Обсуждение: Any way to understand state after data corruption failures during startup.

Поиск
Список
Период
Сортировка

Any way to understand state after data corruption failures during startup.

От
Harinath Kanchu
Дата:
Hello,

If postgres fail when it is starting as standby due to WAL file corruption, can we get the status through any exit code
orby parsing pg_control file ? 

I can see the  “XX001” code is emitted in logs but parsing logs doesn’t seem to be a good option. Please suggest.

Thanks.




Re: Any way to understand state after data corruption failures during startup.

От
Kyotaro Horiguchi
Дата:
At Thu, 16 Jun 2022 19:46:25 -0700, Harinath Kanchu <hkanchu@apple.com> wrote in 
> 
> Hello,
> 
> If postgres fail when it is starting as standby due to WAL file corruption, can we get the status through any exit
codeor by parsing pg_control file ?
 
> 
> I can see the  “XX001” code is emitted in logs but parsing logs doesn’t seem to be a good option. Please suggest.

I'm not sure exactly what is what you want to achieve here, but there
is a negative evidence for the reliability of detecting WAL corruption
failure without searching server logs.

[64923:startup] [XX000] 2022-06-17 13:30:52.493 JST PANIC:  incorrect resource manager data checksum in record at
0/3000060

At least in this case, there's no trace of a server-stop due to WAL
corruption left behind other than in server log.

regards.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center