| From: | "Zhijie Hou (Fujitsu)" <houzj(dot)fnst(at)fujitsu(dot)com> |
|---|---|
| To: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Buildfarm failure on tamandua - "timed out waiting for subscriber to synchronize data" |
| Date: | 2024-03-21 09:47:12 |
| Message-ID: | OS0PR01MB5716DB29841070E0DABCB9E594322@OS0PR01MB5716.jpnprd01.prod.outlook.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi,
There is a failure in 040_standby_failover_slots_sync on tamandua:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tamandua&dt=2024-03-21%2008%3A28%3A58
The reason of the timeout is that the table sync worker for the new table is not started
after executing ALTER SUBSCRIPTION REFRESH PUBLICATION.
We have seen similar failures in other tests as well[1]. AFAICS, the reasons of
them are the same which is because of a race condition in logicalrep apply worker. The
analysis has been posted on another thread[2] and the fix is also being
reviewed.
[1] https://www.postgresql.org/message-id/OSZPR01MB6310D6F48372F52F1D85E1C5FD609%40OSZPR01MB6310.jpnprd01.prod.outlook.com
[2] https://www.postgresql.org/message-id/flat/CALDaNm1XeB3bF%2BVEJZi%3DBT31PZAL_UVys-26%2BYSv_AxCq0G2eg%40mail.gmail.com#87b153fd7676652746406a6f114eb67b
Best Regards,
Hou Zhijie
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Shlok Kyal | 2024-03-21 09:49:16 | Re: speed up a logical replica setup |
| Previous Message | Jelte Fennema-Nio | 2024-03-21 09:44:17 | Re: Flushing large data immediately in pqcomm |