Dear Haiyang, Alexander,
>LGFM. For my observation, the most case failed on AsserTXNOrder is checking empty
>decoded transaction. Maybe we should pay more attention to review ReorderBufferTXNIsEmpty.
While checking on my fresh eyes, I thought the code might be wrong. The first_lsn
would be same as previous one, when the *PREVIOUS transaction* (not cur_txn) was
empty. Thought?
Also I found the crash reported on [1] was not resolved by the patch. I'm
still analyzing but I have not found the good reproducer yet which could be done
by spec file. Can someone find the workload?
[1]:
https://www.postgresql.org/message-id/TYCPR01MB12077573479C5A2471BDE8065F5232%40TYCPR01MB12077.jpnprd01.prod.outlook.com
Best Regards,
Hayato Kuroda
FUJITSU LIMITED
https://www.fujitsu.com/global/
