Re: Customized Options Threshold Error - Mailing list pgsql-hackers

From
Subject Re: Customized Options Threshold Error
Date
Msg-id A9C510524E235E44AE909CD4027AE196BF7D6FB27A@MBX-MSG-SV03.msg.nttdata.co.jp
Whole thread Raw
In response to Re: Customized Options Threshold Error  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
> > When we specify a value which exceeds valid range in "Customized
> > Options" , its behavior is different from "Parameter Interaction via
> Configuration File" behavior.
> 
> > In case of "Parameter Interaction via Configuration File", it finish
> with FATAL error when it get threshold error.
> > But in "Customized Options", it just printout WARNING and continue the
> process with default value.
> 
> I think this is based on a misunderstanding.  Bad values in
> postgresql.conf only result in a FATAL error at initial postmaster
> startup; if you change them and SIGHUP the server, an erroneous new value
> just results in a WARNING.  This is because we don't really want the
> server crashing once it's up.  The case you're showing with registering
> a new background worker is something that happens after the server is
> up, so it's consistent for it to produce a WARNING not a fatal error.

Thanks for the reply.

> registering a new background worker is something that happens after the server is up

I understand that parameters which are set by "Customized Options" are loaded after server is up, same as like "pg_ctl
reload",so even if invalid parameters are specified postmaster just ignore them and results in a WARNING.
 

However, "Customized Options" always treated as "pg_ctl reload" is little bit strange to me.

Regards,

--
Furuya Osamu


pgsql-hackers by date:

Previous
From: Magnus Hagander
Date:
Subject: Re: PostgreSQL doesn't stop propley when --slot option is specified with pg_receivexlog.
Next
From: Ashutosh Bapat
Date:
Subject: Re: postgres_fdw behaves oddly