Fix job restart flow

- Restarted jobs set to 'queued' status
- Cluster master looks for 'queued' jobs, sets to 'processing', then assigns
- Proper job lifecycle: queued -> processing -> assigned -> completed/failed

vidai/cluster_master.py

@@ -1292,7 +1292,7 @@ class ClusterMaster:
                 from .database import get_db_connection
                 conn = get_db_connection()
                 cursor = conn.cursor()
-                cursor.execute('SELECT * FROM processing_queue WHERE status = ? AND (job_id IS NULL OR job_id = ?)', ('processing', ''))
+                cursor.execute('SELECT * FROM processing_queue WHERE status = ?', ('queued',))
                 jobs = cursor.fetchall()
                 conn.close()
 
@@ -1305,6 +1305,10 @@ class ClusterMaster:
 
                 for job_row in jobs:
                     job = dict(job_row)
+                    # Set status to processing before assigning
+                    from .database import update_queue_status
+                    update_queue_status(job['id'], 'processing')
+
                     job['data'] = json.loads(job['data']) if job['data'] else None
                     process_type = 'analysis' if job['request_type'] == 'analyze' else ('training' if job['request_type'] == 'train' else job['request_type'])
                     if should_print_status:
@@ -1313,16 +1317,13 @@ class ClusterMaster:
                     if worker_key:
                         job_id = await self.assign_job_to_worker(worker_key, job['data'])
                         if job_id:
-                            from .database import update_queue_status
                             update_queue_status(job['id'], 'processing', {'job_id': job_id, 'status': 'Assigned to worker'}, job_id=job_id)
                             print(f"Assigned job {job['id']} to worker {worker_key} with job_id {job_id}")
                         else:
                             print(f"Failed to assign job {job['id']} to worker {worker_key}, re-queuing")
-                            from .database import update_queue_status
                             update_queue_status(job['id'], 'queued', error='Failed to assign to worker, re-queued')
                     else:
                         print(f"No suitable worker found for job {job['id']}, re-queuing")
-                        from .database import update_queue_status
                         update_queue_status(job['id'], 'queued', error='No suitable worker found, re-queued')
 
                 await asyncio.sleep(5)  # Poll every 5 seconds

vidai/queue.py

@@ -95,8 +95,8 @@ class QueueManager:
         if job['status'] != 'cancelled':
             return False
 
-        # Reset retry_count and update status to processing (so cluster master picks it up)
-        update_queue_status(queue_id, 'processing', retry_count=0, job_id='')
+        # Reset retry_count and update status to queued (so cluster master picks it up)
+        update_queue_status(queue_id, 'queued', retry_count=0)
 
         return True