diff --git a/src/ai/backend/manager/registry.py b/src/ai/backend/manager/registry.py
index ffd82447ac..283b9f3877 100644
--- a/src/ai/backend/manager/registry.py
+++ b/src/ai/backend/manager/registry.py
@@ -1711,11 +1711,11 @@ async def _update_kernel() -> None:
 
         await execute_with_retry(_update_kernel)
 
-        async with self.agent_cache.rpc_context(
-            agent_alloc_ctx.agent_id,
-            order_key=str(scheduled_session.id),
-        ) as rpc:
-            try:
+        try:
+            async with self.agent_cache.rpc_context(
+                agent_alloc_ctx.agent_id,
+                order_key=str(scheduled_session.id),
+            ) as rpc:
                 get_image_ref = lambda k: image_infos[str(k.image_ref)].image_ref
                 # Issue a batched RPC call to create kernels on this agent
                 # created_infos = await rpc.call.create_kernels(
@@ -1783,42 +1783,44 @@ async def _update_kernel() -> None:
                     [binding.kernel.id for binding in items],
                     agent_alloc_ctx.agent_id,
                 )
-            except (asyncio.TimeoutError, asyncio.CancelledError):
-                log.warning("_create_kernels_in_one_agent(s:{}) cancelled", scheduled_session.id)
-            except Exception as e:
-                # The agent has already cancelled or issued the destruction lifecycle event
-                # for this batch of kernels.
-                ex = e
-                for binding in items:
-                    kernel_id = binding.kernel.id
-
-                    async def _update_failure() -> None:
-                        async with self.db.begin_session() as db_sess:
-                            now = datetime.now(tzutc())
-                            query = (
-                                sa.update(KernelRow)
-                                .where(KernelRow.id == kernel_id)
-                                .values(
-                                    status=KernelStatus.ERROR,
-                                    status_info=f"other-error ({ex!r})",
-                                    status_changed=now,
-                                    terminated_at=now,
-                                    status_history=sql_json_merge(
-                                        KernelRow.status_history,
-                                        (),
-                                        {
-                                            KernelStatus.ERROR.name: (
-                                                now.isoformat()
-                                            ),  # ["PULLING", "PREPARING"]
-                                        },
-                                    ),
-                                    status_data=convert_to_status_data(ex, self.debug),
-                                )
+        except (asyncio.TimeoutError, asyncio.CancelledError):
+            log.warning("_create_kernels_in_one_agent(s:{}) cancelled", scheduled_session.id)
+        except Exception as e:
+            ex = e
+            err_info = convert_to_status_data(ex, self.debug)
+
+            # The agent has already cancelled or issued the destruction lifecycle event
+            # for this batch of kernels.
+            for binding in items:
+                kernel_id = binding.kernel.id
+
+                async def _update_failure() -> None:
+                    async with self.db.begin_session() as db_sess:
+                        now = datetime.now(tzutc())
+                        query = (
+                            sa.update(KernelRow)
+                            .where(KernelRow.id == kernel_id)
+                            .values(
+                                status=KernelStatus.ERROR,
+                                status_info=f"other-error ({ex!r})",
+                                status_changed=now,
+                                terminated_at=now,
+                                status_history=sql_json_merge(
+                                    KernelRow.status_history,
+                                    (),
+                                    {
+                                        KernelStatus.ERROR.name: (
+                                            now.isoformat()
+                                        ),  # ["PULLING", "PREPARING"]
+                                    },
+                                ),
+                                status_data=err_info,
                             )
-                            await db_sess.execute(query)
+                        )
+                        await db_sess.execute(query)
 
-                    await execute_with_retry(_update_failure)
-                raise
+                await execute_with_retry(_update_failure)
+            raise
 
     async def create_cluster_ssh_keypair(self) -> ClusterSSHKeyPair:
         key = rsa.generate_private_key(