Re: Log notice that checkpoint is to be written on shutdown

Hi,

Am Donnerstag, den 02.10.2014, 08:17 -0700 schrieb David G Johnston:
> Michael Banck-2 wrote
> > I've attached a trivial patch for this, should it be added to the next
> > commitfest?
>
> Peeking at this provokes a couple of novice questions:
>
> While apparently it is impossible to have a checkpoint to log at recovery
> end the decision to use the bitwise-OR instead of the local "shutdown" bool
> seemed odd at first...

Good point, using `shutdown' is probably better.  I guess that local
bool had escaped my notice when I first had a look at this a while ago.

> LogCheckpointStart creates the log entry unconditionally - leaving the
> caller responsible for evaluating log_checkpoints - while LogCheckpointEnd
> has the log_checkpoints condition built into it.

I noticed this as well.  My guess would be that this is because
LogCheckpointEnd() also updates the BgWriterStats statistics, and should
do that every time, not just when the checkpoint is getting logged.
Whether or not log_checkpoint is set (and logging should happen) is
evaluated directly after that.

Some refactoring of this might be useful (or there might be a very good
reason why this was done like this), but this seems outside the scope of
this patch.  AIUI, shutdown will not be further delayed after the
checkpoint is finished, so no further logging is required for the
purpose of this patch.

> The rationale makes perfect sense.  +1

Cool!

I have attached an updated patch using the `shutdown' bool.


Michael

--
Michael Banck
Projektleiter / Berater
Tel.: +49 (2161) 4643-171
Fax:  +49 (2161) 4643-100
Email: michael.banck@credativ.de

credativ GmbH, HRB Mönchengladbach 12080
USt-ID-Nummer: DE204566209
Hohenzollernstr. 133, 41061 Mönchengladbach
Geschäftsführung: Dr. Michael Meskes, Jörg Folz, Sascha Heuer