Обсуждение: BUG #17640: fork off the requested timeline before the switchpoint.

Поиск
Список
Период
Сортировка

BUG #17640: fork off the requested timeline before the switchpoint.

От
PG Bug reporting form
Дата:
The following bug has been logged on the website:

Bug reference:      17640
Logged by:          Mukesh Singh
Email address:      mukeshsingh.singh19@gmail.com
PostgreSQL version: 11.5
Operating system:   linux
Description:

2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.969 WAT [63170] LOG:  restored
log file "0000003400004E7500000029" from archive
2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.976 WAT [63170] FATAL:
requested timeline 53 is not a child of this server's history
2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.976 WAT [63170] DETAIL:  Latest
checkpoint is at 4E75/2930CB38 on timeline 52, but in the history of the
requested timeline, the server forked off from that timeline at
4B77/F5FFF080.
2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.977 WAT [63168] LOG:  startup
process (PID 63170) exited with exit code 1
2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.977 WAT [63168] LOG:  aborting
startup due to startup process failure
2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.978 WAT [63168] LOG:  database
system is shut down


Re: BUG #17640: fork off the requested timeline before the switchpoint.

От
Michael Paquier
Дата:
On Fri, Oct 14, 2022 at 05:21:48AM +0000, PG Bug reporting form wrote:
> 2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.969 WAT [63170] LOG:  restored
> log file "0000003400004E7500000029" from archive
> 2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.976 WAT [63170] FATAL:
> requested timeline 53 is not a child of this server's history
> 2022-10-14 06:21:03 WAT 2022-10-14 06:21:03.976 WAT [63170] DETAIL:  Latest
> checkpoint is at 4E75/2930CB38 on timeline 52, but in the history of the
> requested timeline, the server forked off from that timeline at
> 4B77/F5FFF080.

Well, this points out that the timeline you have selected is not the
one your timeline history file likes.  I fail to see why this would be
a bug without more details about your configuration (including
archiving strategy) and what you did with your set of servers when
doing a failover, for example.
--
Michael

Вложения