From: | Fujii Masao <masao(dot)fujii(at)gmail(dot)com> |
---|---|
To: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> |
Cc: | Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: pg_rewind and log messages |
Date: | 2015-04-08 02:16:48 |
Message-ID: | CAHGQGwHAAKuBGa+VN_OW-z==_HwnJNPa_HtUs9yDSQL2DqzMoA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Wed, Apr 8, 2015 at 5:53 AM, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> wrote:
> Heikki Linnakangas wrote:
>> On 04/07/2015 05:59 AM, Michael Paquier wrote:
>
>> >Fix inconsistent handling of logs in pg_rewind
>> >
>> >pg_rewind was handling a couple of things differently compared to the
>> >other src/bin utilities:
>> >- Logging output needs to be flushed on stderr, not stdout
>>
>> Agreed in general. But there's also precedent in printing some stuff to
>> stdout: pg_ctl does that for the status message, like "server starting". As
>> does initdb.
>>
>> I'm pretty unclear on what the rule here is.
>
> One principle that sometimes helps is to consider what happens if you
> use the command as part of a larger pipeline; progress messages can be
> read by some other command further down (and perhaps report them in a
> dialog box, if you embed the program in a GUI, say), but error messages
> should probably be processed differently; normally the pipeline would be
> aborted as a whole.
Make sense.
Regards,
--
Fujii Masao
From | Date | Subject | |
---|---|---|---|
Next Message | Peter Geoghegan | 2015-04-08 02:24:32 | Re: Tuple visibility within a single XID |
Previous Message | Qingqing Zhou | 2015-04-08 02:16:43 | Re: Tuple visibility within a single XID |