFailure: "restore_failed", workspaceRestoreRecovery: recovery } });
    const runtime = { resumeRunLease: vi.fn(async (_input: { lease: EnvironmentLease }) => ({ providerLeaseId, metadata: { remoteCwd: remote } })),
      retryPendingSandboxTeardown: vi.fn(async () => ({ providerLeaseId, state: "stopped" })) };
-    const input = { db, companyId, issueId, runId, actionId, environmentRuntime: runtime as never };
+    const environmentRuntime = environmentRuntimeService(db, { drivers: [{ driver: "sandbox",
+      acquireRunLease: vi.fn(), releaseRunLease: vi.fn(), ...runtime,
+      resolveCapabilities: async () => classifyEnvironmentCapabilities({ verifiedMethods: [], declared: null, narrowing: null }),
+    }] });
+    const input = { db, companyId, issueId, runId, actionId, environmentRuntime };
    return { ...input, input, runtime, host, remote, lease, providerLeaseId };
  }
  const action = async (id: string) => (await db.select().from(issueRecoveryActions).where(eq(issueRecoveryActions.id, id)))[0]!;
@@ -283,6 +288,52 @@ describe("automatic retained workspace recovery", () => {
    expect(await db.select().from(workspaceOperations).where(and(eq(workspaceOperations.companyId, f.companyId), eq(workspaceOperations.phase, "workspace_recovery")))).toHaveLength(1);
  }, 30_000);

+  it("recovers a protected retained source through the assembled environment runtime", async () => {
+    const f = await seed();
+    await db.update(environmentLeases).set({ metadata: { ...f.lease.metadata,
+      workspaceRestoreRecovery: { schema: "paperclip.workspace-restore-recovery.v1", runId: f.runId },
+      remoteExecutionTermination: remoteTerminationReceipt(f.lease, { providerLeaseId: f.providerLeaseId, state: "stopped" }),
+    } }).where(eq(environmentLeases.id, f.lease.id));
+    const runtime = f.input.environmentRuntime;
+    const [lease] = await db.select().from(environmentLeases).where(eq(environmentLeases.id, f.lease.id));
+    const [environment] = await db.select().from(environments).where(eq(environments.id, lease.environmentId!));
+    expect(await runtime.resumeRunLease({ environment: environment as never, lease: lease as EnvironmentLease })).toBeNull();
+    expect(f.runtime.resumeRunLease).not.toHaveBeenCalled();
+    await recoverRetainedWorkspace({ ...f.input, environmentRuntime: runtime });
+    expect(readWorkspaceRecovery((await action(f.actionId)).evidence)?.status).toBe("recovered");
+    expect(await readFile(path.join(f.host, "saved.txt"), "utf8")).toBe("finished work\n");
+    expect(await readFile(path.join(f.host, "newer.txt"), "utf8")).toBe("newer local work\n");
+    expect(f.runtime.resumeRunLease).toHaveBeenCalledOnce();
+    await recoverRetainedWorkspace({ ...f.input, environmentRuntime: runtime });
+    expect(f.runtime.resumeRunLease).toHaveBeenCalledOnce();
+    expect(await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.companyId, f.companyId))).toHaveLength(1);
+    expect(await db.select().from(agentWakeupRequests).where(eq(agentWakeupRequests.companyId, f.companyId))).toHaveLength(0);
+  }, 30_000);
+
+  it.each(["changed_claim", "missing_stop_intent", "wrong_lease", "wrong_environment"])(
+    "does not reopen a retained source without exact recovery authority: %s", async kind => {
+      const f = await seed();
+      const runtime = f.input.environmentRuntime;
+      const resume = runtime.resumeWorkspaceRecoveryLease.bind(runtime);
+      vi.spyOn(runtime, "resumeWorkspaceRecoveryLease").mockImplementationOnce(async input => {
+        if (kind === "changed_claim") await db.update(environmentLeases).set({ metadata: sql`
+          ${environmentLeases.metadata} || '{"pendingCleanupAttemptId":"other-owner"}'::jsonb`
+        }).where(eq(environmentLeases.id, f.lease.id));
+        if (kind === "missing_stop_intent") await db.update(environmentLeases).set({ metadata: sql`
+          ${environmentLeases.metadata} - 'sandboxStopAndRetain'`
+        }).where(eq(environmentLeases.id, f.lease.id));
+        return resume({ ...input,
+          ...(kind === "wrong_lease" ? { leaseId: randomUUID() } : {}),
+          ...(kind === "wrong_environment" ? { environmentId: randomUUID() } : {}),
+        });
+      });
+      await recoverRetainedWorkspace(f.input);
+      expect(readSystemRecovery((await action(f.actionId)).evidence)?.failure?.code).toBe("source_binding_invalid");
+      expect(f.runtime.resumeRunLease).not.toHaveBeenCalled();
+      expect(await readFile(path.join(f.remote, "saved.txt"), "utf8")).toBe("finished work\n");
+      expect(await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.companyId, f.companyId))).toHaveLength(1);
+    }, 30_000);
+
  it("keeps verified copyback files in the preserved workspace on ordinary Retry", async () => {
    const f = await seed();
    await recoverRetainedWorkspace(f.input);
@@ -169,8 +169,8 @@ export async function recoverRetainedWorkspace(input: Source & { db: Db; environ
        }).catch(() => undefined); }, 30_000);
        let stopSettled = false;
        try {
-          const opened = await input.environmentRuntime.resumeRunLease({ environment: environment as unknown as Environment,
-            lease: cleanupLease as EnvironmentLease });
+          const opened = await input.environmentRuntime.resumeWorkspaceRecoveryLease({ source: input,
+            leaseId: cleanupLease.id, environmentId: environment.id, ownership });
          if (opened?.providerLeaseId !== lease.providerLeaseId || opened.metadata?.remoteCwd !== remoteCwd)
            throw new SystemRecoveryFailure("source_binding_invalid", "service_required");
          await inspectWorkspaceRecoverySource(db, input);