From 3e6add90debe8cd261f920fc1564e25540e8935b Mon Sep 17 00:00:00 2001
From: Alexander Knigge <166455923+bigph00t@users.noreply.github.com>
Date: Tue, 27 Jan 2026 23:40:19 -0800
Subject: [PATCH] fix: prevent stale memory_session_id resume crash after
 worker restart (Issue #817) (#839)

When the worker restarts, the SDK context is lost but the database still contains
memory_session_id values from the previous worker instance. The existing guard
(lastPromptNumber > 1) doesn't protect against this because lastPromptNumber is
also loaded from the database.

This fix:
- Clears memory_session_id when initializing a session from DB (not from cache)
- Adds warning log when discarding stale session IDs
- Lets SDK agent capture fresh memory_session_id on first response

The key insight: if a session is not in memory, we're in a new worker instance,
and any database memory_session_id is definitely stale.

Fixes #817
Related to #825

Co-authored-by: bigphoot <bigphoot@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
---
 src/services/worker/SessionManager.ts | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)

diff --git a/src/services/worker/SessionManager.ts b/src/services/worker/SessionManager.ts
index 8915d6ed..34ab6460 100644
--- a/src/services/worker/SessionManager.ts
+++ b/src/services/worker/SessionManager.ts
@@ -106,6 +106,15 @@ export class SessionManager {
       memory_session_id: dbSession.memory_session_id
     });
 
+    // Log warning if we're discarding a stale memory_session_id (Issue #817)
+    if (dbSession.memory_session_id) {
+      logger.warn('SESSION', `Discarding stale memory_session_id from previous worker instance (Issue #817)`, {
+        sessionDbId,
+        staleMemorySessionId: dbSession.memory_session_id,
+        reason: 'SDK context lost on worker restart - will capture new ID'
+      });
+    }
+
     // Use currentUserPrompt if provided, otherwise fall back to database (first prompt)
     const userPrompt = currentUserPrompt || dbSession.user_prompt;
 
@@ -124,11 +133,15 @@ export class SessionManager {
     }
 
     // Create active session
-    // Load memorySessionId from database if previously captured (enables resume across restarts)
+    // CRITICAL: Do NOT load memorySessionId from database here (Issue #817)
+    // When creating a new in-memory session, any database memory_session_id is STALE
+    // because the SDK context was lost when the worker restarted. The SDK agent will
+    // capture a new memorySessionId on the first response and persist it.
+    // Loading stale memory_session_id causes "No conversation found" crashes on resume.
     session = {
       sessionDbId,
       contentSessionId: dbSession.content_session_id,
-      memorySessionId: dbSession.memory_session_id || null,
+      memorySessionId: null,  // Always start fresh - SDK will capture new ID
       project: dbSession.project,
       userPrompt,
       pendingMessages: [],
@@ -143,10 +156,11 @@ export class SessionManager {
       currentProvider: null  // Will be set when generator starts
     };
 
-    logger.debug('SESSION', 'Creating new session object', {
+    logger.debug('SESSION', 'Creating new session object (memorySessionId cleared to prevent stale resume)', {
       sessionDbId,
       contentSessionId: dbSession.content_session_id,
-      memorySessionId: dbSession.memory_session_id || '(none - fresh session)',
+      dbMemorySessionId: dbSession.memory_session_id || '(none in DB)',
+      memorySessionId: '(cleared - will capture fresh from SDK)',
       lastPromptNumber: promptNumber || this.dbManager.getSessionStore().getPromptNumberFromUserPrompts(dbSession.content_session_id)
     });