From: | Andreas Pflug <pgadmin(at)pse-consulting(dot)de> |
---|---|
To: | Greg Sabino Mullane <greg(at)turnstep(dot)com> |
Cc: | pgsql-performance(at)postgresql(dot)org, Josh Berkus <josh(at)agliodbs(dot)com> |
Subject: | Re: Configurator project launched |
Date: | 2005-06-22 08:30:51 |
Message-ID: | 42B921BB.3000705@pse-consulting.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
Greg Sabino Mullane wrote:
>
>
>>>>Um, can't we just get that from pg_settings?
>>>>
>>>>Anyway, I'll be deriving settings from the .conf file, since most of the
>>>>time the Configurator will be run on a new installation.
>>>
>>>Aren't most of the settings all kept in the SHOW variables anyway?
>
>
> As I said, it may not be applicable to this project, but thought I would
> offer. One gotcha the module makes transparent is that in older versions of
> PG, the variables are returned in a different way (via RAISE). My module
> will allow you to get the configuration for any connected database, for
> any configuration file, and the defaults for any known version, and do
> quick comparisons between them all. So you could use it to see what has
> changed between a particular server and its conf file, or the differences
> between two conf files, or the differences between two databases, or even show
> what has changed in the default conf file from 7.4.7 and 8.0.1. It will also
> allow you to rewrite the conf files in a standard way.
Sounds a little similar to what's in pgAdmin CVS right now. The
configuration editor can retrieve the config file and display configured
and active setting concurrently, together with explanations taken from
pg_settings (when not run against a pgsql server but a file current
settings are missing, comments are taken from a pg_settings csv dump).
There's the infrastructure to give hints about all settings, with very
few currently implemented.
I wonder if this could be combined with the configurator somehow.
Currently, integration won't work with Perl, so maybe C for the core and
Perl for the interactive part would be better.
Regards,
Andreas
From | Date | Subject | |
---|---|---|---|
Next Message | Tobias Brox | 2005-06-22 08:39:21 | Re: Querying 19million records very slowly |
Previous Message | Kjell Tore Fossbakk | 2005-06-22 08:18:30 | Re: Querying 19million records very slowly |