[Bug]: 堆栈显示该线程在执行 infinity::RAGAnalyzer::DFS（深度优先搜索）函数时陷入了极深的递归 。

[anole1980]

Is there an existing issue for the same bug?

• I have checked the existing issues.

Version or Commit ID

v0.7.0-dev2

Other environment information

硬件：Linux VM
OS: Debian 12

Actual behavior and How to reproduce it

infinity_stack.txt

Expected behavior

No response

Additional information

infinity进程持续占用约200% CPU，已经十几个小时了，日志中一直重复出现类似下面的内容：

......
[01:36:22.892] [246] [info] Optimize all indexes begin ts: 10395

[01:36:32.893] [246] [info] Optimize all indexes begin ts: 10395

[01:36:34.755] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:36:34.755] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:36:34.765] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:36:34.765] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:36:42.893] [396] [info] Trigger periodic checkpoint, after 30 seconds

[01:36:42.893] [244] [info] checkpoint ts: 10395, txn: 35798

[01:36:42.893] [244] [info] Last checkpoint ts 10393, this checkpoint begin ts: 10395, SKIP CHECKPOINT

[01:36:42.893] [244] [info] Cleaning ts < 10393 dropped entities...

[01:36:42.893] [246] [info] Optimize all indexes begin ts: 10395

[01:36:42.894] [245] [info] Compact begin ts: 10395

[01:36:42.894] [245] [info] Compact begin ts: 10395

[01:36:42.894] [245] [info] Compact begin ts: 10395

[01:36:52.894] [246] [info] Optimize all indexes begin ts: 10397

[01:37:02.895] [246] [info] Optimize all indexes begin ts: 10397

[01:37:12.896] [396] [info] Trigger periodic checkpoint, after 30 seconds

[01:37:12.896] [244] [info] checkpoint ts: 10397, txn: 35807

[01:37:12.896] [246] [info] Optimize all indexes begin ts: 10397

[01:37:12.897] [244] [info] CurrentObjFinalizeNoLock added empty object 019cf446-21eb-7104-9256-f6f3066159eb

[01:37:12.897] [244] [info] PersistenceManager::CheckValid part 1: 0.065299 ms, part2: 0.00036 ms

[01:37:12.897] [238] [info] Checkpoint begin at 10397, cur txn commit_ts: 10398, txn_id: 35807, swap to new wal file

[01:37:12.897] [238] [info] Wal /var/infinity/wal/wal.log swap to new path: /var/infinity/wal/wal.log.10396, error_if_duplicate: true

[01:37:12.897] [238] [info] Open new wal file /var/infinity/wal/wal.log

[01:37:22.897] [246] [info] Optimize all indexes begin ts: 10399

[01:37:23.107] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.107] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.119] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.119] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.243] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.244] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.257] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.258] [23] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.268] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.269] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.279] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.281] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.693] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.694] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.703] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:23.703] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.617] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.618] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.628] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.629] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.645] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.645] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.656] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:24.657] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:27.683] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:27.683] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:27.693] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:27.693] [22] [info] DEBUG: Using cached index readers for table_id: '3'

[01:37:32.897] [246] [info] Optimize all indexes begin ts: 10399

[01:37:42.898] [396] [info] Trigger periodic checkpoint, after 30 seconds

[01:37:42.898] [244] [info] checkpoint ts: 10399, txn: 35881

[01:37:42.898] [244] [info] Last checkpoint ts 10397, this checkpoint begin ts: 10399, SKIP CHECKPOINT

[01:37:42.898] [244] [info] Cleaning ts < 10397 dropped entities...

[01:37:42.898] [246] [info] Optimize all indexes begin ts: 10399

[01:37:52.899] [246] [info] Optimize all indexes begin ts: 10401

[01:38:02.900] [246] [info] Optimize all indexes begin ts: 10401

[01:38:12.900] [396] [info] Trigger periodic checkpoint, after 30 seconds

[01:38:12.901] [244] [info] checkpoint ts: 10401, txn: 35886

[01:38:12.901] [246] [info] Optimize all indexes begin ts: 10401

[01:38:12.901] [244] [info] CurrentObjFinalizeNoLock added empty object 019cf446-21eb-7104-9256-f6f3066159eb

[01:38:12.901] [244] [info] PersistenceManager::CheckValid part 1: 0.0633 ms, part2: 0.00037 ms

[01:38:12.901] [238] [info] Checkpoint begin at 10401, cur txn commit_ts: 10402, txn_id: 35886, swap to new wal file

[01:38:12.901] [238] [info] Wal /var/infinity/wal/wal.log swap to new path: /var/infinity/wal/wal.log.10400, error_if_duplicate: true

[01:38:12.902] [238] [info] Open new wal file /var/infinity/wal/wal.log

[01:38:22.901] [246] [info] Optimize all indexes begin ts: 10403

[01:38:32.902] [246] [info] Optimize all indexes begin ts: 10403
......

[yingfeng]

Can you paste your query here?

[anole1980]

Sorry, no query logged Yingfeng ***@***.***> 于 2026年3月16日周一 20:52写道：

…

*yingfeng* left a comment (infiniflow/infinity#3314) <#3314 (comment)> Can you paste your query here? — Reply to this email directly, view it on GitHub <#3314 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AED4F6GR5DHZ3SFES6RI6WT4Q72INAVCNFSM6AAAAACWS2GTTWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DANRXGQ2DGMJUGM> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[yingfeng]

If you are using RAGFlow, you can also attach the file here

[anole1980]

That file is kind confidential, can not be shown without permission, sorry 😞. But, what I can tell you is this file is a image-based huge pdf, with lots of table in it, and this bug was triggered in index optimization process. Yingfeng ***@***.***> 于 2026年3月16日周一 21:15写道：

…

*yingfeng* left a comment (infiniflow/infinity#3314) <#3314 (comment)> If you are using RAGFlow, you can also attach the file here — Reply to this email directly, view it on GitHub <#3314 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AED4F6FI7OJDVKHN4MRNASL4Q747JAVCNFSM6AAAAACWS2GTTWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DANRXGU4DGOJTGA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[yingfeng]

Then you can try it after #3318 is packaged into latest Infinity, say v0.7.0.dev4