Re: PATCH: regular logging of checkpoint progress - Mailing list pgsql-hackers

From Tomas Vondra
Subject Re: PATCH: regular logging of checkpoint progress
Date
Msg-id 93225ed61ca62fa906649233eb88ce75.squirrel@sq.gransy.com
Whole thread Raw
In response to Re: PATCH: regular logging of checkpoint progress  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: PATCH: regular logging of checkpoint progress
Re: PATCH: regular logging of checkpoint progress
List pgsql-hackers
On 2 Září 2011, 16:13, Tom Lane wrote:
> Peter Eisentraut <peter_e@gmx.net> writes:
>> On fre, 2011-09-02 at 11:01 +0200, Tomas Vondra wrote:
>>> What about logging it with a lower level, e.g. NOTICE instead of the
>>> current LOG? If that's not a solution then a new GUC is needed I
>>> guess.
>
>> Changing the log level is not the appropriate solution.  Make it a
>> configuration parameter.
>
> Frankly, logging as verbose as this is proposed to be is something
> that I can't imagine anybody wanting at all, especially not in
> production environments.  DEBUG3 or somewhere around there seems fine.

Yes, it can be a bit verbose, but I don't think it's that verbose. I'm one
of those this information about checkpoint progress may be a valuable info
- even on production systems - that's why I don't want to put that into
the debug levels.

It should write about 10 lines for each checkpoint (altough I've just
realized the current patch won't do that for the timed checkpoints - it'll
probably log significantly more).

My 'ideal' solution would be either to add another GUC (to turn this
on/off) or allow log_checkpoints to have three values

log_checkpoints = {off, normal, detailed}

where 'normal' provides the current output and 'detail' produces this much
verbose output.

Tomas



pgsql-hackers by date:

Previous
From: Magnus Hagander
Date:
Subject: Re: PATCH: regular logging of checkpoint progress
Next
From: "Tomas Vondra"
Date:
Subject: Re: PATCH: regular logging of checkpoint progress